removed unused import

.gitignore

@@ -35,4 +35,6 @@ webview-ui/src/services/grpc-client.ts
 # Host bridge
 src/hosts/vscode/*/methods.ts
 src/hosts/vscode/*/index.ts
+src/hosts/vscode/client/host-grpc-client.ts
 src/hosts/vscode/host-grpc-service-config.ts
+src/standalone/server-setup.ts

.prettierignore

@@ -5,3 +5,4 @@ webview-ui/build/
 package-lock.json
 src/core/prompts/system.ts
 src/core/prompts/model_prompts/claude4.ts
+evals/

evals/.gitignore

@@ -2,5 +2,21 @@ repositories
 
 results/evals.db
 
+diff-edits/cases/
+diff-edits/results/
+
+# Environment variables
+.env
+
+# backwards compatible
 diff_editing/test_cases/
-diff_editing/test_outputs/
+diff_editing/test_outputs/
+
+*.db
+*.db-wal
+*.db-shm
+
+.cache
+
+# Python bytecode cache
+*__pycache__/

evals/README.md

@@ -17,6 +17,7 @@ The evaluation system consists of two main components:
 
 1. **Test Server**: Enhanced HTTP server in `src/services/test/TestServer.ts` that provides detailed task results
 2. **CLI Tool**: Command-line interface in `evals/cli/` for orchestrating evaluations
+3. **Diff Edit Benchmark**: Separate command using the CLI tool that runs a comprehensive diff editing benchmark suite on real world cases, along with a streamlit dashboard displaying the results. For more details, see the [Diff Edit Benchmark README](./diff-edits/README.md). Make sure you add a `evals/diff-edits/cases` folder with all the conversation jsons. 
 
 ## Directory Structure
 
@@ -40,6 +41,12 @@ cline-repo/
 │   │   │   └── utils/                # Utility functions
 │   │   ├── package.json
 │   │   └── tsconfig.json
+│   ├── diff-edits/                   # Diff editing evaluation suite
+│   │   ├── cases/                    # Test case JSON files
+│   │   ├── results/                  # Evaluation results
+│   │   ├── diff-apply/               # Diff application logic
+│   │   ├── parsing/                  # Assistant message parsing
+│   │   └── prompts/                  # System prompts
 │   ├── repositories/                 # Cloned benchmark repositories
 │   │   ├── exercism/                 # Modified Exercism (from pashpashpash/evals)
 │   │   ├── swe-bench/                # SWE-Bench repository
@@ -148,6 +155,192 @@ Freelance-style programming tasks from the SWELancer benchmark.
 
 Multi-file software engineering tasks from the Multi-SWE-Bench repository.
 
+## Diff Edit Evaluations
+
+The Cline Evaluation System includes a specialized suite for evaluating how well models can make precise edits to files using the `replace_in_file` tool.
+
+### Overview
+
+Diff edit evaluations test a model's ability to:
+
+1. Understand file content and identify specific sections to modify
+2. Generate correct SEARCH/REPLACE blocks for targeted edits
+3. Successfully apply changes without introducing errors
+
+### Directory Structure
+
+```
+diff-edits/
+├── cases/                  # Test case JSON files
+├── results/                # Evaluation results
+├── ClineWrapper.ts         # Wrapper for model interaction
+├── TestRunner.ts           # Main test execution logic
+├── types.ts                # Type definitions
+├── diff-apply/             # Diff application logic
+├── parsing/                # Assistant message parsing
+└── prompts/                # System prompts
+```
+
+### Creating Test Cases
+
+Test cases are defined as JSON files in the `diff-edits/cases/` directory. Each test case should include:
+
+```json
+{
+  "test_id": "example_test_1",
+  "messages": [
+    {
+      "role": "user",
+      "text": "Please fix the bug in this code...",
+      "images": []
+    },
+    {
+      "role": "assistant",
+      "text": "I'll help you fix that bug..."
+    }
+  ],
+  "file_contents": "// Original file content here\nfunction example() {\n  // Code with bug\n}",
+  "file_path": "src/example.js",
+  "system_prompt_details": {
+    "mcp_string": "",
+    "cwd_value": "/path/to/working/directory",
+    "browser_use": false,
+    "width": 900,
+    "height": 600,
+    "os_value": "macOS",
+    "shell_value": "/bin/zsh",
+    "home_value": "/Users/username",
+    "user_custom_instructions": ""
+  },
+  "original_diff_edit_tool_call_message": ""
+}
+```
+
+### Running Diff Edit Evaluations
+
+#### Single Model Evaluation
+
+```bash
+cd evals/cli
+node dist/index.js run-diff-eval --model-ids "anthropic/claude-3-5-sonnet-20241022"
+```
+
+#### Multi-Model Evaluation
+
+Compare multiple models in a single evaluation run:
+
+```bash
+# Compare Claude and Grok models
+node dist/index.js run-diff-eval \
+  --model-ids "anthropic/claude-3-5-sonnet-20241022,x-ai/grok-beta" \
+  --max-cases 10 \
+  --valid-attempts-per-case 3 \
+  --verbose
+
+# Compare multiple Claude variants
+node dist/index.js run-diff-eval \
+  --model-ids "anthropic/claude-3-5-sonnet-20241022,anthropic/claude-3-5-haiku-20241022,anthropic/claude-3-opus-20240229" \
+  --max-cases 5 \
+  --valid-attempts-per-case 2 \
+  --parallel
+```
+
+#### Options
+
+- `--model-ids`: Comma-separated list of model IDs to evaluate (required)
+- `--system-prompt-name`: System prompt to use (default: "basicSystemPrompt")
+- `--valid-attempts-per-case`: Number of attempts per test case per model (default: 1)
+- `--max-cases`: Maximum number of test cases to run (default: all available)
+- `--parsing-function`: Function to parse assistant messages (default: "parseAssistantMessageV2")
+- `--diff-edit-function`: Function to apply diffs (default: "constructNewFileContentV2")
+- `--test-path`: Path to test cases (default: diff-edits/cases)
+- `--thinking-budget`: Tokens allocated for thinking (default: 0)
+- `--parallel`: Run tests in parallel (flag)
+- `--replay`: Use pre-recorded LLM output (flag)
+- `--verbose`: Enable detailed logging (flag)
+
+#### Examples
+
+```bash
+# Quick test with 2 models, 4 cases, 2 attempts each
+node dist/index.js run-diff-eval \
+  --model-ids "anthropic/claude-3-5-sonnet-20241022,x-ai/grok-beta" \
+  --max-cases 4 \
+  --valid-attempts-per-case 2 \
+  --verbose
+
+# Comprehensive evaluation with parallel execution
+node dist/index.js run-diff-eval \
+  --model-ids "anthropic/claude-3-5-sonnet-20241022,anthropic/claude-3-5-haiku-20241022" \
+  --system-prompt-name claude4SystemPrompt \
+  --valid-attempts-per-case 5 \
+  --max-cases 20 \
+  --parallel \
+  --verbose
+```
+
+### Database Storage & Analytics
+
+All evaluation results are automatically stored in a SQLite database (`diff-edits/evals.db`) for advanced analytics and comparison. The database includes:
+
+- **System Prompts**: Versioned system prompt content with hashing for deduplication
+- **Processing Functions**: Versioned parsing and diff-edit function configurations
+- **Files**: Original and edited file content with content-based hashing
+- **Runs**: Evaluation run metadata and configuration
+- **Cases**: Individual test case information with context tokens
+- **Results**: Detailed results with timing, cost, and success metrics
+
+### Interactive Dashboard
+
+Launch the Streamlit dashboard to visualize and analyze evaluation results:
+
+```bash
+cd diff-edits/dashboard
+streamlit run app.py
+```
+
+The dashboard provides:
+
+- **Model Performance Comparison**: Side-by-side comparison of success rates, latency, and costs
+- **Interactive Charts**: Success rate trends, latency vs cost analysis, and performance metrics
+- **Detailed Drill-Down**: Individual result analysis with file content viewing
+- **Run Selection**: Browse and compare different evaluation runs
+- **Real-time Updates**: Automatically refreshes with new evaluation data
+
+#### Dashboard Features
+
+1. **Hero Section**: Overview of current run with key metrics
+2. **Model Cards**: Performance cards with grades and detailed metrics
+3. **Comparison Charts**: Interactive Plotly charts for visual analysis
+4. **Result Explorer**: Detailed view of individual test results including:
+   - Original and edited file content
+   - Raw model output
+   - Parsed tool calls
+   - Timing and cost metrics
+   - Error analysis
+
+#### Quick Start Dashboard
+
+```bash
+# Run a quick evaluation
+node cli/dist/index.js run-diff-eval \
+  --model-ids "anthropic/claude-3-5-sonnet-20241022,x-ai/grok-beta" \
+  --max-cases 4 \
+  --valid-attempts-per-case 2 \
+  --verbose
+
+# Launch dashboard to view results
+cd diff-edits/dashboard && streamlit run app.py
+```
+
+### Legacy Results
+
+For backward compatibility, results are also saved as JSON files in the `diff-edits/results/` directory. The JSON results include:
+- Success/failure status
+- Extracted tool calls
+- Diff edit content
+- Token usage and cost metrics
+
 ## Metrics
 
 The evaluation system collects the following metrics:

evals/cli/package.json

@@ -1,39 +0,0 @@
-{
-	"name": "cline-evaluation-cli",
-	"version": "0.1.0",
-	"description": "CLI tool for orchestrating Cline evaluations across multiple benchmarks",
-	"main": "dist/index.js",
-	"scripts": {
-		"build": "tsc",
-		"start": "node dist/index.js",
-		"dev": "ts-node src/index.ts",
-		"test": "echo \"Error: no test specified\" && exit 1"
-	},
-	"keywords": [
-		"cline",
-		"evaluation",
-		"benchmark"
-	],
-	"author": "",
-	"license": "MIT",
-	"dependencies": {
-		"better-sqlite3": "^11.10.0",
-		"chalk": "^4.1.2",
-		"commander": "^9.4.1",
-		"execa": "^5.1.1",
-		"node-fetch": "^2.7.0",
-		"ora": "^5.4.1",
-		"sqlite": "^4.1.2",
-		"uuid": "^9.0.0",
-		"yargs": "^17.6.2"
-	},
-	"devDependencies": {
-		"@types/better-sqlite3": "^7.6.3",
-		"@types/node": "^18.11.18",
-		"@types/node-fetch": "^2.6.12",
-		"@types/uuid": "^9.0.0",
-		"@types/yargs": "^17.0.19",
-		"ts-node": "^10.9.1",
-		"typescript": "^4.9.4"
-	}
-}

evals/cli/src/commands/runDiffEval.ts

@@ -3,9 +3,9 @@ import chalk from "chalk"
 import path from "path"
 
 interface RunDiffEvalOptions {
-	modelId: string
+	modelIds: string
 	systemPromptName: string
-	numberOfRuns: number
+	validAttemptsPerCase: number
 	parsingFunction: string
 	diffEditFunction: string
 	thinkingBudget: number
@@ -14,22 +14,23 @@ interface RunDiffEvalOptions {
 	testPath: string
 	outputPath: string
 	replay: boolean
+	maxCases?: number
 }
 
 export async function runDiffEvalHandler(options: RunDiffEvalOptions) {
 	console.log(chalk.blue("Starting diff editing evaluation..."))
 
 	// Resolve the path to the TestRunner.ts script relative to the current file
-	const scriptPath = path.resolve(__dirname, "../../../diff_editing/TestRunner.ts")
+	const scriptPath = path.resolve(__dirname, "../../../diff-edits/TestRunner.ts")
 
 	// Construct the arguments array for the execa call
 	const args = [
-		"--model-id",
-		options.modelId,
+		"--model-ids",
+		options.modelIds,
 		"--system-prompt-name",
 		options.systemPromptName,
-		"--number-of-runs",
-		String(options.numberOfRuns),
+		"--valid-attempts-per-case",
+		String(options.validAttemptsPerCase),
 		"--parsing-function",
 		options.parsingFunction,
 		"--diff-edit-function",
@@ -59,12 +60,16 @@ export async function runDiffEvalHandler(options: RunDiffEvalOptions) {
 		args.push("--verbose")
 	}
 
+	if (options.maxCases) {
+		args.push("--max-cases", String(options.maxCases))
+	}
+
 	try {
 		console.log(chalk.gray(`Executing: npx tsx ${scriptPath} ${args.join(" ")}`))
 
 		// Execute the script as a child process
 		// We use 'inherit' to stream the stdout/stderr directly to the user's terminal
-		const subprocess = execa("npx", ["tsx", scriptPath, ...args], {
+		const subprocess = execa("npx", ["tsx", "--tsconfig", path.resolve(__dirname, "../../../tsconfig.json"), scriptPath, ...args], {
 			stdio: "inherit",
 		})
 

evals/cli/src/index.ts

@@ -84,9 +84,10 @@ program
 	.description("Run the diff editing evaluation suite")
 	.option("--test-path <path>", "Path to the directory containing test case JSON files")
 	.option("--output-path <path>", "Path to the directory to save the test output JSON files")
-	.option("--model-id <model_id>", "The model ID to use for the test")
+	.option("--model-ids <model_ids>", "Comma-separated list of model IDs to test")
 	.option("--system-prompt-name <name>", "The name of the system prompt to use", "basicSystemPrompt")
-	.option("-n, --number-of-runs <number>", "Number of times to run each test case", "1")
+	.option("-n, --valid-attempts-per-case <number>", "Number of valid attempts per test case per model (will retry until this many valid attempts are collected)", "1")
+	.option("--max-cases <number>", "Maximum number of test cases to run (limits total cases loaded)")
 	.option("--parsing-function <name>", "The parsing function to use", "parseAssistantMessageV2")
 	.option("--diff-edit-function <name>", "The diff editing function to use", "constructNewFileContentV2")
 	.option("--thinking-budget <tokens>", "Set the thinking tokens budget", "0")
@@ -95,11 +96,11 @@ program
 	.option("-v, --verbose", "Enable verbose logging", false)
 	.action(async (options) => {
 		try {
-			// The logic here simplifies slightly
 			const fullOptions = {
 				...options,
-				numberOfRuns: parseInt(options.numberOfRuns, 10),
+				validAttemptsPerCase: parseInt(options.validAttemptsPerCase, 10),
 				thinkingBudget: parseInt(options.thinkingBudget, 10),
+				maxCases: options.maxCases ? parseInt(options.maxCases, 10) : undefined,
 			}
 			await runDiffEvalHandler(fullOptions)
 		} catch (error) {

evals/diff-edits/ClineWrapper.ts

@@ -39,16 +39,22 @@ interface StreamResult {
 		cacheReadTokens: number
 		totalCost: number
 	}
+	timing?: {
+		timeToFirstTokenMs: number
+		timeToFirstEditMs?: number
+		totalRoundTripMs: number
+	}
 }
 
 /**
- * Process the stream and return full response
+ * Process the stream and return full response with timing data
  */
 async function processStream(
 	handler: OpenRouterHandler,
 	systemPrompt: string,
 	messages: Anthropic.Messages.MessageParam[],
 ): Promise<StreamResult> {
+	const startTime = Date.now()
 	const stream = handler.createMessage(systemPrompt, messages)
 
 	let assistantMessage = ""
@@ -58,12 +64,21 @@ async function processStream(
 	let cacheWriteTokens = 0
 	let cacheReadTokens = 0
 	let totalCost = 0
+	
+	// Timing tracking
+	let timeToFirstTokenMs: number | null = null
+	let timeToFirstEditMs: number | null = null
 
 	for await (const chunk of stream) {
 		if (!chunk) {
 			continue
 		}
 
+		// Capture time to first token (any chunk type)
+		if (timeToFirstTokenMs === null) {
+			timeToFirstTokenMs = Date.now() - startTime
+		}
+
 		switch (chunk.type) {
 			case "usage":
 				inputTokens += chunk.inputTokens
@@ -79,10 +94,25 @@ async function processStream(
 				break
 			case "text":
 				assistantMessage += chunk.text
+				
+				// Try to detect first tool call by parsing accumulated message
+				if (timeToFirstEditMs === null) {
+					try {
+						const parsed = parseAssistantMessageV2(assistantMessage)
+						const hasToolCall = parsed.some(block => block.type === "tool_use")
+						if (hasToolCall) {
+							timeToFirstEditMs = Date.now() - startTime
+						}
+					} catch {
+						// Parsing failed, continue accumulating
+					}
+				}
 				break
 		}
 	}
 
+	const totalRoundTripMs = Date.now() - startTime
+
 	return {
 		assistantMessage,
 		reasoningMessage,
@@ -93,6 +123,11 @@ async function processStream(
 			cacheReadTokens,
 			totalCost,
 		},
+		timing: {
+			timeToFirstTokenMs: timeToFirstTokenMs || 0,
+			timeToFirstEditMs: timeToFirstEditMs || undefined,
+			totalRoundTripMs,
+		},
 	}
 }
 
@@ -245,7 +280,22 @@ export async function runSingleEvaluation(input: TestInput): Promise<TestResult>
 		}
 
 		// check that we are editing the correct file path
+		console.log(`Expected file path: "${originalFilePath}"`);
+		console.log(`Actual file path used: "${diffToolPath}"`);
 		if (diffToolPath !== originalFilePath) {
+			console.log(`❌ File path mismatch detected!`);
+			// Enhanced logging:
+			if (streamResult?.assistantMessage) {
+				console.log(`   Full model output (assistantMessage):`);
+				console.log(`   -----------------------------------------`);
+				console.log(`   ${streamResult.assistantMessage}`);
+				console.log(`   -----------------------------------------`);
+			}
+			if (toolCall) {
+				console.log(`   Parsed tool call that caused mismatch:`);
+				console.log(`   ${JSON.stringify(toolCall, null, 2)}`);
+				console.log(`   -----------------------------------------`);
+			}
 			return {
 				success: false,
 				streamResult: streamResult,

evals/diff-edits/README.md

@@ -0,0 +1,71 @@
+# A Note on Cline's Diff Evaluation Setup
+
+Hey there, this note explains what we're doing with Cline's diff evaluation (evals) system. It's all about checking how well various AI models (which users connect to Cline via their own API keys), prompts, and diffing tools can handle file changes.
+
+## What We're Trying to Figure Out
+
+The main idea here is to figure out which AI models (configured by users) are best at making `replace_in_file` tool calls that work correctly. This helps us understand model capabilities and also speeds up our own experiments with prompts and diffing algorithms to make Cline better over time. We want to know a few key things.
+
+First, can the model create diffs, which are just sets of SEARCH and REPLACE blocks, that apply cleanly to a file? This is what we call `diffEditSuccess`.
+
+Second, how do different LLMs, like Claude or Grok, stack up against each other when they try to make these diff edits? We use a standard set of real-world test cases for this.
+
+Third, do different system prompts, say our `basicSystemPrompt` versus the `claude4SystemPrompt`, change how well a model does at diff editing?
+
+Fourth, we're also looking at different ways to apply the diffs themselves. We have a few algorithms like `constructNewFileContentV1`, `V2`, and `V3`, and we want to see which ones are more robust when fed model-generated diffs.
+
+Fifth, we track how fast the model starts making an edit. The `timeToFirstEditMs` metric gives us a hint about how quickly a user would see changes happening in their editor.
+
+And finally, we keep an eye on how many tokens are used and what it costs for each model and each try. This helps us compare how efficient they are.
+
+Right now, these evals are mostly about whether the diff *applies* correctly. That means, do the SEARCH blocks find a match, and can the REPLACE blocks be put in without an error? We're not yet deeply analyzing if the change is valid code or matches what the user *wanted* semantically. That's a problem for another day, and will require a lot more scaffolding.
+
+## How We Run These Tests
+
+Two prerequisites:
+
+1. Make sure you have an `evals/.env` file with `OPENROUTER_API_KEY=<your-openrouter-key>` 
+
+2. Make sure you add a `evals/diff-edits/cases` folder with all the conversation jsons prior to running this. 
+
+
+Our testing strategy is based on replaying situations from actual user sessions where diff edits were tried.
+
+It starts with our test cases. Each one is a JSON file in `./cases` that has the conversation history that led to a diff edit, the original file content and its path, and the info needed to rebuild the system prompt from that original session.
+
+Then, for every test run, we set up a specific configuration. This includes which LLM we're testing, which system prompt it gets, which function we use to parse the model's raw output, and which function we use to actually apply the diff. Here's the command I've been using:
+
+```bash
+npm run diff-eval -- --model-ids "anthropic/claude-3-5-sonnet-20241022,x-ai/grok-3-beta" --max-cases 4 --valid-attempts-per-case 2 --verbose --parallel
+```
+
+This will build the eval script, run it, and then open the streamlit dashboard to show the results.
+
+The `TestRunner.ts` script is the main coordinator. For each test case and setup, `ClineWrapper.ts` takes over and sends the conversation and system prompt to the LLM. We then watch the model's response as it streams in and parse it to find any tool calls.
+
+We're specifically looking for the model to make a single `replace_in_file` tool call. Multiple edits in one tool call are allowed, and recorded (in case you want to filter results by number of edits in a single tool call and compare success rate for that slice across different models/system prompts/etc). If it does, and it's for the correct file, we grab the diff content it produced. Then, the chosen diff application algorithm tries to apply that diff to the original file. We record whether this worked or not as `diffEditSuccess`.
+
+We record a bunch of data for every attempt into a database. This includes details about the model and prompt, token counts, costs, the raw output from the model, the parsed tool calls, whether it succeeded or failed, any error messages, and timing info. For a detailed explanation of the database schema, see [database.md](./database.md).
+
+A big part of this is how we handle "valid attempts," which I'll explain next.
+
+## Keeping it Fair with "Valid Attempts"
+
+LLMs can be unpredictable. If we replay an old scenario, a new model, or even the same model later, might do something completely different than what happened originally. It might call another tool or ask a question instead of trying a diff edit.
+
+Since we really want to test the *diff editing* part, we need a way to make sure we're comparing fairly. That's why we have this idea of "valid attempts."
+
+An attempt is "valid" for this benchmark if the model actually tries to do what we're interested in. This means two things. One, it must call the `replace_in_file` tool. Two, it must target the *same file path* that was targeted in the original recorded conversation for that test case.
+
+If the model does something else, like calling a different tool or picking the wrong file, we don't count that attempt against its diff editing score. Instead, we consider it an "invalid attempt" for *this specific benchmark* and simply re-run that test case with that model. We keep doing this until we've collected a set number of these "valid attempts."
+
+For example, if we ask for 5 valid attempts per test case, the system will keep re-rolling for that case until the model has tried to edit the correct file using the `replace_in_file` tool 5 times. Only then do we look at how many of those 5 valid attempts actually resulted in a successful diff application (`diffEditSuccess`).
+
+This way, if we're comparing two models and one gets a 10% success rate on its valid diff edit attempts, and another gets 90%, we have a much clearer picture of their actual diff-generating capabilities. It avoids muddying the waters with attempts where the model didn't even try to perform the specific action we're evaluating. This approach helps us isolate and measure the diff-editing skill more directly, despite the non-deterministic nature of these models.
+
+## Some known edge cases
+
+I noticed that some of the current conversation jsons in the `./cases` folder are a little big bogus. Here's a running list of these areas:
+
+- ~~What if the conversation json was using a model with a massive context window, like Google Gemini's 1M context window, and we're now re-rolling that case on a smaller context window model like claude 4 (200k) or grok-3 (128k)? It's def gonna fail, and we shouldn't just keep trying. We should have a smart system for selecting which cases we can use given the arguments passed in. For example, if we pass in claude and grok with `max cases = 2`, we shouldn't just pick the first two jsons in the folder. We should go through, use a tokenizer, and make sure it would fit with some padding like 20k tokens. Use tiktoken. 20k padding will be sufficient even though different models tokenize differently.~~
+- There are some weird jsons, where something weird happen, where essentially there's a fluke. Maybe the user was using an extremely dumb model that just hallucinated a fake filepath. Now when we try to reroll that case, we never get a valid case. This can easily be handled by making sure that the file is present before selecting that eval for testing. By "file is present" I mean, that file_contents is present in the eval. Additionally, we should in the streamlit dashboard show cases where getting a valid attempt is a challenge, so we can review those cases more easily and throw them out if they're bogus. Maybe a special tab/page in the dashboard for this purpose. Across all runs / cases, what the most consistently problematic cases are. Pop one open to see the case formatted json with all the user/assistant turns.

evals/diff-edits/TestRunner.ts

@@ -0,0 +1,930 @@
+import { runSingleEvaluation, TestInput, TestResult } from "./ClineWrapper"
+import { basicSystemPrompt } from "./prompts/basicSystemPrompt-06-06-25"
+import { claude4SystemPrompt } from "./prompts/claude4SystemPrompt-06-06-25"
+import { formatResponse } from "./helpers"
+import { Anthropic } from "@anthropic-ai/sdk"
+import * as fs from "fs"
+import * as path from "path"
+import { Command } from "commander"
+import { InputMessage, ProcessedTestCase, TestCase, TestConfig, SystemPromptDetails, ConstructSystemPromptFn } from "./types"
+import { loadOpenRouterModelData, EvalOpenRouterModelInfo } from "./openRouterModelsHelper" // Added import
+import {
+	getDatabase,
+	upsertSystemPrompt,
+	upsertProcessingFunctions,
+	upsertFile,
+	createBenchmarkRun,
+	createCase,
+	insertResult,
+	DatabaseClient,
+	CreateResultInput,
+} from "./database"
+
+// Load environment variables from .env file
+import * as dotenv from "dotenv"
+dotenv.config({ path: path.join(__dirname, "../.env") })
+
+// tiktoken for token counting
+import { get_encoding } from "tiktoken";
+const encoding = get_encoding("cl100k_base"); 
+
+let openRouterModelDataGlobal: Record<string, EvalOpenRouterModelInfo> = {}; // Global to store fetched data
+
+function log(isVerbose: boolean, message: string) {
+	if (isVerbose) {
+		console.log(message)
+	}
+}
+
+const systemPromptGeneratorLookup: Record<string, ConstructSystemPromptFn> = {
+	basicSystemPrompt: basicSystemPrompt,
+	claude4SystemPrompt: claude4SystemPrompt,
+}
+
+type TestResultSet = { [test_id: string]: (TestResult & { test_id?: string })[] }
+
+class NodeTestRunner {
+	private apiKey: string | undefined
+	private currentRunId: string | null = null
+	private systemPromptHash: string | null = null
+	private processingFunctionsHash: string | null = null
+	private caseIdMap: Map<string, string> = new Map() // test_id -> case_id mapping
+
+	constructor(isReplay: boolean) {
+		if (!isReplay) {
+			this.apiKey = process.env.OPENROUTER_API_KEY
+			if (!this.apiKey) {
+				throw new Error("OPENROUTER_API_KEY environment variable not set for a non-replay run.")
+			}
+		}
+	}
+
+	/**
+	 * Initialize database run and store system prompt and processing functions
+	 */
+	async initializeDatabaseRun(testConfig: TestConfig, testCases: ProcessedTestCase[], isVerbose: boolean): Promise<string> {
+		try {
+			// Generate a sample system prompt to hash (using first test case)
+			const sampleSystemPrompt = testCases.length > 0 
+				? this.constructSystemPrompt(testCases[0].system_prompt_details, testConfig.system_prompt_name)
+				: "default-system-prompt";
+
+			// Store system prompt
+			this.systemPromptHash = await upsertSystemPrompt({
+				name: testConfig.system_prompt_name,
+				content: sampleSystemPrompt
+			});
+
+			// Store processing functions
+			this.processingFunctionsHash = await upsertProcessingFunctions({
+				name: `${testConfig.parsing_function}-${testConfig.diff_edit_function}`,
+				parsing_function: testConfig.parsing_function,
+				diff_edit_function: testConfig.diff_edit_function
+			});
+
+			// Create benchmark run
+			const runDescription = `Model: ${testConfig.model_id}, Cases: ${testCases.length}, Runs per case: ${testConfig.number_of_runs}`;
+			this.currentRunId = await createBenchmarkRun({
+				description: runDescription,
+				system_prompt_hash: this.systemPromptHash
+			});
+
+			log(isVerbose, `✓ Database run initialized: ${this.currentRunId}`);
+			
+			// Create case records
+			await this.createDatabaseCases(testCases, isVerbose);
+
+			return this.currentRunId;
+		} catch (error) {
+			console.error("Failed to initialize database run:", error);
+			throw error;
+		}
+	}
+
+	/**
+	 * Initialize multi-model database run (one run for all models)
+	 */
+	async initializeMultiModelRun(testCases: ProcessedTestCase[], systemPromptName: string, parsingFunction: string, diffEditFunction: string, runDescription: string, isVerbose: boolean): Promise<string> {
+		try {
+			// Generate a sample system prompt to hash (using first test case)
+			const sampleSystemPrompt = testCases.length > 0 
+				? this.constructSystemPrompt(testCases[0].system_prompt_details, systemPromptName)
+				: "default-system-prompt";
+
+			// Store system prompt
+			this.systemPromptHash = await upsertSystemPrompt({
+				name: systemPromptName,
+				content: sampleSystemPrompt
+			});
+
+			// Store processing functions
+			this.processingFunctionsHash = await upsertProcessingFunctions({
+				name: `${parsingFunction}-${diffEditFunction}`,
+				parsing_function: parsingFunction,
+				diff_edit_function: diffEditFunction
+			});
+
+			// Create benchmark run
+			this.currentRunId = await createBenchmarkRun({
+				description: runDescription,
+				system_prompt_hash: this.systemPromptHash
+			});
+
+			log(isVerbose, `✓ Multi-model database run initialized: ${this.currentRunId}`);
+			
+			// Create case records
+			await this.createDatabaseCases(testCases, isVerbose);
+
+			return this.currentRunId;
+		} catch (error) {
+			console.error("Failed to initialize multi-model database run:", error);
+			throw error;
+		}
+	}
+
+	/**
+	 * Create database case records for all test cases
+	 */
+	async createDatabaseCases(testCases: ProcessedTestCase[], isVerbose: boolean): Promise<void> {
+		if (!this.currentRunId || !this.systemPromptHash) {
+			throw new Error("Database run not initialized");
+		}
+
+		for (const testCase of testCases) {
+			try {
+				// Store file content if available
+				let fileHash: string | undefined;
+				if (testCase.file_contents && testCase.file_path) {
+					fileHash = await upsertFile({
+						filepath: testCase.file_path,
+						content: testCase.file_contents
+					});
+				}
+
+				// Calculate tokens in context (approximate)
+				const tokensInContext = this.estimateTokens(testCase.messages);
+
+				// Create case record
+				const caseId = await createCase({
+					run_id: this.currentRunId,
+					description: testCase.test_id,
+					system_prompt_hash: this.systemPromptHash,
+					task_id: testCase.test_id,
+					tokens_in_context: tokensInContext,
+					file_hash: fileHash
+				});
+
+				this.caseIdMap.set(testCase.test_id, caseId);
+			} catch (error) {
+				console.error(`Failed to create database case for ${testCase.test_id}:`, error);
+				// Continue with other cases
+			}
+		}
+
+		log(isVerbose, `✓ Created ${this.caseIdMap.size} database case records`);
+	}
+
+	/**
+	 * Store test result in database
+	 */
+	async storeResultInDatabase(result: TestResult, testId: string, modelId: string): Promise<void> {
+		if (!this.currentRunId || !this.processingFunctionsHash) {
+			return; // Skip if database not initialized
+		}
+
+		const caseId = this.caseIdMap.get(testId);
+		if (!caseId) {
+			return; // Skip if case not found
+		}
+
+		try {
+			// Map error string to error enum (simple mapping)
+			const errorEnum = this.mapErrorToEnum(result.error);
+
+			// Store diff edit content if available
+			let fileEditedHash: string | undefined;
+			if (result.diffEdit) {
+				fileEditedHash = await upsertFile({
+					filepath: `diff-edit-${testId}`,
+					content: result.diffEdit
+				});
+			}
+
+			// Calculate basic metrics from diff edit if available
+			let numEdits = 0;
+			let numLinesAdded = 0;
+			let numLinesDeleted = 0;
+			
+			if (result.diffEdit) {
+				// Simple parsing to count edits - count SEARCH/REPLACE blocks
+				const searchBlocks = (result.diffEdit.match(/------- SEARCH/g) || []).length;
+				numEdits = searchBlocks;
+				
+				// Count added/deleted lines (rough approximation)
+				const lines = result.diffEdit.split('\n');
+				for (const line of lines) {
+					if (line.startsWith('+') && !line.startsWith('+++')) {
+						numLinesAdded++;
+					} else if (line.startsWith('-') && !line.startsWith('---')) {
+						numLinesDeleted++;
+					}
+				}
+			}
+
+			const resultInput: CreateResultInput = {
+				run_id: this.currentRunId,
+				case_id: caseId,
+				model_id: modelId,
+				processing_functions_hash: this.processingFunctionsHash,
+				succeeded: result.success && (result.diffEditSuccess ?? false),
+				error_enum: errorEnum,
+				num_edits: numEdits || undefined,
+				num_lines_deleted: numLinesDeleted || undefined,
+				num_lines_added: numLinesAdded || undefined,
+				time_to_first_token_ms: result.streamResult?.timing?.timeToFirstTokenMs,
+				time_to_first_edit_ms: result.streamResult?.timing?.timeToFirstEditMs,
+				time_round_trip_ms: result.streamResult?.timing?.totalRoundTripMs,
+				cost_usd: result.streamResult?.usage?.totalCost,
+				completion_tokens: result.streamResult?.usage?.outputTokens,
+				raw_model_output: result.streamResult?.assistantMessage,
+				file_edited_hash: fileEditedHash,
+				parsed_tool_call_json: result.toolCalls ? JSON.stringify(result.toolCalls) : undefined
+			};
+
+			await insertResult(resultInput);
+		} catch (error) {
+			console.error(`Failed to store result in database for ${testId}:`, error);
+			// Continue execution - don't fail the test run
+		}
+	}
+
+	/**
+	 * Estimate token count for messages (rough approximation)
+	 */
+	public estimateTokens(messages: Anthropic.Messages.MessageParam[]): number { // Made public
+		let totalText = "";
+		for (const message of messages) {
+			if (Array.isArray(message.content)) {
+				for (const block of message.content) {
+					if (block.type === 'text') {
+						totalText += block.text + "\n";
+					}
+				}
+			} else if (typeof message.content === 'string') {
+				totalText += message.content + "\n";
+			}
+		}
+		return encoding.encode(totalText).length;
+	}
+
+	/**
+	 * Map error string to error enum
+	 */
+	private mapErrorToEnum(error?: string): number | undefined {
+		if (!error) return undefined;
+		
+		const errorMap: Record<string, number> = {
+			'no_tool_calls': 1,
+			'parsing_error': 2,
+			'diff_edit_error': 3,
+			'missing_original_diff_edit_tool_call_message': 4,
+			'api_error': 5,
+			'wrong_tool_call': 6,
+			'wrong_file_edited': 7,
+			'multi_tool_calls': 8,
+			'tool_call_params_undefined': 9,
+			'other_error': 99
+		};
+
+		return errorMap[error] || 99; // 99 for unknown errors
+	}
+
+	/**
+	 * convert our messages array into a properly formatted Anthropic messages array
+	 */
+	transformMessages(messages: InputMessage[]): Anthropic.Messages.MessageParam[] {
+		return messages.map((msg) => {
+			// Use TextBlockParam here for constructing the input message
+			const content: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[] = []
+
+			if (msg.text) {
+				// This object now correctly matches the TextBlockParam type
+				content.push({ type: "text", text: msg.text })
+			}
+
+			if (msg.images && Array.isArray(msg.images)) {
+				const imageBlocks = formatResponse.imageBlocks(msg.images)
+				content.push(...imageBlocks)
+			}
+
+			return {
+				role: msg.role,
+				content: content,
+			}
+		})
+	}
+
+	/**
+	 * Generate the system prompt on the fly
+	 */
+	constructSystemPrompt(systemPromptDetails: SystemPromptDetails, systemPromptName: string) {
+		const systemPromptGenerator = systemPromptGeneratorLookup[systemPromptName]
+
+		const { cwd_value, browser_use, width, height, os_value, shell_value, home_value, mcp_string, user_custom_instructions } =
+			systemPromptDetails
+
+		const systemPrompt = systemPromptGenerator(
+			cwd_value,
+			browser_use,
+			width,
+			height,
+			os_value,
+			shell_value,
+			home_value,
+			mcp_string,
+			user_custom_instructions,
+		)
+
+		return systemPrompt
+	}
+
+	/**
+	 * Loads our test cases from a directory of json files
+	 */
+	loadTestCases(testDirectoryPath: string, isVerbose: boolean): TestCase[] {
+		const testCasesArray: TestCase[] = []
+		const dirents = fs.readdirSync(testDirectoryPath, { withFileTypes: true })
+
+		for (const dirent of dirents) {
+			if (dirent.isFile() && dirent.name.endsWith(".json")) {
+				const testFilePath = path.join(testDirectoryPath, dirent.name)
+				const fileContent = fs.readFileSync(testFilePath, "utf8")
+				const testCase: TestCase = JSON.parse(fileContent)
+
+				// Use the filename (without extension) as the test_id if not provided
+				if (!testCase.test_id) {
+					testCase.test_id = path.parse(dirent.name).name
+				}
+
+				// Filter out cases with missing file_contents
+				if (!testCase.file_contents || testCase.file_contents.trim() === "") {
+					log(isVerbose, `Skipping case ${testCase.test_id}: missing or empty file_contents.`);
+					continue;
+				}
+				testCasesArray.push(testCase)
+			}
+		}
+		return testCasesArray
+	}
+
+	/**
+	 * Saves the test results to the specified output directory.
+	 */
+	saveTestResults(results: TestResultSet, outputPath: string) {
+		// Ensure output directory exists
+		if (!fs.existsSync(outputPath)) {
+			fs.mkdirSync(outputPath, { recursive: true })
+		}
+
+		// Write each test result to its own file
+		for (const testId in results) {
+			const outputFilePath = path.join(outputPath, `${testId}.json`)
+			const testResult = results[testId]
+			fs.writeFileSync(outputFilePath, JSON.stringify(testResult, null, 2))
+		}
+	}
+
+	/**
+	 * Run a single test example
+	 */
+	async runSingleTest(testCase: ProcessedTestCase, testConfig: TestConfig, isVerbose: boolean = false): Promise<TestResult> {
+		if (testConfig.replay && !testCase.original_diff_edit_tool_call_message) {
+			return {
+				success: false,
+				error: "missing_original_diff_edit_tool_call_message",
+				errorString: `Test case ${testCase.test_id} is missing 'original_diff_edit_tool_call_message' for replay.`,
+			}
+		}
+
+		const customSystemPrompt = this.constructSystemPrompt(testCase.system_prompt_details, testConfig.system_prompt_name)
+
+		// messages don't include system prompt and are everything up to the first replace_in_file tool call which results in a diff edit error
+		const input: TestInput = {
+			apiKey: this.apiKey,
+			systemPrompt: customSystemPrompt,
+			messages: testCase.messages,
+			modelId: testConfig.model_id,
+			originalFile: testCase.file_contents,
+			originalFilePath: testCase.file_path,
+			parsingFunction: testConfig.parsing_function,
+			diffEditFunction: testConfig.diff_edit_function,
+			thinkingBudgetTokens: testConfig.thinking_tokens_budget,
+			originalDiffEditToolCallMessage: testConfig.replay ? testCase.original_diff_edit_tool_call_message : undefined,
+		}
+
+		if (isVerbose) {
+			log(isVerbose, `    Sending request to ${testConfig.model_id} for test case ${testCase.test_id}...`);
+		}
+		
+		return await runSingleEvaluation(input)
+	}
+
+	/**
+	 * Runs all the text examples synchonously
+	 */
+	async runAllTests(testCases: ProcessedTestCase[], testConfig: TestConfig, isVerbose: boolean): Promise<TestResultSet> {
+		const results: TestResultSet = {}
+
+		// Initialize database run
+		try {
+			await this.initializeDatabaseRun(testConfig, testCases, isVerbose);
+		} catch (error) {
+			log(isVerbose, `Warning: Failed to initialize database: ${error}`);
+		}
+
+		for (const testCase of testCases) {
+			results[testCase.test_id] = []
+
+			log(isVerbose, `-Running test: ${testCase.test_id}`)
+			for (let i = 0; i < testConfig.number_of_runs; i++) {
+				log(isVerbose, `  Attempt ${i+1}/${testConfig.number_of_runs} for ${testCase.test_id}...`);
+				const result = await this.runSingleTest(testCase, testConfig, isVerbose)
+				results[testCase.test_id].push(result)
+				
+				// Log result status
+				if (isVerbose) {
+					if (result.success) {
+						log(isVerbose, `  ✓ Attempt ${i+1} completed successfully`);
+					} else {
+						log(isVerbose, `  ✗ Attempt ${i+1} failed (error: ${result.error || 'unknown'})`);
+					}
+				}
+				
+				// Store result in database
+				try {
+					await this.storeResultInDatabase(result, testCase.test_id, testConfig.model_id);
+				} catch (error) {
+					log(isVerbose, `Warning: Failed to store result in database: ${error}`);
+				}
+			}
+		}
+		return results
+	}
+
+	/**
+	 * Runs all of the text examples asynchronously, with concurrency limit
+	 */
+	async runAllTestsParallel(
+		testCases: ProcessedTestCase[],
+		testConfig: TestConfig,
+		isVerbose: boolean,
+		maxConcurrency: number = 20,
+	): Promise<TestResultSet> {
+		const results: TestResultSet = {}
+		testCases.forEach((tc) => {
+			results[tc.test_id] = []
+		})
+
+		// Initialize database run
+		try {
+			await this.initializeDatabaseRun(testConfig, testCases, isVerbose);
+		} catch (error) {
+			log(isVerbose, `Warning: Failed to initialize database: ${error}`);
+		}
+
+		// Create a flat list of all individual runs we need to execute
+		const allRuns = testCases.flatMap((testCase) =>
+			Array(testConfig.number_of_runs)
+				.fill(null)
+				.map(() => testCase),
+		)
+
+		for (let i = 0; i < allRuns.length; i += maxConcurrency) {
+			const batch = allRuns.slice(i, i + maxConcurrency)
+
+			const batchPromises = batch.map((testCase) => {
+				log(isVerbose, `  Running test for ${testCase.test_id}...`);
+				return this.runSingleTest(testCase, testConfig, isVerbose).then((result) => ({
+					...result,
+					test_id: testCase.test_id,
+				}))
+			})
+
+			const batchResults = await Promise.all(batchPromises)
+
+			// Calculate the total cost for this batch
+			const batchCost = batchResults.reduce((total, result) => {
+				return total + (result.streamResult?.usage?.totalCost || 0)
+			}, 0)
+
+			// Populate the results dictionary and store in database
+			for (const result of batchResults) {
+				if (result.test_id) {
+					results[result.test_id].push(result)
+					
+					// Store result in database
+					try {
+						await this.storeResultInDatabase(result, result.test_id, testConfig.model_id);
+					} catch (error) {
+						log(isVerbose, `Warning: Failed to store result in database: ${error}`);
+					}
+				}
+			}
+
+			const batchNumber = i / maxConcurrency + 1
+			const totalBatches = Math.ceil(allRuns.length / maxConcurrency)
+			log(isVerbose, `-Completed batch ${batchNumber} of ${totalBatches}... (Batch Cost: $${batchCost.toFixed(6)})`)
+		}
+
+		return results
+	}
+
+	/**
+	 * Check if a test result is a valid attempt (no error_enum 1, 6, or 7)
+	 */
+	isValidAttempt(result: TestResult): boolean {
+		// Invalid if error is one of: no_tool_calls, wrong_tool_call, wrong_file_edited
+		const invalidErrors = ['no_tool_calls', 'wrong_tool_call', 'wrong_file_edited'];
+		return !invalidErrors.includes(result.error || '');
+	}
+
+	/**
+	 * Runs all tests for a specific model (assumes database run already initialized)
+	 * Keeps retrying until we get the requested number of valid attempts per case
+	 */
+	async runAllTestsForModel(testCases: ProcessedTestCase[], testConfig: TestConfig, isVerbose: boolean): Promise<TestResultSet> {
+		const results: TestResultSet = {}
+
+		for (const testCase of testCases) {
+			results[testCase.test_id] = []
+			let validAttempts = 0;
+			let totalAttempts = 0;
+
+			log(isVerbose, `-Running test: ${testCase.test_id}`)
+			
+			// Keep trying until we get the requested number of valid attempts
+			while (validAttempts < testConfig.number_of_runs) {
+				totalAttempts++;
+				log(isVerbose, `  Attempt ${totalAttempts} for ${testCase.test_id} (${validAttempts}/${testConfig.number_of_runs} valid so far)...`);
+				
+				const result = await this.runSingleTest(testCase, testConfig, isVerbose)
+				results[testCase.test_id].push(result)
+				
+				// Check if this was a valid attempt
+				const isValid = this.isValidAttempt(result);
+				if (isValid) {
+					validAttempts++;
+					log(isVerbose, `  ✓ Valid attempt ${validAttempts}/${testConfig.number_of_runs} completed (${result.success ? 'SUCCESS' : 'FAILED'})`);
+				} else {
+					log(isVerbose, `  ✗ Invalid attempt (error: ${result.error || 'unknown'})`);
+				}
+				
+				// Store result in database
+				try {
+					await this.storeResultInDatabase(result, testCase.test_id, testConfig.model_id);
+				} catch (error) {
+					log(isVerbose, `Warning: Failed to store result in database: ${error}`);
+				}
+				
+				// Safety check to prevent infinite loops - limit to 10 attempts per valid attempt requested
+				if (totalAttempts >= testConfig.number_of_runs * 10) {
+					log(isVerbose, `  ⚠️ Reached maximum attempts (${totalAttempts}) for test case ${testCase.test_id}. Only got ${validAttempts}/${testConfig.number_of_runs} valid attempts.`);
+					break;
+				}
+			}
+			
+			log(isVerbose, `  ✓ Completed test case ${testCase.test_id}: ${validAttempts}/${testConfig.number_of_runs} valid attempts (${totalAttempts} total attempts)`);
+		}
+		return results
+	}
+
+	/**
+	 * Print output of the tests
+	 */
+	printSummary(results: TestResultSet, isVerbose: boolean) {
+		let totalRuns = 0
+		let totalPasses = 0
+		let totalInputTokens = 0
+		let totalOutputTokens = 0
+		let totalCost = 0
+		let runsWithUsageData = 0
+		let totalDiffEditSuccesses = 0
+		let totalRunsWithToolCalls = 0
+		const testCaseIds = Object.keys(results)
+
+		log(isVerbose, "\n=== TEST SUMMARY ===")
+
+		for (const testId of testCaseIds) {
+			const testResults = results[testId]
+			const passedCount = testResults.filter((r) => r.success && r.diffEditSuccess).length
+			const runCount = testResults.length
+
+			totalRuns += runCount
+			totalPasses += passedCount
+
+			const runsWithToolCalls = testResults.filter((r) => r.success === true).length
+			const diffEditSuccesses = passedCount
+			totalRunsWithToolCalls += runsWithToolCalls
+			totalDiffEditSuccesses += diffEditSuccesses
+
+			// Accumulate token and cost data
+			for (const result of testResults) {
+				if (result.streamResult?.usage) {
+					totalInputTokens += result.streamResult.usage.inputTokens
+					totalOutputTokens += result.streamResult.usage.outputTokens
+					totalCost += result.streamResult.usage.totalCost
+					runsWithUsageData++
+				}
+			}
+
+			log(isVerbose, `\n--- Test Case: ${testId} ---`)
+			log(isVerbose, `  Runs: ${runCount}`)
+			log(isVerbose, `  Passed: ${passedCount}`)
+			log(isVerbose, `  Success Rate: ${runCount > 0 ? ((passedCount / runCount) * 100).toFixed(1) : "N/A"}%`)
+		}
+
+		log(isVerbose, "\n\n=== OVERALL SUMMARY ===")
+		log(isVerbose, `Total Test Cases: ${testCaseIds.length}`)
+		log(isVerbose, `Total Runs Executed: ${totalRuns}`)
+		log(isVerbose, `Overall Passed: ${totalPasses}`)
+		log(isVerbose, `Overall Failed: ${totalRuns - totalPasses}`)
+		log(isVerbose, `Overall Success Rate: ${totalRuns > 0 ? ((totalPasses / totalRuns) * 100).toFixed(1) : "N/A"}%`)
+
+		log(isVerbose, "\n\n=== OVERALL DIFF EDIT SUCCESS RATE ===")
+		if (totalRunsWithToolCalls > 0) {
+			const diffSuccessRate = (totalDiffEditSuccesses / totalRunsWithToolCalls) * 100
+			log(isVerbose, `Total Runs with Successful Tool Calls: ${totalRunsWithToolCalls}`)
+			log(isVerbose, `Total Runs with Successful Diff Edits: ${totalDiffEditSuccesses}`)
+			log(isVerbose, `Diff Edit Success Rate: ${diffSuccessRate.toFixed(1)}%`)
+		} else {
+			log(isVerbose, "No successful tool calls to analyze for diff edit success.")
+		}
+
+		log(isVerbose, "\n\n=== TOKEN & COST ANALYSIS ===")
+		if (runsWithUsageData > 0) {
+			log(isVerbose, `Total Input Tokens: ${totalInputTokens.toLocaleString()}`)
+			log(isVerbose, `Total Output Tokens: ${totalOutputTokens.toLocaleString()}`)
+			log(isVerbose, `Total Cost: $${totalCost.toFixed(6)}`)
+			log(isVerbose, "---")
+			log(
+				isVerbose,
+				`Avg Input Tokens / Run: ${(totalInputTokens / runsWithUsageData).toLocaleString(undefined, {
+					maximumFractionDigits: 0,
+				})}`,
+			)
+			log(
+				isVerbose,
+				`Avg Output Tokens / Run: ${(totalOutputTokens / runsWithUsageData).toLocaleString(undefined, {
+					maximumFractionDigits: 0,
+				})}`,
+			)
+			log(isVerbose, `Avg Cost / Run: $${(totalCost / runsWithUsageData).toFixed(6)}`)
+		} else {
+			log(isVerbose, "No usage data available to analyze.")
+		}
+	}
+}
+
+async function main() {
+	interface EvaluationTask {
+		modelId: string;
+		testCase: ProcessedTestCase;
+		testConfig: TestConfig;
+	}
+
+	const program = new Command()
+
+	const defaultTestPath = path.join(__dirname, "cases")
+	const defaultOutputPath = path.join(__dirname, "results")
+
+	program
+		.name("TestRunner")
+		.description("Run evaluation tests for diff editing")
+		.version("1.0.0")
+		.option("--test-path <path>", "Path to the directory containing test case JSON files", defaultTestPath)
+		.option("--output-path <path>", "Path to the directory to save the test output JSON files", defaultOutputPath)
+		.option("--model-ids <model_ids>", "Comma-separated list of model IDs to test")
+		.option("--system-prompt-name <name>", "The name of the system prompt to use", "basicSystemPrompt")
+		.option("-n, --valid-attempts-per-case <number>", "Number of valid attempts per test case per model (will retry until this many valid attempts are collected)", "1")
+		.option("--max-cases <number>", "Maximum number of test cases to run (limits total cases loaded)")
+		.option("--parsing-function <name>", "The parsing function to use", "parseAssistantMessageV2")
+		.option("--diff-edit-function <name>", "The diff editing function to use", "constructNewFileContentV2")
+		.option("--thinking-budget <tokens>", "Set the thinking tokens budget", "0")
+		.option("--parallel", "Run tests in parallel", false)
+		.option("--replay", "Run evaluation from a pre-recorded LLM output, skipping the API call", false)
+		.option("-v, --verbose", "Enable verbose logging", false)
+		.option("--max-concurrency <number>", "Maximum number of parallel requests", "80")
+
+
+	program.parse(process.argv)
+
+	const options = program.opts()
+	const isVerbose = options.verbose
+	const testPath = options.testPath
+	const outputPath = options.outputPath
+	const maxConcurrency = parseInt(options.maxConcurrency, 10);
+
+	// Parse model IDs from comma-separated string
+	const modelIds = options.modelIds ? options.modelIds.split(',').map(id => id.trim()) : [];
+	if (modelIds.length === 0) {
+		console.error("Error: --model-ids is required and must contain at least one model ID");
+		process.exit(1);
+	}
+
+	const validAttemptsPerCase = parseInt(options.validAttemptsPerCase, 10);
+	
+	try {
+		const startTime = Date.now()
+
+		// Load OpenRouter model data first
+		openRouterModelDataGlobal = await loadOpenRouterModelData(isVerbose);
+		if (Object.keys(openRouterModelDataGlobal).length === 0 && isVerbose) {
+			log(isVerbose, "Warning: Could not load OpenRouter model data. Context window filtering might be affected for OpenRouter models.");
+		}
+
+		const runner = new NodeTestRunner(options.replay)
+		let allLoadedTestCases = runner.loadTestCases(testPath, isVerbose) // Pass isVerbose
+		
+		const allProcessedTestCasesGlobal: ProcessedTestCase[] = allLoadedTestCases.map((tc) => ({
+			...tc,
+			messages: runner.transformMessages(tc.messages),
+		}));
+
+		log(isVerbose, `-Loaded ${allLoadedTestCases.length} initial test cases.`)
+		log(isVerbose, `-Testing ${modelIds.length} model(s): ${modelIds.join(', ')}`)
+		log(isVerbose, `-Target: ${validAttemptsPerCase} valid attempts per test case per model (will retry until this many valid attempts are collected)`)
+		if (options.replay) {
+			log(isVerbose, `-Running in REPLAY mode. No API calls will be made.`)
+		}
+		log(isVerbose, "Starting tests...\n")
+
+		// Determine the smallest context window among all specified models
+		let smallestContextWindow = Infinity;
+		for (const modelId of modelIds) {
+			let modelInfo = openRouterModelDataGlobal[modelId];
+			if (!modelInfo) {
+				const foundKey = Object.keys(openRouterModelDataGlobal).find(
+					key => key.includes(modelId) || modelId.includes(key)
+				);
+				if (foundKey) modelInfo = openRouterModelDataGlobal[foundKey];
+			}
+			const currentModelContext = modelInfo?.contextWindow;
+			if (currentModelContext && currentModelContext > 0) {
+				if (currentModelContext < smallestContextWindow) {
+					smallestContextWindow = currentModelContext;
+				}
+			} else {
+				log(isVerbose, `Warning: Context window for model ${modelId} is unknown or zero. It will not constrain the test case selection.`);
+			}
+		}
+
+		if (smallestContextWindow === Infinity) {
+			log(isVerbose, "Warning: Could not determine a common smallest context window. Proceeding with all loaded cases, context issues may occur.");
+		} else {
+			log(isVerbose, `Smallest common context window (with padding consideration) across specified models: ${smallestContextWindow} (target for filtering: ${smallestContextWindow - 20000})`);
+		}
+		
+		let eligibleCasesForThisRun = [...allLoadedTestCases];
+		if (smallestContextWindow !== Infinity && smallestContextWindow > 20000) { // Only filter if a valid smallest window is found
+			const originalCaseCount = eligibleCasesForThisRun.length;
+			eligibleCasesForThisRun = eligibleCasesForThisRun.filter(tc => {
+				const systemPromptText = runner.constructSystemPrompt(tc.system_prompt_details, options.systemPromptName);
+				const systemPromptTokens = encoding.encode(systemPromptText).length;
+				const messagesTokens = runner.estimateTokens(runner.transformMessages(tc.messages));
+				const totalInputTokens = systemPromptTokens + messagesTokens;
+				return totalInputTokens + 20000 <= smallestContextWindow; // 20k padding
+			});
+			log(isVerbose, `Filtered to ${eligibleCasesForThisRun.length} cases (from ${originalCaseCount}) to fit smallest context window of ${smallestContextWindow} (with padding).`);
+		}
+
+		// Apply max-cases limit if specified, to the context-filtered list
+		if (options.maxCases && options.maxCases > 0 && eligibleCasesForThisRun.length > options.maxCases) {
+			log(isVerbose, `Limiting to ${options.maxCases} test cases (out of ${eligibleCasesForThisRun.length} eligible).`);
+			eligibleCasesForThisRun = eligibleCasesForThisRun.slice(0, options.maxCases);
+		}
+
+		if (eligibleCasesForThisRun.length === 0) {
+			log(isVerbose, `No eligible test cases found after filtering for all specified models. Exiting.`);
+			process.exit(0);
+		}
+		
+		const processedEligibleCasesForRun: ProcessedTestCase[] = eligibleCasesForThisRun.map((tc) => ({
+			...tc,
+			messages: runner.transformMessages(tc.messages),
+		}));
+
+		// Initialize ONE database run for ALL models using the commonly eligible cases
+		const runDescription = `Models: ${modelIds.join(', ')}, Common Cases: ${processedEligibleCasesForRun.length}, Valid attempts per case: ${validAttemptsPerCase}`;
+		await runner.initializeMultiModelRun(processedEligibleCasesForRun, options.systemPromptName, options.parsingFunction, options.diffEditFunction, runDescription, isVerbose);
+
+		// Create a global task queue
+		const globalTaskQueue: EvaluationTask[] = modelIds.flatMap(modelId => 
+			processedEligibleCasesForRun.map(testCase => ({
+				modelId,
+				testCase,
+				testConfig: {
+					model_id: modelId,
+					system_prompt_name: options.systemPromptName,
+					number_of_runs: validAttemptsPerCase,
+					parsing_function: options.parsingFunction,
+					diff_edit_function: options.diffEditFunction,
+					thinking_tokens_budget: parseInt(options.thinkingBudget, 10),
+					replay: options.replay,
+				}
+			}))
+		);
+
+		const results: TestResultSet = {};
+		const taskStates: Record<string, { valid: number; total: number; pending: number }> = {};
+
+		globalTaskQueue.forEach(({ modelId, testCase }) => {
+			const taskId = `${modelId}-${testCase.test_id}`;
+			taskStates[taskId] = { valid: 0, total: 0, pending: 0 };
+			if (!results[testCase.test_id]) {
+				results[testCase.test_id] = [];
+			}
+		});
+
+		let remainingTasks = [...globalTaskQueue];
+
+		while (remainingTasks.length > 0) {
+			const batch: EvaluationTask[] = [];
+			for (const task of remainingTasks) {
+				if (batch.length >= maxConcurrency) break;
+				const taskId = `${task.modelId}-${task.testCase.test_id}`;
+				if ((taskStates[taskId].valid + taskStates[taskId].pending) < validAttemptsPerCase) {
+					batch.push(task);
+					taskStates[taskId].pending++;
+				}
+			}
+
+			if (batch.length === 0) {
+				await new Promise(resolve => setTimeout(resolve, 100));
+				continue;
+			}
+
+			const batchPromises = batch.map(task => {
+				const taskId = `${task.modelId}-${task.testCase.test_id}`;
+				taskStates[taskId].total++;
+				log(isVerbose, `  Attempt ${taskStates[taskId].total} for ${task.testCase.test_id} with ${task.modelId} (${taskStates[taskId].valid} valid, ${taskStates[taskId].pending - 1} pending)...`);
+				return runner.runSingleTest(task.testCase, task.testConfig, isVerbose).then(result => ({
+					...result,
+					test_id: task.testCase.test_id,
+					modelId: task.modelId,
+				}));
+			});
+
+			const batchResults = await Promise.all(batchPromises);
+
+			for (const result of batchResults) {
+				const taskId = `${result.modelId}-${result.test_id}`;
+				taskStates[taskId].pending--;
+				results[result.test_id].push(result);
+
+				if (runner.isValidAttempt(result)) {
+					taskStates[taskId].valid++;
+					log(isVerbose, `  ✓ Valid attempt ${taskStates[taskId].valid}/${validAttemptsPerCase} for ${result.test_id} with ${result.modelId} completed (${result.success ? 'SUCCESS' : 'FAILED'})`);
+				} else {
+					log(isVerbose, `  ✗ Invalid attempt for ${result.test_id} with ${result.modelId} (error: ${result.error || 'unknown'})`);
+				}
+
+				await runner.storeResultInDatabase(result, result.test_id, result.modelId);
+			}
+
+			remainingTasks = remainingTasks.filter(task => {
+				const taskId = `${task.modelId}-${task.testCase.test_id}`;
+				if (taskStates[taskId].total >= validAttemptsPerCase * 10) {
+					log(isVerbose, `  ⚠️ Reached maximum attempts for ${task.testCase.test_id} with ${task.modelId}.`);
+					return false;
+				}
+				return taskStates[taskId].valid < validAttemptsPerCase;
+			});
+
+			const batchCost = batchResults.reduce((total, result) => total + (result.streamResult?.usage?.totalCost || 0), 0);
+			log(isVerbose, `-Completed batch... (Batch Cost: $${batchCost.toFixed(6)}, Remaining tasks: ${remainingTasks.length})`);
+		}
+
+		// Print summary for each model
+		for (const modelId of modelIds) {
+			const modelResults: TestResultSet = {};
+			Object.keys(results).forEach(testId => {
+				modelResults[testId] = results[testId].filter(r => (r as any).modelId === modelId);
+			});
+			log(isVerbose, `\n=== Results for Model: ${modelId} ===`);
+			runner.printSummary(modelResults, isVerbose);
+		}
+
+		const endTime = Date.now()
+		const durationSeconds = ((endTime - startTime) / 1000).toFixed(2)
+		log(isVerbose, `\n-Total execution time: ${durationSeconds} seconds`)
+
+		log(isVerbose, `\n✓ All results stored in database. Use the dashboard to view results.`)
+	} catch (error) {
+		console.error("\nError running tests:", error)
+		process.exit(1)
+	}
+}
+
+if (require.main === module) {
+	main()
+}

evals/diff-edits/dashboard/.streamlit/config.toml

@@ -0,0 +1,8 @@
+[theme]
+base="dark"
+
+[browser]
+gatherUsageStats = false
+
+[server]
+headless = true

evals/diff-edits/dashboard/README.md

@@ -0,0 +1,159 @@
+# 🚀 The Sickest Diff Edits Evaluation Dashboard Ever!
+
+A beautiful, modern Streamlit dashboard for visualizing and analyzing diff editing evaluation results with deep drill-down capabilities.
+
+## ✨ Features
+
+### 🎯 **Smart Model Comparison**
+- **Latest Run Focus**: Automatically loads and displays your most recent evaluation run
+- **Beautiful Performance Cards**: Each model gets a stunning card with performance grades (A+ to C)
+- **Best Performer Highlighting**: The top model gets special styling and a trophy 🏆
+- **Interactive Charts**: Success rate comparisons and latency vs cost analysis
+
+### 🔍 **Deep Drill-Down Analysis**
+- **Individual Result Inspection**: Click any model to see detailed results
+- **Side-by-Side File Views**: See original file content with line numbers
+- **Parsed Tool Call Analysis**: View exactly what the model tried to do
+- **Error Analysis**: Detailed error information for failed attempts
+- **Success Metrics**: Line changes, edit counts, and timing breakdowns
+
+### 🎨 **Aesthetic Design**
+- **Modern UI**: Custom CSS with Inter font, gradients, and shadows
+- **Responsive Layout**: Looks great on any screen size
+- **Color-Coded Performance**: Green for excellent, yellow for good, red for poor
+- **Smooth Animations**: Hover effects and transitions
+- **Professional Styling**: Clean, modern design that looks amazing
+
+### 📊 **Comprehensive Metrics**
+- **Success Rates**: Color-coded percentages with performance grades
+- **Timing Analysis**: First token, first edit, and round trip times
+- **Cost Tracking**: Per-result and total cost analysis
+- **Token Metrics**: Context tokens and completion tokens
+- **Edit Statistics**: Number of edits, lines added/deleted
+
+## 🚀 Quick Start
+
+1. **Install dependencies**:
+   ```bash
+   cd diff-edits/dashboard
+   pip install -r requirements.txt
+   ```
+
+2. **Launch the dashboard**:
+   ```bash
+   streamlit run app.py
+   ```
+   
+   Or use the convenient launch script:
+   ```bash
+   ./launch.sh
+   ```
+
+3. **Open your browser** to http://localhost:8501
+
+## 🎯 Dashboard Sections
+
+### **Hero Section**
+- Beautiful gradient header with run information
+- Key metrics overview (models tested, total results, success rate, cost)
+
+### **Model Performance Cards**
+- Each model displayed as a beautiful card
+- Large success rate display with color coding
+- Performance grade badges (A+, A, B+, B, C+, C)
+- Key metrics: latency, cost, results count, first token time
+- "Drill Down" button for detailed analysis
+
+### **Performance Analytics**
+- Interactive bar chart showing success rates
+- Scatter plot of latency vs cost with bubble sizes
+- Hover details and zoom capabilities
+
+### **Detailed Analysis (Drill-Down)**
+- Model-specific success rate, latency, and cost metrics
+- Individual result selector with status icons
+- Tabbed interface for different views:
+
+#### 📄 **File & Edits Tab**
+- **Side-by-side view**: Original file content with line numbers
+- **Edit analysis**: Success/failure status with detailed metrics
+- **Error display**: Clear error information for failed attempts
+- **Success metrics**: Lines added/deleted, number of edits
+- **Parsed tool calls**: JSON view of what the model attempted
+
+#### 🤖 **Raw Output Tab**
+- Complete raw model output in a code viewer
+- Monospace font for easy reading
+
+#### 🔧 **Parsed Tool Call Tab**
+- Pretty-printed JSON of parsed tool calls
+- Diff block visualization for replace_in_file calls
+- Error handling for malformed JSON
+
+#### 📊 **Metrics Tab**
+- Detailed timing metrics (first token, first edit, round trip)
+- Token and cost information
+- Context size and completion tokens
+
+## 🛠 **Technical Features**
+
+### **Smart Data Loading**
+- Automatic latest run detection
+- Efficient SQL queries with proper JOINs
+- Streamlit caching for performance
+- Error handling for missing data
+
+### **Interactive Navigation**
+- Session state management for drill-down views
+- Back button to return to overview
+- Smooth transitions between views
+
+### **Beautiful Styling**
+- Custom CSS with Google Fonts (Inter)
+- Gradient backgrounds and shadows
+- Hover effects and animations
+- Color-coded performance indicators
+- Professional card-based layout
+
+### **Responsive Design**
+- Works on desktop, tablet, and mobile
+- Flexible column layouts
+- Scalable text and metrics
+
+## 🎨 **Design Philosophy**
+
+This dashboard follows modern design principles:
+- **Clarity**: Information is easy to find and understand
+- **Beauty**: Visually appealing with professional styling
+- **Functionality**: Deep drill-down capabilities for detailed analysis
+- **Performance**: Fast loading with efficient data queries
+- **Usability**: Intuitive navigation and clear visual hierarchy
+
+## 📊 **Data Visualization**
+
+- **Plotly Charts**: Interactive, professional-looking visualizations
+- **Color Coding**: Consistent color scheme for performance levels
+- **Performance Badges**: A+ to C grading system
+- **Status Icons**: ✅ for success, ❌ for failure
+- **Metric Cards**: Clean, card-based metric display
+
+## 🔧 **Customization**
+
+The dashboard is highly customizable:
+- **CSS Styling**: Easy to modify colors, fonts, and layouts
+- **Performance Grades**: Adjustable thresholds for A/B/C grades
+- **Metrics Display**: Add or remove metrics as needed
+- **Chart Types**: Easily swap chart types or add new visualizations
+
+## 🚀 **Future Enhancements**
+
+Potential additions:
+- **Historical Trends**: Compare performance across multiple runs
+- **Export Functionality**: Download results as CSV/PDF
+- **Real-time Updates**: Auto-refresh for ongoing evaluations
+- **Custom Filters**: Filter by date range, model type, etc.
+- **Comparison Mode**: Side-by-side model comparisons
+
+---
+
+**This is the sickest eval dashboard ever!** 🔥 It combines beautiful design with powerful analysis capabilities, making it easy to understand model performance at a glance while providing deep drill-down capabilities for detailed investigation.

evals/diff-edits/dashboard/app.py

@@ -0,0 +1,948 @@
+import streamlit as st
+import sqlite3
+import pandas as pd
+import plotly.express as px
+import plotly.graph_objects as go
+from plotly.subplots import make_subplots
+import numpy as np
+from datetime import datetime
+import os
+import json
+import difflib
+# import mimetypes # No longer needed here if guess_language_from_filepath handles it
+from utils import get_database_connection, guess_language_from_filepath # Import from utils
+
+# Page config
+st.set_page_config(
+    page_title="Diff Edits Evaluation Dashboard",
+    page_icon="📊",
+    layout="wide",
+    initial_sidebar_state="expanded"
+)
+
+# Custom CSS for beautiful styling
+st.markdown("""
+<style>
+    /* Import Google Fonts */
+    @import url('https://fonts.googleapis.com/css2?family=Azeret+Mono:wght@400;700&display=swap');
+    
+    /* Global Styles */
+    .main {
+        font-family: 'Azeret Mono', monospace;
+    }
+    
+    /* Hero Section */
+    .hero-container {
+        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+        padding: 2rem;
+        border-radius: 15px;
+        margin-bottom: 2rem;
+        color: white;
+        text-align: center;
+    }
+    
+    .hero-title {
+        font-size: 3rem;
+        font-weight: 700;
+        margin-bottom: 0.5rem;
+        text-shadow: 2px 2px 4px rgba(0,0,0,0.3);
+    }
+    
+    .hero-subtitle {
+        font-size: 1.2rem;
+        font-weight: 300;
+        opacity: 0.9;
+    }
+    
+    /* Model Performance Cards */
+    .model-card {
+        background: white;
+        border-radius: 15px;
+        padding: 1.5rem;
+        margin: 1rem 0;
+        box-shadow: 0 8px 32px rgba(0,0,0,0.1);
+        border: 1px solid rgba(255,255,255,0.2);
+        transition: transform 0.3s ease, box-shadow 0.3s ease;
+    }
+    
+    .model-card:hover {
+        transform: translateY(-5px);
+        box-shadow: 0 12px 40px rgba(0,0,0,0.15);
+    }
+    
+    .model-card.best-performer {
+        border: 2px solid #00D4AA;
+        background: linear-gradient(135deg, #f0fdf4 0%, #ecfdf5 100%);
+    }
+    
+    .model-name {
+        font-size: 1.5rem;
+        font-weight: 600;
+        margin-bottom: 1rem;
+        color: #1f2937;
+    }
+    
+    .success-rate {
+        font-size: 3rem;
+        font-weight: 700;
+        margin-bottom: 0.5rem;
+    }
+    
+    .success-rate.excellent { color: #10b981; }
+    .success-rate.good { color: #f59e0b; }
+    .success-rate.poor { color: #ef4444; }
+    
+    .metric-row {
+        display: flex;
+        justify-content: space-between;
+        margin: 0.5rem 0;
+        padding: 0.5rem;
+        background: rgba(0,0,0,0.02);
+        border-radius: 8px;
+    }
+    
+    .metric-label {
+        font-weight: 500;
+        color: #6b7280;
+    }
+    
+    .metric-value {
+        font-weight: 600;
+        color: #1f2937;
+    }
+    
+    /* Performance Badge */
+    .performance-badge {
+        display: inline-block;
+        padding: 0.25rem 0.75rem;
+        border-radius: 20px;
+        font-weight: 600;
+        font-size: 0.875rem;
+        margin-left: 1rem;
+    }
+    
+    .badge-a { background: #10b981; color: white; }
+    .badge-b { background: #f59e0b; color: white; }
+    .badge-c { background: #ef4444; color: white; }
+    
+    /* Comparison Charts */
+    .chart-container {
+        background: white;
+        border-radius: 15px;
+        padding: 1.5rem;
+        margin: 1rem 0;
+        box-shadow: 0 4px 20px rgba(0,0,0,0.08);
+    }
+    
+    /* Result Detail Modal */
+    .result-detail {
+        background: white;
+        border-radius: 15px;
+        padding: 2rem;
+        margin: 1rem 0;
+        box-shadow: 0 8px 32px rgba(0,0,0,0.1);
+    }
+    
+    .file-viewer {
+        background: #f8fafc;
+        border: 1px solid #e2e8f0;
+        border-radius: 8px;
+        padding: 1rem;
+        font-family: 'Monaco', 'Menlo', 'Ubuntu Mono', monospace;
+        font-size: 0.875rem;
+        line-height: 1.5;
+        overflow-x: auto;
+    }
+    
+    .diff-added {
+        background-color: #dcfce7;
+        color: #166534;
+    }
+    
+    .diff-removed {
+        background-color: #fef2f2;
+        color: #dc2626;
+    }
+    
+    .error-display {
+        background: #fef2f2;
+        border: 1px solid #fecaca;
+        border-radius: 8px;
+        padding: 1rem;
+        color: #dc2626;
+        font-family: monospace;
+    }
+    
+    /* Sidebar Styling */
+    .sidebar .sidebar-content {
+        background: linear-gradient(180deg, #f8fafc 0%, #f1f5f9 100%);
+    }
+    
+    /* Custom Metrics */
+    .custom-metric {
+        text-align: center;
+        padding: 1rem;
+        background: white;
+        border-radius: 10px;
+        box-shadow: 0 2px 10px rgba(0,0,0,0.05);
+        margin: 0.5rem 0;
+    }
+    
+    .custom-metric-value {
+        font-size: 2rem;
+        font-weight: 700;
+        color: #1f2937;
+    }
+    
+    .custom-metric-label {
+        font-size: 0.875rem;
+        color: #6b7280;
+        font-weight: 500;
+        margin-top: 0.25rem;
+    }
+</style>
+""", unsafe_allow_html=True)
+
+# Enhanced data loading functions
+@st.cache_data
+def load_all_runs():
+    """Load all evaluation runs"""
+    conn = get_database_connection()
+    
+    query = """
+    SELECT run_id, description, created_at, system_prompt_hash
+    FROM runs 
+    ORDER BY created_at DESC
+    """
+    
+    return pd.read_sql_query(query, conn)
+
+@st.cache_data
+def load_run_comparison(run_id):
+    """Load a specific run with model comparison data"""
+    conn = get_database_connection()
+    
+    # Get the run details
+    run_query = f"""
+    SELECT run_id, description, created_at, system_prompt_hash
+    FROM runs 
+    WHERE run_id = '{run_id}'
+    """
+    run_data = pd.read_sql_query(run_query, conn)
+    
+    if run_data.empty:
+        return None, None
+    
+    # Get model performance for this run
+    model_perf_query = f"""
+    SELECT 
+        res.model_id,
+        COUNT(*) as total_results,
+        AVG(CASE WHEN res.succeeded THEN 1.0 ELSE 0.0 END) as success_rate,
+        AVG(res.cost_usd) as avg_cost,
+        SUM(res.cost_usd) as total_cost,
+        AVG(res.time_to_first_token_ms) as avg_first_token_ms,
+        AVG(res.time_to_first_edit_ms) as avg_first_edit_ms,
+        AVG(res.time_round_trip_ms) as avg_round_trip_ms,
+        AVG(res.completion_tokens) as avg_completion_tokens,
+        AVG(res.num_edits) as avg_num_edits,
+        MIN(res.time_round_trip_ms) as min_round_trip_ms,
+        MAX(res.time_round_trip_ms) as max_round_trip_ms
+    FROM results res
+    JOIN cases c ON res.case_id = c.case_id
+    WHERE c.run_id = '{run_id}'
+      AND (res.error_enum NOT IN (1, 6, 7) OR res.error_enum IS NULL)  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY res.model_id
+    ORDER BY success_rate DESC, avg_round_trip_ms ASC
+    """
+    
+    model_performance = pd.read_sql_query(model_perf_query, conn)
+    
+    return run_data.iloc[0], model_performance
+
+@st.cache_data
+def load_latest_run_comparison():
+    """Load the latest run with model comparison data"""
+    conn = get_database_connection()
+    
+    # Get the latest run
+    latest_run_query = """
+    SELECT run_id, description, created_at, system_prompt_hash
+    FROM runs 
+    ORDER BY created_at DESC 
+    LIMIT 1
+    """
+    latest_run = pd.read_sql_query(latest_run_query, conn)
+    
+    if latest_run.empty:
+        return None, None
+    
+    return load_run_comparison(latest_run.iloc[0]['run_id'])
+
+@st.cache_data
+def load_detailed_results(run_id, model_id=None, valid_only=False):
+    """Load detailed results for drill-down analysis"""
+    conn = get_database_connection()
+    
+    where_clause = f"WHERE c.run_id = '{run_id}'"
+    if model_id:
+        where_clause += f" AND res.model_id = '{model_id}'"
+    
+    # Option to filter out invalid attempts
+    if valid_only:
+        where_clause += " AND (res.error_enum NOT IN (1, 6, 7) OR res.error_enum IS NULL)"
+    
+    query = f"""
+    SELECT 
+        res.*,
+        c.task_id,
+        c.description as case_description,
+        c.tokens_in_context,
+        sp.name as system_prompt_name,
+        pf.name as processing_functions_name,
+        orig_f.filepath as original_filepath,
+        orig_f.content as original_file_content,
+        edit_f.filepath as edited_filepath,
+        edit_f.content as edited_file_content
+    FROM results res
+    JOIN cases c ON res.case_id = c.case_id
+    LEFT JOIN system_prompts sp ON c.system_prompt_hash = sp.hash
+    LEFT JOIN processing_functions pf ON res.processing_functions_hash = pf.hash
+    LEFT JOIN files orig_f ON c.file_hash = orig_f.hash
+    LEFT JOIN files edit_f ON res.file_edited_hash = edit_f.hash
+    {where_clause}
+    ORDER BY res.created_at DESC
+    """
+    
+    return pd.read_sql_query(query, conn)
+
+def get_performance_grade(success_rate):
+    """Get performance grade based on success rate"""
+    if success_rate >= 0.9:
+        return "A+", "excellent"
+    elif success_rate >= 0.8:
+        return "A", "excellent"
+    elif success_rate >= 0.7:
+        return "B+", "good"
+    elif success_rate >= 0.6:
+        return "B", "good"
+    elif success_rate >= 0.5:
+        return "C+", "good"
+    else:
+        return "C", "poor"
+
+def render_hero_section(current_run, model_performance):
+    """Render the hero section with key metrics"""
+    run_title = current_run['description'] if current_run['description'] else f"Run {current_run['run_id'][:8]}..."
+    st.markdown(f"""
+    <div class="hero-container">
+        <div class="hero-title">Diff Edit Evaluation Results</div>
+        <div class="hero-subtitle">A comprehensive analysis of model performance on code editing tasks.</div>
+        <div class="hero-subtitle" style="font-size: 0.9rem; margin-top: 10px;">
+            <strong>Current Run:</strong> {run_title} • {current_run['created_at']}
+        </div>
+    </div>
+    """, unsafe_allow_html=True)
+    
+    # Key metrics row
+    col1, col2, col3, col4 = st.columns(4)
+    
+    total_results = model_performance['total_results'].sum()
+    overall_success = model_performance['success_rate'].mean()
+    total_cost = model_performance['total_cost'].sum()
+    avg_latency = model_performance['avg_round_trip_ms'].mean()
+    
+    with col1:
+        st.markdown(f"""
+        <div class="custom-metric">
+            <div class="custom-metric-value">{len(model_performance)}</div>
+            <div class="custom-metric-label">Models Tested</div>
+        </div>
+        """, unsafe_allow_html=True)
+    
+    with col2:
+        st.markdown(f"""
+        <div class="custom-metric">
+            <div class="custom-metric-value">{total_results}</div>
+            <div class="custom-metric-label">Valid Results</div>
+        </div>
+        """, unsafe_allow_html=True)
+    
+    with col3:
+        success_color = "#10b981" if overall_success > 0.8 else "#f59e0b" if overall_success > 0.6 else "#ef4444"
+        st.markdown(f"""
+        <div class="custom-metric">
+            <div class="custom-metric-value" style="color: {success_color}">{overall_success:.1%}</div>
+            <div class="custom-metric-label">Avg Success Rate</div>
+        </div>
+        """, unsafe_allow_html=True)
+    
+    with col4:
+        st.markdown(f"""
+        <div class="custom-metric">
+            <div class="custom-metric-value">${total_cost:.3f}</div>
+            <div class="custom-metric-label">Total Cost</div>
+        </div>
+        """, unsafe_allow_html=True)
+
+def render_model_comparison_cards(model_performance):
+    """Render beautiful model comparison cards"""
+    st.markdown("## Model Leaderboard")
+    
+    # Find best performer
+    best_model = model_performance.iloc[0]['model_id']
+    
+    for idx, model in model_performance.iterrows():
+        is_best = model['model_id'] == best_model
+        grade, grade_class = get_performance_grade(model['success_rate'])
+        
+        # Create a container for each model
+        with st.container():
+            col1, col2 = st.columns([3, 1])
+            
+            with col1:
+                # Use Streamlit's native components instead of raw HTML
+                if is_best:
+                    st.success(f"**{model['model_id']}** - Best Performer")
+                else:
+                    st.info(f"**{model['model_id']}**")
+                
+                # Success rate with color coding
+                success_rate = model['success_rate']
+                if success_rate >= 0.8:
+                    st.success(f"**Success Rate:** {success_rate:.1%} ({grade})")
+                elif success_rate >= 0.6:
+                    st.warning(f"**Success Rate:** {success_rate:.1%} ({grade})")
+                else:
+                    st.error(f"**Success Rate:** {success_rate:.1%} ({grade})")
+                
+                # Metrics in columns
+                metric_col1, metric_col2, metric_col3, metric_col4 = st.columns(4)
+                
+                with metric_col1:
+                    st.metric("Avg Latency", f"{model['avg_round_trip_ms']:.0f}ms")
+                
+                with metric_col2:
+                    st.metric("Avg Cost", f"${model['avg_cost']:.4f}")
+                
+                with metric_col3:
+                    st.metric("Valid Results", f"{model['total_results']}")
+                
+                with metric_col4:
+                    st.metric("First Token", f"{model['avg_first_token_ms']:.0f}ms")
+            
+            with col2:
+                st.write("")  # Add some spacing
+                if st.button(f"Drill Down", key=f"drill_{model['model_id']}", use_container_width=True):
+                    st.session_state.drill_down_model = model['model_id']
+            
+            st.divider()  # Add a divider between models
+
+def render_comparison_charts(model_performance):
+    """Render interactive comparison charts"""
+    st.markdown("## Performance Analysis")
+    
+    col1, col2 = st.columns(2)
+
+    with col1:
+        # Time to First Edit
+        fig_first_edit = px.bar(
+            model_performance,
+            x='model_id',
+            y='avg_first_edit_ms',
+            title="Time to First Edit",
+            labels={'avg_first_edit_ms': 'Time to First Edit (ms)', 'model_id': 'Model'},
+            color='avg_first_edit_ms',
+            color_continuous_scale='bluered',
+            text='avg_first_edit_ms',
+            template='plotly_dark'
+        )
+        fig_first_edit.update_traces(texttemplate='%{text:.0f}ms', textposition='outside')
+        fig_first_edit.update_layout(
+            showlegend=False,
+            plot_bgcolor='rgba(0,0,0,0)',
+            paper_bgcolor='rgba(0,0,0,0)',
+            font=dict(family="Azeret Mono, monospace"),
+            margin=dict(t=50)
+        )
+        st.plotly_chart(fig_first_edit, use_container_width=True)
+
+    with col2:
+        # Latency vs Cost Scatter
+        fig_scatter = px.scatter(
+            model_performance,
+            x='avg_round_trip_ms',
+            y='avg_cost',
+            size='total_results',
+            color='success_rate',
+            hover_name='model_id',
+            title="Latency vs Cost Analysis",
+            labels={
+                'avg_round_trip_ms': 'Avg Round Trip (ms)',
+                'avg_cost': 'Avg Cost ($)',
+                'success_rate': 'Success Rate',
+                'total_results': 'Valid Results'
+            },
+            color_continuous_scale='RdYlGn',
+            template='plotly_dark'
+        )
+        fig_scatter.update_layout(
+            plot_bgcolor='rgba(0,0,0,0)',
+            paper_bgcolor='rgba(0,0,0,0)',
+            font=dict(family="Azeret Mono, monospace")
+        )
+        st.plotly_chart(fig_scatter, use_container_width=True)
+
+def render_detailed_analysis(run_id, model_id):
+    """Render detailed drill-down analysis"""
+    st.markdown(f"## Detailed Analysis: {model_id}")
+    
+    # Load all results (including invalid attempts)
+    detailed_results = load_detailed_results(run_id, model_id)
+    
+    # Also load only valid results for metrics
+    valid_results = load_detailed_results(run_id, model_id, valid_only=True)
+    
+    if detailed_results.empty:
+        st.warning("No detailed results found.")
+        return
+    
+    # Show total vs valid results
+    st.info(f"Showing all {len(detailed_results)} results ({len(valid_results)} valid, {len(detailed_results) - len(valid_results)} invalid)")
+    
+    # Results overview
+    col1, col2, col3 = st.columns(3)
+    
+    with col1:
+        success_count = valid_results['succeeded'].sum()
+        total_count = len(valid_results)
+        st.metric("Success Rate", f"{success_count}/{total_count} ({success_count/total_count:.1%} of valid results)")
+    
+    with col2:
+        avg_latency = detailed_results['time_round_trip_ms'].mean()
+        st.metric("Avg Latency", f"{avg_latency:.0f}ms")
+    
+    with col3:
+        total_cost = detailed_results['cost_usd'].sum()
+        st.metric("Total Cost", f"${total_cost:.4f}")
+    
+    # Interactive results table
+    st.markdown("### 📋 Individual Results")
+    
+    # Add result selector with indicators for valid/invalid attempts
+    result_options = []
+    for idx, row in detailed_results.iterrows():
+        # Check if this is a valid result
+        is_valid = (row['error_enum'] not in [1, 6, 7]) if not pd.isna(row['error_enum']) else True
+        
+        # Create status indicator
+        if is_valid:
+            status = "✅" if row['succeeded'] else "❌"
+        else:
+            status = "⚠️"  # Warning symbol for invalid results
+            
+        # Add validity indicator to the option text
+        validity_text = "" if is_valid else " [INVALID RESULT]"
+        result_options.append(f"{status} {row['task_id']} - {row['time_round_trip_ms']:.0f}ms{validity_text}")
+    
+    selected_result_idx = st.selectbox(
+        "Select a result to analyze:",
+        range(len(result_options)),
+        format_func=lambda x: result_options[x]
+    )
+    
+    if selected_result_idx is not None:
+        render_result_detail(detailed_results.iloc[selected_result_idx])
+
+def render_result_detail(result):
+    """Render detailed view of a single result"""
+    st.markdown("### 🔬 Result Deep Dive")
+    
+    # Check if this is a valid result
+    is_valid = (result['error_enum'] not in [1, 6, 7]) if not pd.isna(result['error_enum']) else True
+    
+    # Show validity warning if needed
+    if not is_valid:
+        st.warning("⚠️ **This is an invalid result** - The model didn't properly call the diff edit tool or edited the wrong file. This result is excluded from success rate calculations.")
+    
+    # Result metadata
+    col1, col2, col3, col4 = st.columns(4)
+    
+    with col1:
+        status_icon = "✅" if result['succeeded'] else "❌"
+        st.markdown(f"**Status:** {status_icon} {'Success' if result['succeeded'] else 'Failed'}")
+    
+    with col2:
+        st.markdown(f"**Task ID:** {result['task_id']}")
+    
+    with col3:
+        st.markdown(f"**Round Trip:** {result['time_round_trip_ms']:.0f}ms")
+    
+    with col4:
+        st.markdown(f"**Cost:** ${result['cost_usd']:.4f}")
+    
+    # Tabbed interface for different views
+    tab1, tab2, tab3, tab4 = st.tabs(["📄 File & Edits", "🤖 Raw Output", "🔧 Parsed Tool Call", "📊 Metrics"])
+    
+    with tab1:
+        render_file_and_edits_view(result)
+    
+    with tab2:
+        render_raw_output_view(result)
+    
+    with tab3:
+        render_parsed_tool_call_view(result)
+    
+    with tab4:
+        render_metrics_view(result)
+
+def render_file_and_edits_view(result):
+    """Render side-by-side file and edits view"""
+    st.markdown("#### 📄 File Content & Edit Analysis")
+    
+    # Check if we have original file content
+    has_original = not pd.isna(result['original_file_content']) and result['original_file_content']
+    has_edited = not pd.isna(result['edited_file_content']) and result['edited_file_content']
+    
+    if not has_original and not has_edited:
+        st.warning("No file content available for this result.")
+        return
+    
+    col1, col2 = st.columns(2)
+    
+    with col1:
+        st.markdown("**Original File:**")
+        if has_original:
+            filepath = result['original_filepath'] if not pd.isna(result['original_filepath']) else 'Unknown file'
+            st.markdown(f"📁 `{filepath}`")
+            
+            # Display full original file content in a scrollable code block
+            with st.expander("View Original File Content", expanded=True):
+                # Prepare content for the copy button (needs JS-specific escaping)
+                raw_content_for_copy = result['original_file_content']
+                # Escape for JavaScript template literal: backticks, backslashes, newlines
+                js_escaped_content = raw_content_for_copy.replace('\\', '\\\\') \
+                                                       .replace('`', '\\`') \
+                                                       .replace('\r\n', '\\n') \
+                                                       .replace('\n', '\\n') \
+                                                       .replace('\r', '\\n')
+
+                unique_suffix = str(result.name if hasattr(result, 'name') else result['task_id']).replace('-', '_').replace('.', '_')
+                button_id = f"copyBtnOriginal_{unique_suffix}"
+                
+                copy_button_html = f"""
+                    <button id="{button_id}" onclick="copyOriginalToClipboard(`{js_escaped_content}`, '{button_id}')" style="margin-bottom: 10px; padding: 5px 10px; border-radius: 5px; border: 1px solid #ccc; cursor: pointer;">Copy Original File</button>
+                    <script>
+                        if (!window.copyOriginalToClipboard) {{
+                            window.copyOriginalToClipboard = async function(text, buttonId) {{
+                                try {{
+                                    await navigator.clipboard.writeText(text);
+                                    const button = document.getElementById(buttonId);
+                                    button.innerText = 'Copied!';
+                                    button.style.backgroundColor = '#d4edda'; // Optional: success feedback
+                                    setTimeout(() => {{ 
+                                        button.innerText = 'Copy Original File'; 
+                                        button.style.backgroundColor = '';
+                                    }}, 2000);
+                                }} catch (err) {{
+                                    console.error('Failed to copy original: ', err);
+                                    const button = document.getElementById(buttonId);
+                                    button.innerText = 'Copy Failed!';
+                                    button.style.backgroundColor = '#f8d7da'; // Optional: error feedback
+                                    setTimeout(() => {{ 
+                                        button.innerText = 'Copy Original File'; 
+                                        button.style.backgroundColor = '';
+                                    }}, 2000);
+                                }}
+                            }}
+                        }}
+                    </script>
+                    """
+                st.components.v1.html(copy_button_html, height=50)
+
+                # Prepare content for st.code (needs actual newlines)
+                content_for_display = result['original_file_content']
+                # Iteratively replace common escaped newline sequences with actual newlines
+                # This handles cases like "\\n" -> "\n" and then "\n" (if it was literally "\n")
+                # Order might matter if there are multiple levels of escaping, but this covers common ones.
+                content_for_display = content_for_display.replace('\\\\r\\\\n', '\r\n').replace('\\\\n', '\n') # Double escaped
+                content_for_display = content_for_display.replace('\\r\\n', '\r\n').replace('\\n', '\n')     # Single escaped
+
+                language = guess_language_from_filepath(filepath)
+                st.code(content_for_display, language=language, line_numbers=False)
+
+        else:
+            st.warning("Original file content not available")
+    
+    with col2:
+        st.markdown("**Edit Analysis:**")
+        
+        if not result['succeeded']:
+            # Show error information
+            st.error("❌ **Edit Failed**")
+            
+            if not pd.isna(result['error_enum']):
+                st.markdown(f"**Error Code:** {result['error_enum']}")
+        else:
+            # Show successful edit information
+            st.success("✅ **Edit Successful**")
+            
+            # Show edit metrics
+            metric_col1, metric_col2, metric_col3 = st.columns(3)
+            
+            with metric_col1:
+                if not pd.isna(result['num_edits']):
+                    st.metric("Edits", int(result['num_edits']))
+            
+            with metric_col2:
+                if not pd.isna(result['num_lines_added']):
+                    st.metric("Added", int(result['num_lines_added']))
+            
+            with metric_col3:
+                if not pd.isna(result['num_lines_deleted']):
+                    st.metric("Deleted", int(result['num_lines_deleted']))
+            
+            # Show edited file if available
+            if has_edited:
+                st.markdown("**Edited File:**")
+                with st.expander("View Edited File Content"):
+                    edited_lines = result['edited_file_content'].split('\n')
+                    for i, line in enumerate(edited_lines[:50], 1):
+                        st.text(f"{i:3d} | {line}")
+                    
+                    if len(edited_lines) > 50:
+                        st.text(f"... ({len(edited_lines) - 50} more lines)")
+        
+        # Show parsed tool call if available
+        if not pd.isna(result['parsed_tool_call_json']):
+            with st.expander("View Parsed Tool Call"):
+                try:
+                    parsed_call = json.loads(result['parsed_tool_call_json'])
+                    st.json(parsed_call)
+                except:
+                    st.text(result['parsed_tool_call_json'])
+
+def render_raw_output_view(result):
+    """Render raw model output"""
+    st.markdown("#### 🤖 Raw Model Output")
+    
+    if pd.isna(result['raw_model_output']) or not result['raw_model_output']:
+        st.warning("No raw output available for this result.")
+        return
+    
+    st.markdown("""
+    <div class="file-viewer">
+    """, unsafe_allow_html=True)
+    
+    st.text(result['raw_model_output'])
+    
+    st.markdown("</div>", unsafe_allow_html=True)
+
+def render_parsed_tool_call_view(result):
+    """Render parsed tool call analysis"""
+    st.markdown("#### 🔧 Parsed Tool Call Analysis")
+    
+    if pd.isna(result['parsed_tool_call_json']) or not result['parsed_tool_call_json']:
+        st.warning("No parsed tool call available for this result.")
+        return
+    
+    try:
+        parsed_call = json.loads(result['parsed_tool_call_json'])
+        
+        # Pretty print the JSON
+        st.json(parsed_call)
+        
+        # If it's a replace_in_file call, show the diff blocks
+        if isinstance(parsed_call, dict) and 'diff' in parsed_call:
+            st.markdown("**Diff Blocks:**")
+            st.code(parsed_call['diff'], language='diff')
+            
+    except json.JSONDecodeError:
+        st.markdown("**Raw Parsed Call (Invalid JSON):**")
+        st.text(result['parsed_tool_call_json'])
+
+def render_metrics_view(result):
+    """Render detailed metrics for the result"""
+    st.markdown("#### 📊 Detailed Metrics")
+    
+    col1, col2 = st.columns(2)
+    
+    with col1:
+        st.markdown("**Timing Metrics:**")
+        if not pd.isna(result['time_to_first_token_ms']):
+            st.metric("Time to First Token", f"{result['time_to_first_token_ms']:.0f}ms")
+        
+        if not pd.isna(result['time_to_first_edit_ms']):
+            st.metric("Time to First Edit", f"{result['time_to_first_edit_ms']:.0f}ms")
+        
+        if not pd.isna(result['time_round_trip_ms']):
+            st.metric("Round Trip Time", f"{result['time_round_trip_ms']:.0f}ms")
+    
+    with col2:
+        st.markdown("**Token & Cost Metrics:**")
+        if not pd.isna(result['completion_tokens']):
+            st.metric("Completion Tokens", int(result['completion_tokens']))
+        
+        if not pd.isna(result['cost_usd']):
+            st.metric("Cost", f"${result['cost_usd']:.4f}")
+        
+        if not pd.isna(result['tokens_in_context']):
+            st.metric("Context Tokens", int(result['tokens_in_context']))
+
+def guess_language_from_filepath(filepath):
+    """Guess the language for syntax highlighting from filepath."""
+    if not filepath or pd.isna(filepath):
+        return None
+    
+    extension_map = {
+        '.py': 'python',
+        '.js': 'javascript',
+        '.ts': 'typescript',
+        '.java': 'java',
+        '.cs': 'csharp',
+        '.cpp': 'cpp',
+        '.c': 'c',
+        '.html': 'html',
+        '.css': 'css',
+        '.json': 'json',
+        '.sql': 'sql',
+        '.md': 'markdown',
+        '.rb': 'ruby',
+        '.php': 'php',
+        '.go': 'go',
+        '.rs': 'rust',
+        '.swift': 'swift',
+        '.kt': 'kotlin',
+        '.sh': 'bash',
+        '.yaml': 'yaml',
+        '.yml': 'yaml',
+        '.xml': 'xml',
+    }
+    
+    _, ext = os.path.splitext(filepath)
+def main():
+    # Add a note about valid attempts
+    st.sidebar.markdown("""
+    ### Note on Metrics
+    Success rates are calculated based on **valid results only**. 
+    
+    Invalid results (where the model didn't call the diff edit tool or edited the wrong file) are excluded from calculations.
+    """)
+    
+    # Initialize session state
+    if 'drill_down_model' not in st.session_state:
+        st.session_state.drill_down_model = None
+    if 'selected_run_id' not in st.session_state:
+        st.session_state.selected_run_id = None
+    
+    # Load all runs for sidebar
+    all_runs = load_all_runs()
+    
+    if all_runs.empty:
+        st.error("No evaluation runs found in the database.")
+        st.stop()
+    
+    # Sidebar for run selection
+    with st.sidebar:
+        st.markdown("## 📊 Evaluation Runs")
+        st.markdown("Select a run to analyze:")
+        
+        # Create run options with nice formatting
+        run_options = []
+        run_ids = []
+        
+        for idx, run in all_runs.iterrows():
+            # Format the run description nicely
+            date_str = run['created_at'][:10]  # Get just the date part
+            time_str = run['created_at'][11:16]  # Get just the time part
+            
+            if run['description']:
+                display_name = f"🚀 {run['description']}"
+            else:
+                display_name = f"📅 Run {run['run_id'][:8]}..."
+            
+            run_options.append(f"{display_name}\n📅 {date_str} {time_str}")
+            run_ids.append(run['run_id'])
+        
+        # Default to latest run if no selection
+        if st.session_state.selected_run_id is None:
+            default_index = 0  # Latest run is first
+            st.session_state.selected_run_id = run_ids[0]
+        else:
+            try:
+                default_index = run_ids.index(st.session_state.selected_run_id)
+            except ValueError:
+                default_index = 0
+                st.session_state.selected_run_id = run_ids[0]
+        
+        selected_run_idx = st.selectbox(
+            "Choose run:",
+            range(len(run_options)),
+            format_func=lambda x: run_options[x],
+            index=default_index,
+            key="run_selector"
+        )
+        
+        # Update selected run if changed
+        if run_ids[selected_run_idx] != st.session_state.selected_run_id:
+            st.session_state.selected_run_id = run_ids[selected_run_idx]
+            st.session_state.drill_down_model = None  # Reset drill down when changing runs
+            st.rerun()
+        
+        # Show run details in sidebar
+        selected_run = all_runs.iloc[selected_run_idx]
+        st.markdown("---")
+        st.markdown("### 📋 Run Details")
+        st.markdown(f"**Run ID:** `{selected_run['run_id'][:12]}...`")
+        st.markdown(f"**Created:** {selected_run['created_at']}")
+        if selected_run['description']:
+            st.markdown(f"**Description:** {selected_run['description']}")
+    
+    # Load data for selected run
+    current_run, model_performance = load_run_comparison(st.session_state.selected_run_id)
+    
+    if current_run is None or model_performance.empty:
+        st.error("No data found for the selected run.")
+        st.stop()
+    
+    # Render main dashboard
+    render_hero_section(current_run, model_performance)
+    
+    # Check if we're in drill-down mode
+    if st.session_state.drill_down_model:
+        col1, col2 = st.columns([1, 4])
+        with col1:
+            if st.button("Back to Overview", use_container_width=True):
+                st.session_state.drill_down_model = None
+                st.rerun()
+        
+        render_detailed_analysis(current_run['run_id'], st.session_state.drill_down_model)
+    else:
+        # Success Rate Comparison
+        fig_success = px.bar(
+            model_performance,
+            x='model_id',
+            y='success_rate',
+            title="Success Rate by Model",
+            labels={'success_rate': 'Success Rate', 'model_id': 'Model'},
+            color='success_rate',
+            color_continuous_scale='RdYlGn',
+            text='success_rate',
+            template='plotly_dark'
+        )
+        fig_success.update_traces(texttemplate='%{text:.1%}', textposition='outside')
+        fig_success.update_layout(
+            showlegend=False,
+            plot_bgcolor='rgba(0,0,0,0)',
+            paper_bgcolor='rgba(0,0,0,0)',
+            font=dict(family="Azeret Mono, monospace"),
+            yaxis_range=[0,1],  # Set y-axis from 0% to 100%
+            margin=dict(t=50)  # Add top margin to prevent clipping
+        )
+        st.plotly_chart(fig_success, use_container_width=True)
+        
+        render_model_comparison_cards(model_performance)
+        render_comparison_charts(model_performance)
+
+if __name__ == "__main__":
+    main()

evals/diff-edits/dashboard/launch.sh

@@ -0,0 +1,33 @@
+#!/bin/bash
+
+# Diff Edits Evaluation Dashboard Launcher
+echo "🚀 Starting Diff Edits Evaluation Dashboard..."
+
+# Check if we're in the right directory
+if [ ! -f "app.py" ]; then
+    echo "❌ Error: app.py not found. Please run this script from the dashboard directory."
+    exit 1
+fi
+
+# Check if database exists
+if [ ! -f "../evals.db" ]; then
+    echo "⚠️  Warning: Database file ../evals.db not found."
+    echo "   Make sure you've run some evaluations first to populate the database."
+    echo "   You can run: node ../cli/dist/index.js run-diff-eval --model-id anthropic/claude-sonnet-4 --max-cases 1"
+    echo ""
+fi
+
+# Check if requirements are installed
+echo "📦 Checking Python dependencies..."
+if ! python -c "import streamlit, plotly, pandas" 2>/dev/null; then
+    echo "📥 Installing required packages..."
+    pip install -r requirements.txt
+fi
+
+echo "🌐 Launching Streamlit dashboard..."
+echo "   Dashboard will open in your browser at http://localhost:8501"
+echo "   Press Ctrl+C to stop the dashboard"
+echo ""
+
+# Launch Streamlit
+streamlit run app.py

evals/diff-edits/dashboard/pages/02_Bad_Cases.py

@@ -0,0 +1,183 @@
+import streamlit as st
+import pandas as pd
+import json
+import os # Need to import os for load_case_raw_data
+from utils import get_database_connection, guess_language_from_filepath # Absolute import
+
+st.set_page_config(
+    page_title="Case Health Inspector",
+    page_icon="🧑‍⚕️",
+    layout="wide"
+)
+
+st.title("Case Health Inspector")
+st.markdown("Identify test cases that are frequently problematic across different models and runs.")
+
+@st.cache_data
+def load_problematic_cases_summary():
+    conn = get_database_connection()
+    query = """
+    WITH case_attempts AS (
+        SELECT
+            c.task_id,
+            c.description AS case_description,
+            f_orig.filepath AS original_filepath, -- Get from files table
+            r.run_id,
+            r.model_id,
+            r.result_id,
+            (CASE WHEN (r.error_enum NOT IN (1, 6, 7) OR r.error_enum IS NULL) THEN 1 ELSE 0 END) AS is_valid_attempt,
+            (CASE WHEN (r.error_enum NOT IN (1, 6, 7) OR r.error_enum IS NULL) THEN r.succeeded ELSE NULL END) AS succeeded_on_valid
+        FROM cases c
+        JOIN results r ON c.case_id = r.case_id
+        LEFT JOIN files f_orig ON c.file_hash = f_orig.hash -- Join to get original filepath
+    ),
+    case_summary AS (
+        SELECT
+            task_id,
+            case_description,
+            original_filepath, -- This is now f_orig.filepath
+            COUNT(DISTINCT run_id) AS num_benchmark_runs,
+            COUNT(result_id) AS total_attempts,
+            SUM(is_valid_attempt) AS total_valid_attempts,
+            SUM(succeeded_on_valid) AS total_successful_valid_attempts
+        FROM case_attempts
+        GROUP BY task_id, case_description, original_filepath -- original_filepath is f_orig.filepath
+    )
+    SELECT
+        task_id,
+        case_description,
+        original_filepath, -- This is f_orig.filepath from case_summary
+        num_benchmark_runs,
+        total_attempts,
+        total_valid_attempts,
+        CAST(total_valid_attempts AS REAL) * 100.0 / total_attempts AS percent_valid_attempts,
+        CASE 
+            WHEN total_valid_attempts > 0 THEN CAST(total_successful_valid_attempts AS REAL) * 100.0 / total_valid_attempts
+            ELSE 0 
+        END AS success_rate_on_valid
+    FROM case_summary
+    ORDER BY percent_valid_attempts ASC, success_rate_on_valid ASC;
+    """
+    df = pd.read_sql_query(query, conn)
+    return df
+
+@st.cache_data
+def load_case_raw_data(task_id):
+    """Loads the original JSON data for a given task_id."""
+    # This assumes test cases are stored in ../cases relative to this script's parent (dashboard)
+    # So, ../../cases from this script's location (pages/02_Bad_Cases.py)
+    # Correct path from this script (pages/02_Bad_Cases.py) to cases/
+    # os.path.dirname(__file__) -> pages
+    # os.path.join(..., '..') -> dashboard
+    # os.path.join(..., '..', '..') -> diff-edits
+    # os.path.join(..., '..', '..', 'cases') -> diff-edits/cases
+    cases_dir = os.path.join(os.path.dirname(__file__), '..', '..', 'cases')
+    
+    # The task_id is usually the filename without .json
+    # However, some task_ids might have suffixes or be different.
+    # We need a robust way to find the file. For now, assume task_id is filename base.
+    # This might need adjustment if task_id format varies significantly from filename.
+    
+    # Try direct match first
+    potential_filename = f"{task_id}.json"
+    filepath = os.path.join(cases_dir, potential_filename)
+
+    if not os.path.exists(filepath):
+        # If direct match fails, list files and try to find one that starts with task_id
+        # This is a simple fallback, might need more robust matching if task_ids are complex
+        try:
+            for f_name in os.listdir(cases_dir):
+                if f_name.startswith(task_id) and f_name.endswith(".json"):
+                    filepath = os.path.join(cases_dir, f_name)
+                    break
+            else: # No break means no file found
+                 return None # File not found
+        except FileNotFoundError:
+            return None # Cases directory itself not found
+
+    if not os.path.exists(filepath): # Check again after potential find
+        return None
+
+    try:
+        with open(filepath, 'r') as f:
+            return json.load(f)
+    except Exception as e:
+        st.error(f"Error loading case file {filepath}: {e}")
+        return None
+
+def render_problematic_cases_page():
+    summary_df = load_problematic_cases_summary()
+
+    if summary_df.empty:
+        st.warning("No case summary data found. Run some evaluations first.")
+        return
+
+    st.markdown("### Cases Overview")
+    st.dataframe(summary_df.style.format({
+        "percent_valid_attempts": "{:.1f}%",
+        "success_rate_on_valid": "{:.1f}%"
+    }), use_container_width=True)
+
+    st.markdown("---")
+    st.markdown("### Case Drill Down")
+    
+    selected_task_id = st.selectbox(
+        "Select a Case ID (task_id) to inspect:",
+        options=[""] + summary_df['task_id'].tolist() # Add a blank option
+    )
+
+    if selected_task_id:
+        case_data = summary_df[summary_df['task_id'] == selected_task_id].iloc[0]
+        st.subheader(f"Details for Case: {case_data['task_id']}")
+        st.markdown(f"**Description:** {case_data['case_description']}")
+        st.markdown(f"**Original Filepath:** `{case_data['original_filepath']}`")
+        
+        raw_json_data = load_case_raw_data(selected_task_id)
+        if raw_json_data:
+            with st.expander("View Raw Case JSON Data", expanded=False):
+                st.json(raw_json_data)
+            
+            if 'file_contents' in raw_json_data and raw_json_data['file_contents']:
+                with st.expander("View Original File Content (from Case JSON)", expanded=True):
+                    # Prepare content for the copy button
+                    raw_content_for_copy = raw_json_data['file_contents']
+                    js_escaped_content = raw_content_for_copy.replace('\\', '\\\\') \
+                                                           .replace('`', '\\`') \
+                                                           .replace('\r\n', '\\n') \
+                                                           .replace('\n', '\\n') \
+                                                           .replace('\r', '\\n')
+                    button_id = f"copyBtnCase_{selected_task_id.replace('-', '_').replace('.', '_')}"
+                    copy_button_html = f"""
+                        <button id="{button_id}" onclick="copyCaseContentToClipboard(`{js_escaped_content}`, '{button_id}')" style="margin-bottom: 10px; padding: 5px 10px; border-radius: 5px; border: 1px solid #ccc; cursor: pointer;">Copy File Content</button>
+                        <script>
+                            if (!window.copyCaseContentToClipboard) {{
+                                window.copyCaseContentToClipboard = async function(text, buttonId) {{
+                                    try {{
+                                        await navigator.clipboard.writeText(text);
+                                        const button = document.getElementById(buttonId);
+                                        button.innerText = 'Copied!';
+                                        setTimeout(() => {{ button.innerText = 'Copy File Content'; }}, 2000);
+                                    }} catch (err) {{ console.error('Failed to copy: ', err); const button = document.getElementById(buttonId); button.innerText = 'Copy Failed!'; setTimeout(() => {{ button.innerText = 'Copy File Content'; }}, 2000); }}
+                                }}
+                            }}
+                        </script>
+                        """
+                    st.components.v1.html(copy_button_html, height=50)
+
+                    # Prepare content for st.code
+                    content_for_display = raw_json_data['file_contents']
+                    content_for_display = content_for_display.replace('\\\\r\\\\n', '\r\n').replace('\\\\n', '\n')
+                    content_for_display = content_for_display.replace('\\r\\n', '\r\n').replace('\\n', '\n')
+                    
+                    language = guess_language_from_filepath(case_data['original_filepath'])
+                    st.code(content_for_display, language=language, line_numbers=False)
+            else:
+                st.warning("Original file content not found in case JSON.")
+        else:
+            st.error(f"Could not load raw JSON data for case: {selected_task_id}")
+        
+        # Placeholder for more detailed stats (per-model performance on this case, error breakdown)
+        st.markdown("*(Further per-model statistics and error breakdowns for this case can be added here.)*")
+
+if __name__ == "__main__":
+    render_problematic_cases_page()

evals/diff-edits/dashboard/requirements.txt

@@ -0,0 +1,4 @@
+streamlit>=1.28.0
+plotly>=5.17.0
+pandas>=2.0.0
+numpy>=1.24.0

evals/diff-edits/dashboard/utils.py

@@ -0,0 +1,51 @@
+import streamlit as st
+import sqlite3
+import pandas as pd
+import os
+
+@st.cache_resource
+def get_database_connection():
+    # Assuming the script is run from the dashboard directory,
+    # evals.db is two levels up from there.
+    # __file__ is utils.py, its dirname is dashboard.
+    # os.path.dirname(__file__) -> dashboard/
+    # os.path.join(..., '..') -> diff-edits/
+    # os.path.join(..., '..', 'evals.db') -> diff-edits/evals.db
+    db_path = os.path.join(os.path.dirname(__file__), '..', 'evals.db')
+    if not os.path.exists(db_path):
+        st.error(f"Database not found. Expected at: {os.path.abspath(db_path)}")
+        st.stop()
+    return sqlite3.connect(db_path, check_same_thread=False)
+
+def guess_language_from_filepath(filepath):
+    """Guess the language for syntax highlighting from filepath."""
+    if not filepath or pd.isna(filepath):
+        return None
+    
+    extension_map = {
+        '.py': 'python',
+        '.js': 'javascript',
+        '.ts': 'typescript',
+        '.java': 'java',
+        '.cs': 'csharp',
+        '.cpp': 'cpp',
+        '.c': 'c',
+        '.html': 'html',
+        '.css': 'css',
+        '.json': 'json',
+        '.sql': 'sql',
+        '.md': 'markdown',
+        '.rb': 'ruby',
+        '.php': 'php',
+        '.go': 'go',
+        '.rs': 'rust',
+        '.swift': 'swift',
+        '.kt': 'kotlin',
+        '.sh': 'bash',
+        '.yaml': 'yaml',
+        '.yml': 'yaml',
+        '.xml': 'xml',
+    }
+    
+    _, ext = os.path.splitext(str(filepath)) # Ensure filepath is string
+    return extension_map.get(ext.lower(), None)

evals/diff-edits/database.md

@@ -0,0 +1,84 @@
+# Diff Edit Evaluation Database Schema
+
+This document provides an overview of the SQLite database schema used for the diff edit evaluation suite. The database is designed to capture every aspect of the evaluation runs in a structured way, allowing for detailed, multi-dimensional analysis and ensuring full reproducibility of our findings.
+
+## Data Model Overview
+
+The database is composed of several interconnected tables that work together to provide a comprehensive picture of each evaluation. The core of the model revolves around `runs`, `cases`, and `results`.
+
+### `runs`
+
+A `run` represents a single, top-level execution of the evaluation script (e.g., one invocation of `npm run diff-eval`). It serves as the main container for a complete benchmark session.
+
+-   **Purpose**: To group all the results from a single benchmark execution, allowing for high-level comparison between different runs over time.
+-   **Key Columns**:
+    -   `run_id`: A unique identifier for the entire run.
+    -   `description`: A human-readable summary of the run's configuration (e.g., which models were tested, how many cases, etc.).
+    -   `system_prompt_hash`: A foreign key that links this run to the specific system prompt that was used, ensuring we can track performance changes based on prompt modifications.
+
+### `cases`
+
+A `case` represents a single test scenario that is presented to a model. It corresponds to one of the JSON files in the `cases/` directory and links that static definition to a specific benchmark `run`.
+
+-   **Purpose**: To track the individual test scenarios within a given run.
+-   **Key Columns**:
+    -   `case_id`: A unique identifier for the case *within* a specific run.
+    -   `run_id`: A foreign key linking back to the parent `run`.
+    -   `task_id`: The original, persistent identifier for the test case (typically from the JSON filename).
+    -   `file_hash`: A foreign key linking to the original, un-edited file content for this case.
+
+### `results`
+
+This is the most granular and important table in the database. A `result` represents the outcome of a single attempt by a specific model on a specific case.
+
+-   **Purpose**: To store the detailed outcome of every single model attempt, providing the raw data for all quantitative and qualitative analysis.
+-   **Key Columns**:
+    -   `result_id`: A unique identifier for the individual attempt.
+    -   `run_id`, `case_id`, `model_id`, `processing_functions_hash`: A set of foreign keys that precisely situate this result within the context of a specific run, case, model, and set of helper functions.
+    -   `succeeded`: A boolean indicating if the generated diff was applied successfully.
+    -   `error_enum`: A numeric code representing the specific type of error if the attempt failed (e.g., `1` for `no_tool_calls`, `7` for `wrong_file_edited`).
+    -   `num_edits`, `num_lines_deleted`, `num_lines_added`: Quantitative metrics about the structure of the generated diff.
+    -   `time_to_first_token_ms`, `time_to_first_edit_ms`, `time_round_trip_ms`: High-precision timing data to measure model latency.
+    -   `cost_usd`, `completion_tokens`: Cost and token usage metrics for efficiency analysis.
+    -   `raw_model_output`, `file_edited_hash`, `parsed_tool_call_json`: The rich, qualitative data. This includes the model's full, raw response and the parsed tool calls, which are invaluable for debugging and understanding the model's reasoning.
+
+---
+
+## Supporting Tables
+
+The following tables store versioned, deduplicated content to ensure data integrity and efficiency.
+
+### `system_prompts`
+
+-   **Purpose**: Stores the versioned content of the system prompts used in evaluations.
+-   **Key Columns**:
+    -   `hash`: A unique hash of the prompt's content, which acts as the primary key. This prevents duplicate storage of the same prompt.
+    -   `name`: A human-readable name for the prompt (e.g., `basicSystemPrompt`, `claude4SystemPrompt`).
+    -   `content`: The full text of the system prompt.
+
+### `processing_functions`
+
+-   **Purpose**: Stores the versioned combinations of parsing and diff-editing functions.
+-   **Key Columns**:
+    -   `hash`: A unique hash of the function combination name.
+    -   `name`: A human-readable name (e.g., `parseV2-diffV2`).
+    -   `parsing_function`: The name of the function used to parse the model's output.
+    -   `diff_edit_function`: The name of the function used to apply the diff.
+
+### `files`
+
+-   **Purpose**: Stores the content of all files involved in the tests, including the original source files and the diffs generated by the models.
+-   **Key Columns**:
+    -   `hash`: A content-based hash of the file, ensuring that identical files are only stored once.
+    -   `filepath`: The original path of the file.
+    -   `content`: The full content of the file.
+
+## The Bigger Picture
+
+This relational schema provides a powerful foundation for sophisticated analysis. It moves beyond simple pass/fail metrics and allows us to explore the nuanced interactions between models, prompts, and the code they operate on. With this database, we can answer critical questions like:
+
+-   "How does prompt engineering affect not just success rate, but also latency and cost?"
+-   "Are certain models more prone to specific types of errors (e.g., hallucinating file paths vs. failing to call a tool)?"
+-   "Which of our internal diffing algorithms is the most robust against a wide range of model-generated edits?"
+
+Ultimately, this data model enables us to move from simply *measuring* performance to truly *understanding* it, providing the insights needed to build more capable and reliable AI engineering systems.

evals/diff-edits/database/client.ts

@@ -0,0 +1,135 @@
+import Database from 'better-sqlite3';
+import * as fs from 'fs';
+import * as path from 'path';
+import * as crypto from 'crypto';
+
+export class DatabaseClient {
+  private static instance: DatabaseClient;
+  private db: Database.Database;
+  private dbPath: string;
+
+  private constructor() {
+    // Get database path from environment or use default
+    this.dbPath = process.env.DIFF_EVALS_DB_PATH || path.join(__dirname, '../evals.db');
+    
+    // Ensure directory exists
+    const dbDir = path.dirname(this.dbPath);
+    if (!fs.existsSync(dbDir)) {
+      fs.mkdirSync(dbDir, { recursive: true });
+    }
+
+    // Initialize database connection
+    this.db = new Database(this.dbPath);
+    
+    // Enable WAL mode for concurrent access
+    this.db.pragma('journal_mode = WAL');
+    
+    // Enable foreign key constraints
+    this.db.pragma('foreign_keys = ON');
+    
+    // Initialize schema if needed
+    this.initializeSchema();
+  }
+
+  static getInstance(): DatabaseClient {
+    if (!DatabaseClient.instance) {
+      DatabaseClient.instance = new DatabaseClient();
+    }
+    return DatabaseClient.instance;
+  }
+
+  private initializeSchema(): void {
+    // Check if tables exist by trying to query one of them
+    try {
+      this.db.prepare('SELECT COUNT(*) FROM system_prompts LIMIT 1').get();
+      // If we get here, tables exist
+      return;
+    } catch (error) {
+      // Tables don't exist, create them
+      console.log('Initializing database schema...');
+      this.createTables();
+    }
+  }
+
+  private createTables(): void {
+    const schemaPath = path.join(__dirname, 'schema.sql');
+    const schema = fs.readFileSync(schemaPath, 'utf8');
+    
+    // Execute the entire schema as one block
+    this.db.transaction(() => {
+      this.db.exec(schema);
+    })();
+
+    console.log('Database schema initialized successfully');
+  }
+
+  getDatabase(): Database.Database {
+    return this.db;
+  }
+
+  getDatabasePath(): string {
+    return this.dbPath;
+  }
+
+  // Utility method to generate SHA-256 hash
+  static generateHash(content: string): string {
+    return crypto.createHash('sha256').update(content).digest('hex');
+  }
+
+  // Utility method to generate UUID-like ID
+  static generateId(): string {
+    return crypto.randomUUID();
+  }
+
+  // Transaction wrapper
+  transaction<T>(fn: () => T): T {
+    return this.db.transaction(fn)();
+  }
+
+  // Close database connection (for cleanup)
+  close(): void {
+    if (this.db) {
+      this.db.close();
+    }
+  }
+
+  // Get database info
+  getInfo(): { path: string; size: number; tables: string[] } {
+    const stats = fs.statSync(this.dbPath);
+    const tables = this.db
+      .prepare("SELECT name FROM sqlite_master WHERE type='table' ORDER BY name")
+      .all()
+      .map((row: any) => row.name);
+
+    return {
+      path: this.dbPath,
+      size: stats.size,
+      tables
+    };
+  }
+
+  // Vacuum database (cleanup and optimize)
+  vacuum(): void {
+    this.db.exec('VACUUM');
+  }
+
+  // Get database statistics
+  getStats(): { [tableName: string]: number } {
+    const tables = ['system_prompts', 'processing_functions', 'files', 'runs', 'cases', 'results'];
+    const stats: { [tableName: string]: number } = {};
+
+    for (const table of tables) {
+      try {
+        const result = this.db.prepare(`SELECT COUNT(*) as count FROM ${table}`).get() as { count: number };
+        stats[table] = result.count;
+      } catch (error) {
+        stats[table] = 0;
+      }
+    }
+
+    return stats;
+  }
+}
+
+// Export singleton instance getter
+export const getDatabase = () => DatabaseClient.getInstance();

evals/diff-edits/database/index.ts

@@ -0,0 +1,23 @@
+// Main database module exports
+export { DatabaseClient, getDatabase } from './client';
+export * from './types';
+export * from './operations';
+export * from './queries';
+
+// Re-export commonly used functions for convenience
+export {
+  upsertSystemPrompt,
+  upsertProcessingFunctions,
+  upsertFile,
+  createBenchmarkRun,
+  createCase,
+  insertResult,
+  getRunStats
+} from './operations';
+
+export {
+  getSuccessRatesByModel,
+  getModelComparisons,
+  getDatabaseSummary,
+  getErrorDistribution
+} from './queries';

evals/diff-edits/database/operations.ts

@@ -0,0 +1,348 @@
+import { DatabaseClient } from './client';
+import {
+  SystemPrompt,
+  ProcessingFunctions,
+  FileRecord,
+  BenchmarkRun,
+  Case,
+  Result,
+  CreateSystemPromptInput,
+  CreateProcessingFunctionsInput,
+  CreateFileInput,
+  CreateBenchmarkRunInput,
+  CreateCaseInput,
+  CreateResultInput
+} from './types';
+
+const db = DatabaseClient.getInstance();
+
+// System Prompts Operations
+export async function upsertSystemPrompt(input: CreateSystemPromptInput): Promise<string> {
+  const hash = DatabaseClient.generateHash(input.content);
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT OR IGNORE INTO system_prompts (hash, name, content)
+    VALUES (?, ?, ?)
+  `);
+  
+  stmt.run(hash, input.name, input.content);
+  return hash;
+}
+
+export async function getSystemPromptByHash(hash: string): Promise<SystemPrompt | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM system_prompts WHERE hash = ?
+  `);
+  
+  const result = stmt.get(hash) as SystemPrompt | undefined;
+  return result || null;
+}
+
+// Processing Functions Operations
+export async function upsertProcessingFunctions(input: CreateProcessingFunctionsInput): Promise<string> {
+  const hash = DatabaseClient.generateHash(input.parsing_function + input.diff_edit_function);
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT OR IGNORE INTO processing_functions (hash, name, parsing_function, diff_edit_function)
+    VALUES (?, ?, ?, ?)
+  `);
+  
+  stmt.run(hash, input.name, input.parsing_function, input.diff_edit_function);
+  return hash;
+}
+
+export async function getProcessingFunctionsByHash(hash: string): Promise<ProcessingFunctions | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM processing_functions WHERE hash = ?
+  `);
+  
+  const result = stmt.get(hash) as ProcessingFunctions | undefined;
+  return result || null;
+}
+
+// Files Operations
+export async function upsertFile(input: CreateFileInput): Promise<string> {
+  const hash = DatabaseClient.generateHash(input.content);
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT OR IGNORE INTO files (hash, filepath, content, tokens)
+    VALUES (?, ?, ?, ?)
+  `);
+  
+  stmt.run(hash, input.filepath, input.content, input.tokens || null);
+  return hash;
+}
+
+export async function getFileByHash(hash: string): Promise<FileRecord | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM files WHERE hash = ?
+  `);
+  
+  const result = stmt.get(hash) as FileRecord | undefined;
+  return result || null;
+}
+
+// Benchmark Runs Operations
+export async function createBenchmarkRun(input: CreateBenchmarkRunInput): Promise<string> {
+  const runId = DatabaseClient.generateId();
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT INTO runs (run_id, description, system_prompt_hash)
+    VALUES (?, ?, ?)
+  `);
+  
+  stmt.run(runId, input.description || null, input.system_prompt_hash);
+  return runId;
+}
+
+export async function getBenchmarkRun(runId: string): Promise<BenchmarkRun | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM runs WHERE run_id = ?
+  `);
+  
+  const result = stmt.get(runId) as BenchmarkRun | undefined;
+  return result || null;
+}
+
+export async function getAllBenchmarkRuns(): Promise<BenchmarkRun[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM runs ORDER BY created_at DESC
+  `);
+  
+  return stmt.all() as BenchmarkRun[];
+}
+
+// Cases Operations
+export async function createCase(input: CreateCaseInput): Promise<string> {
+  const caseId = DatabaseClient.generateId();
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT INTO cases (case_id, run_id, description, system_prompt_hash, task_id, tokens_in_context, file_hash)
+    VALUES (?, ?, ?, ?, ?, ?, ?)
+  `);
+  
+  stmt.run(
+    caseId,
+    input.run_id,
+    input.description,
+    input.system_prompt_hash,
+    input.task_id,
+    input.tokens_in_context,
+    input.file_hash || null
+  );
+  
+  return caseId;
+}
+
+export async function getCasesByRun(runId: string): Promise<Case[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM cases WHERE run_id = ? ORDER BY created_at
+  `);
+  
+  return stmt.all(runId) as Case[];
+}
+
+export async function getCaseById(caseId: string): Promise<Case | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM cases WHERE case_id = ?
+  `);
+  
+  const result = stmt.get(caseId) as Case | undefined;
+  return result || null;
+}
+
+// Results Operations
+export async function insertResult(input: CreateResultInput): Promise<string> {
+  const resultId = DatabaseClient.generateId();
+  
+  const stmt = db.getDatabase().prepare(`
+    INSERT INTO results (
+      result_id, run_id, case_id, model_id, processing_functions_hash,
+      succeeded, error_enum, num_edits, num_lines_deleted, num_lines_added,
+      time_to_first_token_ms, time_to_first_edit_ms, time_round_trip_ms,
+      cost_usd, completion_tokens, raw_model_output, file_edited_hash,
+      parsed_tool_call_json
+    ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+  `);
+  
+  stmt.run(
+    resultId,
+    input.run_id,
+    input.case_id,
+    input.model_id,
+    input.processing_functions_hash,
+    input.succeeded ? 1 : 0, // Convert boolean to integer
+    input.error_enum || null,
+    input.num_edits || null,
+    input.num_lines_deleted || null,
+    input.num_lines_added || null,
+    input.time_to_first_token_ms || null,
+    input.time_to_first_edit_ms || null,
+    input.time_round_trip_ms || null,
+    input.cost_usd || null,
+    input.completion_tokens || null,
+    input.raw_model_output || null,
+    input.file_edited_hash || null,
+    input.parsed_tool_call_json || null
+  );
+  
+  return resultId;
+}
+
+export async function getResultsByRun(runId: string): Promise<Result[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM results WHERE run_id = ? ORDER BY created_at
+  `);
+  
+  return stmt.all(runId) as Result[];
+}
+
+export async function getResultsByCase(caseId: string): Promise<Result[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM results WHERE case_id = ? ORDER BY created_at
+  `);
+  
+  return stmt.all(caseId) as Result[];
+}
+
+export async function getResultById(resultId: string): Promise<Result | null> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM results WHERE result_id = ?
+  `);
+  
+  const result = stmt.get(resultId) as Result | undefined;
+  return result || null;
+}
+
+// Batch operations for performance
+export async function insertResultsBatch(inputs: CreateResultInput[]): Promise<string[]> {
+  const stmt = db.getDatabase().prepare(`
+    INSERT INTO results (
+      result_id, run_id, case_id, model_id, processing_functions_hash,
+      succeeded, error_enum, num_edits, num_lines_deleted, num_lines_added,
+      time_to_first_token_ms, time_to_first_edit_ms, time_round_trip_ms,
+      cost_usd, completion_tokens, raw_model_output, file_edited_hash,
+      parsed_tool_call_json
+    ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+  `);
+  
+  return db.transaction(() => {
+    const resultIds: string[] = [];
+    
+    for (const input of inputs) {
+      const resultId = DatabaseClient.generateId();
+      
+      stmt.run(
+        resultId,
+        input.run_id,
+        input.case_id,
+        input.model_id,
+        input.processing_functions_hash,
+        input.succeeded ? 1 : 0, // Convert boolean to integer
+        input.error_enum || null,
+        input.num_edits || null,
+        input.num_lines_deleted || null,
+        input.num_lines_added || null,
+        input.time_to_first_token_ms || null,
+        input.time_to_first_edit_ms || null,
+        input.time_round_trip_ms || null,
+        input.cost_usd || null,
+        input.completion_tokens || null,
+        input.raw_model_output || null,
+        input.file_edited_hash || null,
+        input.parsed_tool_call_json || null
+      );
+      
+      resultIds.push(resultId);
+    }
+    
+    return resultIds;
+  });
+}
+
+export async function createCasesBatch(inputs: CreateCaseInput[]): Promise<string[]> {
+  const stmt = db.getDatabase().prepare(`
+    INSERT INTO cases (case_id, run_id, description, system_prompt_hash, task_id, tokens_in_context)
+    VALUES (?, ?, ?, ?, ?, ?)
+  `);
+  
+  return db.transaction(() => {
+    const caseIds: string[] = [];
+    
+    for (const input of inputs) {
+      const caseId = DatabaseClient.generateId();
+      
+      stmt.run(
+        caseId,
+        input.run_id,
+        input.description,
+        input.system_prompt_hash,
+        input.task_id,
+        input.tokens_in_context
+      );
+      
+      caseIds.push(caseId);
+    }
+    
+    return caseIds;
+  });
+}
+
+// Utility functions
+export async function getRunStats(runId: string): Promise<{
+  total_cases: number;
+  total_results: number;
+  success_rate: number;
+  avg_cost: number;
+  avg_latency: number;
+}> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      COUNT(DISTINCT c.case_id) as total_cases,
+      COUNT(r.result_id) as total_results,
+      AVG(CASE WHEN r.succeeded THEN 1.0 ELSE 0.0 END) as success_rate,
+      AVG(r.cost_usd) as avg_cost,
+      AVG(r.time_round_trip_ms) as avg_latency
+    FROM cases c
+    LEFT JOIN results r ON c.case_id = r.case_id
+    WHERE c.run_id = ?
+  `);
+  
+  const result = stmt.get(runId) as any;
+  return {
+    total_cases: result.total_cases || 0,
+    total_results: result.total_results || 0,
+    success_rate: result.success_rate || 0,
+    avg_cost: result.avg_cost || 0,
+    avg_latency: result.avg_latency || 0
+  };
+}
+
+// Count valid attempts for a specific case and model
+export async function getValidAttemptCount(caseId: string, modelId: string): Promise<number> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT COUNT(*) as count
+    FROM results 
+    WHERE case_id = ? 
+      AND model_id = ?
+      AND error_enum NOT IN (1, 6, 7)  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+  `);
+  
+  const result = stmt.get(caseId, modelId) as { count: number };
+  return result.count;
+}
+
+// Get valid results for a specific case and model (for analysis)
+export async function getValidResults(caseId: string, modelId: string, limit?: number): Promise<Result[]> {
+  const limitClause = limit ? `LIMIT ${limit}` : '';
+  const stmt = db.getDatabase().prepare(`
+    SELECT * FROM results 
+    WHERE case_id = ? 
+      AND model_id = ?
+      AND error_enum NOT IN (1, 6, 7)  -- Only valid attempts
+    ORDER BY created_at
+    ${limitClause}
+  `);
+  
+  return stmt.all(caseId, modelId) as Result[];
+}

evals/diff-edits/database/queries.ts

@@ -0,0 +1,309 @@
+import { DatabaseClient } from './client';
+import {
+  ModelSuccessRate,
+  ModelLatency,
+  CostAnalysis,
+  ErrorDistribution,
+  FailedCase,
+  PerformanceTrend,
+  ModelComparison
+} from './types';
+
+const db = DatabaseClient.getInstance();
+
+// Performance analysis queries
+export async function getSuccessRatesByModel(): Promise<ModelSuccessRate[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      model_id,
+      COUNT(*) as total_runs,
+      SUM(CASE WHEN succeeded THEN 1 ELSE 0 END) as successful_runs,
+      ROUND(AVG(CASE WHEN succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate
+    FROM results
+    WHERE error_enum NOT IN (1, 6, 7) OR error_enum IS NULL  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY model_id
+    ORDER BY success_rate DESC, total_runs DESC
+  `);
+  
+  return stmt.all() as ModelSuccessRate[];
+}
+
+export async function getAverageLatencyByModel(): Promise<ModelLatency[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      model_id,
+      ROUND(AVG(time_to_first_token_ms), 2) as avg_time_to_first_token_ms,
+      ROUND(AVG(time_to_first_edit_ms), 2) as avg_time_to_first_edit_ms,
+      ROUND(AVG(time_round_trip_ms), 2) as avg_time_round_trip_ms
+    FROM results
+    WHERE time_to_first_token_ms IS NOT NULL
+    GROUP BY model_id
+    ORDER BY avg_time_round_trip_ms ASC
+  `);
+  
+  return stmt.all() as ModelLatency[];
+}
+
+export async function getCostAnalysisByRun(): Promise<CostAnalysis[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      run_id,
+      model_id,
+      ROUND(SUM(cost_usd), 4) as total_cost_usd,
+      ROUND(AVG(cost_usd), 4) as avg_cost_per_case,
+      SUM(completion_tokens) as total_completion_tokens
+    FROM results
+    WHERE cost_usd IS NOT NULL
+    GROUP BY run_id, model_id
+    ORDER BY total_cost_usd DESC
+  `);
+  
+  return stmt.all() as CostAnalysis[];
+}
+
+// Error analysis queries
+export async function getErrorDistribution(): Promise<ErrorDistribution[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      error_enum,
+      COUNT(*) as count,
+      ROUND(COUNT(*) * 100.0 / (SELECT COUNT(*) FROM results WHERE succeeded = 0), 2) as percentage
+    FROM results
+    WHERE succeeded = 0 AND error_enum IS NOT NULL
+    GROUP BY error_enum
+    ORDER BY count DESC
+  `);
+  
+  return stmt.all() as ErrorDistribution[];
+}
+
+export async function getFailedCasesByError(errorEnum?: number): Promise<FailedCase[]> {
+  let query = `
+    SELECT 
+      r.case_id,
+      r.model_id,
+      r.error_enum,
+      c.description,
+      r.raw_model_output
+    FROM results r
+    JOIN cases c ON r.case_id = c.case_id
+    WHERE r.succeeded = 0
+  `;
+  
+  const params: any[] = [];
+  if (errorEnum !== undefined) {
+    query += ` AND r.error_enum = ?`;
+    params.push(errorEnum);
+  }
+  
+  query += ` ORDER BY r.created_at DESC LIMIT 100`;
+  
+  const stmt = db.getDatabase().prepare(query);
+  return stmt.all(...params) as FailedCase[];
+}
+
+// Trend analysis queries
+export async function getPerformanceTrends(days: number = 30): Promise<PerformanceTrend[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      DATE(r.created_at) as date,
+      r.model_id,
+      ROUND(AVG(CASE WHEN r.succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(r.time_round_trip_ms), 2) as avg_latency_ms,
+      ROUND(AVG(r.cost_usd), 4) as avg_cost_usd
+    FROM results r
+    WHERE r.created_at >= datetime('now', '-' || ? || ' days')
+      AND (r.error_enum NOT IN (1, 6, 7) OR r.error_enum IS NULL)  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY DATE(r.created_at), r.model_id
+    ORDER BY date DESC, model_id
+  `);
+  
+  return stmt.all(days) as PerformanceTrend[];
+}
+
+export async function getModelComparisons(): Promise<ModelComparison[]> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      model_id,
+      ROUND(AVG(CASE WHEN succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(time_round_trip_ms), 2) as avg_latency_ms,
+      ROUND(AVG(cost_usd), 4) as avg_cost_usd,
+      COUNT(*) as total_runs
+    FROM results
+    WHERE error_enum NOT IN (1, 6, 7) OR error_enum IS NULL  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY model_id
+    HAVING total_runs >= 10
+    ORDER BY success_rate DESC, avg_latency_ms ASC
+  `);
+  
+  return stmt.all() as ModelComparison[];
+}
+
+// Advanced analysis queries
+export async function getTopPerformingCases(limit: number = 10): Promise<Array<{
+  case_id: string;
+  description: string;
+  success_rate: number;
+  avg_latency_ms: number;
+  total_runs: number;
+}>> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      c.case_id,
+      c.description,
+      ROUND(AVG(CASE WHEN r.succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(r.time_round_trip_ms), 2) as avg_latency_ms,
+      COUNT(r.result_id) as total_runs
+    FROM cases c
+    JOIN results r ON c.case_id = r.case_id
+    WHERE r.error_enum NOT IN (1, 6, 7) OR r.error_enum IS NULL  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY c.case_id, c.description
+    HAVING total_runs >= 5
+    ORDER BY success_rate DESC, avg_latency_ms ASC
+    LIMIT ?
+  `);
+  
+  return stmt.all(limit) as Array<{
+    case_id: string;
+    description: string;
+    success_rate: number;
+    avg_latency_ms: number;
+    total_runs: number;
+  }>;
+}
+
+export async function getWorstPerformingCases(limit: number = 10): Promise<Array<{
+  case_id: string;
+  description: string;
+  success_rate: number;
+  avg_latency_ms: number;
+  total_runs: number;
+}>> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      c.case_id,
+      c.description,
+      ROUND(AVG(CASE WHEN r.succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(r.time_round_trip_ms), 2) as avg_latency_ms,
+      COUNT(r.result_id) as total_runs
+    FROM cases c
+    JOIN results r ON c.case_id = r.case_id
+    WHERE r.error_enum NOT IN (1, 6, 7) OR r.error_enum IS NULL  -- Exclude: no_tool_calls, wrong_tool_call, wrong_file_edited
+    GROUP BY c.case_id, c.description
+    HAVING total_runs >= 5
+    ORDER BY success_rate ASC, avg_latency_ms DESC
+    LIMIT ?
+  `);
+  
+  return stmt.all(limit) as Array<{
+    case_id: string;
+    description: string;
+    success_rate: number;
+    avg_latency_ms: number;
+    total_runs: number;
+  }>;
+}
+
+export async function getModelPerformanceByTimeOfDay(): Promise<Array<{
+  model_id: string;
+  hour: number;
+  success_rate: number;
+  avg_latency_ms: number;
+  total_runs: number;
+}>> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      model_id,
+      CAST(strftime('%H', created_at) AS INTEGER) as hour,
+      ROUND(AVG(CASE WHEN succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(time_round_trip_ms), 2) as avg_latency_ms,
+      COUNT(*) as total_runs
+    FROM results
+    GROUP BY model_id, hour
+    HAVING total_runs >= 5
+    ORDER BY model_id, hour
+  `);
+  
+  return stmt.all() as Array<{
+    model_id: string;
+    hour: number;
+    success_rate: number;
+    avg_latency_ms: number;
+    total_runs: number;
+  }>;
+}
+
+export async function getRunComparison(runId1: string, runId2: string): Promise<{
+  run1: { run_id: string; success_rate: number; avg_latency_ms: number; avg_cost_usd: number; total_cases: number };
+  run2: { run_id: string; success_rate: number; avg_latency_ms: number; avg_cost_usd: number; total_cases: number };
+}> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      run_id,
+      ROUND(AVG(CASE WHEN succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) as success_rate,
+      ROUND(AVG(time_round_trip_ms), 2) as avg_latency_ms,
+      ROUND(AVG(cost_usd), 4) as avg_cost_usd,
+      COUNT(DISTINCT case_id) as total_cases
+    FROM results
+    WHERE run_id IN (?, ?)
+    GROUP BY run_id
+  `);
+  
+  const results = stmt.all(runId1, runId2) as Array<{
+    run_id: string;
+    success_rate: number;
+    avg_latency_ms: number;
+    avg_cost_usd: number;
+    total_cases: number;
+  }>;
+  
+  const run1 = results.find(r => r.run_id === runId1);
+  const run2 = results.find(r => r.run_id === runId2);
+  
+  if (!run1 || !run2) {
+    throw new Error('One or both runs not found');
+  }
+  
+  return { run1, run2 };
+}
+
+// Summary statistics
+export async function getDatabaseSummary(): Promise<{
+  total_runs: number;
+  total_cases: number;
+  total_results: number;
+  valid_results: number;
+  unique_models: number;
+  overall_success_rate: number;
+  date_range: { earliest: string; latest: string };
+}> {
+  const stmt = db.getDatabase().prepare(`
+    SELECT 
+      (SELECT COUNT(*) FROM runs) as total_runs,
+      (SELECT COUNT(*) FROM cases) as total_cases,
+      (SELECT COUNT(*) FROM results) as total_results,
+      (SELECT COUNT(*) FROM results WHERE error_enum NOT IN (1, 6, 7) OR error_enum IS NULL) as valid_results,
+      (SELECT COUNT(DISTINCT model_id) FROM results) as unique_models,
+      (SELECT ROUND(AVG(CASE WHEN succeeded THEN 1.0 ELSE 0.0 END) * 100, 2) 
+       FROM results 
+       WHERE error_enum NOT IN (1, 6, 7) OR error_enum IS NULL) as overall_success_rate,
+      (SELECT MIN(created_at) FROM results) as earliest,
+      (SELECT MAX(created_at) FROM results) as latest
+    FROM results
+    LIMIT 1
+  `);
+  
+  const result = stmt.get() as any;
+  return {
+    total_runs: result.total_runs || 0,
+    total_cases: result.total_cases || 0,
+    total_results: result.total_results || 0,
+    valid_results: result.valid_results || 0,
+    unique_models: result.unique_models || 0,
+    overall_success_rate: result.overall_success_rate || 0,
+    date_range: {
+      earliest: result.earliest || '',
+      latest: result.latest || ''
+    }
+  };
+}

evals/diff-edits/database/schema.sql

@@ -0,0 +1,78 @@
+PRAGMA foreign_keys = ON;
+
+CREATE TABLE system_prompts (
+    hash TEXT PRIMARY KEY,
+    name TEXT NOT NULL,
+    content TEXT NOT NULL,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP
+);
+
+CREATE TABLE processing_functions (
+    hash TEXT PRIMARY KEY,
+    name TEXT NOT NULL,
+    parsing_function TEXT NOT NULL,
+    diff_edit_function TEXT NOT NULL,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP
+);
+
+CREATE TABLE files (
+    hash TEXT PRIMARY KEY,
+    filepath TEXT NOT NULL,
+    content TEXT NOT NULL,
+    tokens INTEGER,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP
+);
+
+CREATE TABLE runs (
+    run_id TEXT PRIMARY KEY,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP,
+    description TEXT,
+    system_prompt_hash TEXT NOT NULL,
+    FOREIGN KEY (system_prompt_hash) REFERENCES system_prompts(hash)
+);
+
+CREATE TABLE cases (
+    case_id TEXT PRIMARY KEY,
+    run_id TEXT NOT NULL,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP,
+    description TEXT NOT NULL,
+    system_prompt_hash TEXT NOT NULL,
+    task_id TEXT NOT NULL,
+    tokens_in_context INTEGER,
+    file_hash TEXT,
+    FOREIGN KEY (run_id) REFERENCES runs(run_id),
+    FOREIGN KEY (system_prompt_hash) REFERENCES system_prompts(hash),
+    FOREIGN KEY (file_hash) REFERENCES files(hash)
+);
+
+CREATE TABLE results (
+    result_id TEXT PRIMARY KEY,
+    run_id TEXT NOT NULL,
+    case_id TEXT NOT NULL,
+    model_id TEXT NOT NULL,
+    processing_functions_hash TEXT NOT NULL,
+    succeeded BOOLEAN NOT NULL,
+    error_enum INTEGER,
+    num_edits INTEGER,
+    num_lines_deleted INTEGER,
+    num_lines_added INTEGER,
+    time_to_first_token_ms INTEGER,
+    time_to_first_edit_ms INTEGER,
+    time_round_trip_ms INTEGER,
+    cost_usd REAL,
+    completion_tokens INTEGER,
+    raw_model_output TEXT,
+    file_edited_hash TEXT,
+    parsed_tool_call_json TEXT,
+    created_at DATETIME DEFAULT CURRENT_TIMESTAMP,
+    FOREIGN KEY (run_id) REFERENCES runs(run_id),
+    FOREIGN KEY (case_id) REFERENCES cases(case_id),
+    FOREIGN KEY (processing_functions_hash) REFERENCES processing_functions(hash)
+);
+
+CREATE INDEX idx_results_run_model ON results(run_id, model_id);
+CREATE INDEX idx_results_case_model ON results(case_id, model_id);
+CREATE INDEX idx_results_success ON results(succeeded);
+CREATE INDEX idx_cases_run ON cases(run_id);
+CREATE INDEX idx_results_created_at ON results(created_at);
+CREATE INDEX idx_runs_created_at ON runs(created_at);

evals/diff-edits/database/test.ts

@@ -0,0 +1,53 @@
+// Simple test to verify database functionality
+import { getDatabase } from './client';
+import { upsertSystemPrompt, createBenchmarkRun, getDatabaseSummary } from './index';
+
+async function testDatabase() {
+  console.log('Testing database functionality...');
+  
+  try {
+    // Test database connection
+    const db = getDatabase();
+    console.log('✓ Database connection established');
+    console.log('Database path:', db.getDatabasePath());
+    
+    // Test database info
+    const info = db.getInfo();
+    console.log('✓ Database info:', info);
+    
+    // Test database stats
+    const stats = db.getStats();
+    console.log('✓ Database stats:', stats);
+    
+    // Test system prompt creation
+    const systemPromptHash = await upsertSystemPrompt({
+      name: 'test-prompt',
+      content: 'This is a test system prompt for database verification.'
+    });
+    console.log('✓ System prompt created with hash:', systemPromptHash);
+    
+    // Test benchmark run creation
+    const runId = await createBenchmarkRun({
+      description: 'Test run for database verification',
+      system_prompt_hash: systemPromptHash
+    });
+    console.log('✓ Benchmark run created with ID:', runId);
+    
+    // Test database summary
+    const summary = await getDatabaseSummary();
+    console.log('✓ Database summary:', summary);
+    
+    console.log('\n🎉 All database tests passed!');
+    
+  } catch (error) {
+    console.error('❌ Database test failed:', error);
+    process.exit(1);
+  }
+}
+
+// Run test if this file is executed directly
+if (require.main === module) {
+  testDatabase();
+}
+
+export { testDatabase };

evals/diff-edits/database/types.ts

@@ -0,0 +1,169 @@
+// Database type definitions for diff-edits evaluation system
+
+export interface SystemPrompt {
+  hash: string;
+  name: string;
+  content: string;
+  created_at: string;
+}
+
+export interface ProcessingFunctions {
+  hash: string;
+  name: string;
+  parsing_function: string;
+  diff_edit_function: string;
+  created_at: string;
+}
+
+export interface FileRecord {
+  hash: string;
+  filepath: string;
+  content: string;
+  tokens?: number;
+  created_at: string;
+}
+
+export interface BenchmarkRun {
+  run_id: string;
+  created_at: string;
+  description?: string;
+  system_prompt_hash: string;
+}
+
+export interface Case {
+	case_id: string
+	run_id: string
+	created_at: string
+	description: string
+	system_prompt_hash: string
+	task_id: string
+	tokens_in_context: number
+	file_hash?: string
+}
+
+export interface Result {
+  result_id: string;
+  run_id: string;
+  case_id: string;
+  model_id: string;
+  processing_functions_hash: string;
+  succeeded: boolean;
+  error_enum?: number;
+  num_edits?: number;
+  num_lines_deleted?: number;
+  num_lines_added?: number;
+  time_to_first_token_ms?: number;
+  time_to_first_edit_ms?: number;
+  time_round_trip_ms?: number;
+  cost_usd?: number;
+  completion_tokens?: number;
+  raw_model_output?: string;
+  file_edited_hash?: string;
+  parsed_tool_call_json?: string;
+  created_at: string;
+}
+
+// Input types for creating records
+export interface CreateSystemPromptInput {
+  name: string;
+  content: string;
+}
+
+export interface CreateProcessingFunctionsInput {
+  name: string;
+  parsing_function: string;
+  diff_edit_function: string;
+}
+
+export interface CreateFileInput {
+  filepath: string;
+  content: string;
+  tokens?: number;
+}
+
+export interface CreateBenchmarkRunInput {
+  description?: string;
+  system_prompt_hash: string;
+}
+
+export interface CreateCaseInput {
+  run_id: string;
+  description: string;
+  system_prompt_hash: string;
+  task_id: string;
+  tokens_in_context: number;
+  file_hash?: string;
+}
+
+export interface CreateResultInput {
+  run_id: string;
+  case_id: string;
+  model_id: string;
+  processing_functions_hash: string;
+  succeeded: boolean;
+  error_enum?: number;
+  num_edits?: number;
+  num_lines_deleted?: number;
+  num_lines_added?: number;
+  time_to_first_token_ms?: number;
+  time_to_first_edit_ms?: number;
+  time_round_trip_ms?: number;
+  cost_usd?: number;
+  completion_tokens?: number;
+  raw_model_output?: string;
+  file_edited_hash?: string;
+  parsed_tool_call_json?: string;
+}
+
+// Analysis result types
+export interface ModelSuccessRate {
+  model_id: string;
+  total_runs: number;
+  successful_runs: number;
+  success_rate: number;
+}
+
+export interface ModelLatency {
+  model_id: string;
+  avg_time_to_first_token_ms: number;
+  avg_time_to_first_edit_ms: number;
+  avg_time_round_trip_ms: number;
+}
+
+export interface CostAnalysis {
+  run_id: string;
+  model_id: string;
+  total_cost_usd: number;
+  avg_cost_per_case: number;
+  total_completion_tokens: number;
+}
+
+export interface ErrorDistribution {
+  error_enum: number;
+  count: number;
+  percentage: number;
+}
+
+export interface FailedCase {
+  case_id: string;
+  model_id: string;
+  error_enum: number;
+  description: string;
+  raw_model_output?: string;
+}
+
+export interface PerformanceTrend {
+  date: string;
+  model_id: string;
+  success_rate: number;
+  avg_latency_ms: number;
+  avg_cost_usd: number;
+}
+
+export interface ModelComparison {
+  model_id: string;
+  success_rate: number;
+  avg_latency_ms: number;
+  avg_cost_usd: number;
+  total_runs: number;
+}

evals/diff-edits/openRouterModelsHelper.ts

@@ -0,0 +1,98 @@
+import axios from "axios";
+import path from "path";
+import fs from "fs/promises";
+
+// Minimal type for what we need from OpenRouter model info in evals
+export interface EvalOpenRouterModelInfo {
+    id: string;
+    contextWindow: number;
+    inputPrice?: number; // Price per million tokens
+    outputPrice?: number; // Price per million tokens
+    // Add any other fields if they become necessary for evals
+}
+
+function logHelper(isVerbose: boolean, message: string) {
+    if (isVerbose) {
+        console.log(`[OpenRouterModelsHelper] ${message}`);
+    }
+}
+
+/**
+ * Ensures the cache directory exists within evals and returns its path
+ */
+async function ensureEvalCacheDirectoryExists(): Promise<string> {
+    // Cache directory within evals, e.g., evals/.cache/
+    const cacheDir = path.join(__dirname, "..", ".cache");
+    await fs.mkdir(cacheDir, { recursive: true });
+    return cacheDir;
+}
+
+/**
+ * Fetches, parses, and caches OpenRouter model data.
+ * Tries to read from a local cache first.
+ * @param isVerbose Enable verbose logging
+ * @returns A record of model IDs to their info.
+ */
+export async function loadOpenRouterModelData(isVerbose: boolean = false): Promise<Record<string, EvalOpenRouterModelInfo>> {
+    const cacheDir = await ensureEvalCacheDirectoryExists();
+    const cacheFilePath = path.join(cacheDir, "openRouterModels.json");
+    let models: Record<string, EvalOpenRouterModelInfo> = {};
+
+    try {
+        const stats = await fs.stat(cacheFilePath).catch(() => null);
+        // Use cache if less than 24 hours old
+        if (stats && (Date.now() - stats.mtimeMs < 24 * 60 * 60 * 1000)) {
+            logHelper(isVerbose, "Using cached OpenRouter model data.");
+            const fileContents = await fs.readFile(cacheFilePath, "utf8");
+            models = JSON.parse(fileContents);
+            if (Object.keys(models).length > 0) {
+                return models;
+            }
+            logHelper(isVerbose, "Cache was empty or invalid, fetching fresh data.");
+        } else if (stats) {
+            logHelper(isVerbose, "Cached OpenRouter model data is stale, fetching fresh data.");
+        } else {
+            logHelper(isVerbose, "No cached OpenRouter model data found, fetching fresh data.");
+        }
+    } catch (e) {
+        logHelper(isVerbose, `Error accessing cache, fetching fresh data: ${e}`);
+    }
+
+    try {
+        const response = await axios.get("https://openrouter.ai/api/v1/models");
+        if (response.data?.data) {
+            const rawModels = response.data.data;
+            const parsedModels: Record<string, EvalOpenRouterModelInfo> = {};
+            const parsePrice = (price: any) => price ? parseFloat(price) * 1_000_000 : undefined;
+
+            for (const rawModel of rawModels) {
+                parsedModels[rawModel.id] = {
+                    id: rawModel.id,
+                    contextWindow: rawModel.context_length ?? 0,
+                    inputPrice: parsePrice(rawModel.pricing?.prompt),
+                    outputPrice: parsePrice(rawModel.pricing?.completion),
+                };
+            }
+            await fs.writeFile(cacheFilePath, JSON.stringify(parsedModels, null, 2));
+            logHelper(isVerbose, `Fetched and cached ${Object.keys(parsedModels).length} OpenRouter models.`);
+            return parsedModels;
+        } else {
+            logHelper(isVerbose, "Invalid response structure from OpenRouter API.");
+        }
+    } catch (error) {
+        logHelper(isVerbose, `Error fetching OpenRouter models: ${error}. Attempting to use stale cache if available.`);
+        // Attempt to read stale cache as a last resort if fetching failed
+        try {
+            const fileContents = await fs.readFile(cacheFilePath, "utf8");
+            models = JSON.parse(fileContents);
+            if (Object.keys(models).length > 0) {
+                logHelper(isVerbose, "Successfully loaded stale cache after fetch failure.");
+                return models;
+            }
+        } catch (cacheError) {
+            logHelper(isVerbose, `Failed to read stale cache: ${cacheError}. Proceeding without OpenRouter model data.`);
+        }
+    }
+    // Return empty if all attempts fail, so the caller can decide how to handle it
+    return {};
+}

evals/diff-edits/run_and_open_dashboard.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+
+# Get the directory of this script to make paths robust
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+# The 'evals' directory is the parent of the script's directory
+EVALS_DIR=$(dirname "$SCRIPT_DIR")
+
+# Navigate to the evals directory to ensure npm commands run correctly
+cd "$EVALS_DIR"
+
+# Re-install dependencies and build the CLI
+echo "Ensuring dependencies are up to date and building CLI..."
+npm install && npm run build:cli
+
+# Check if the build was successful before proceeding
+if [ $? -ne 0 ]; then
+    echo "CLI build failed. Aborting evaluation."
+    exit 1
+fi
+
+# Run the evaluation script, passing all arguments from the command line
+echo "Running evaluation..."
+node ./cli/dist/index.js run-diff-eval "$@"
+
+# Check the exit code of the evaluation script
+if [ $? -eq 0 ]; then
+  # If the script succeeded, open the dashboard in the background
+  echo "Evaluation complete. Starting dashboard..."
+  (cd "$SCRIPT_DIR/dashboard" && streamlit run app.py &)
+else
+  # If the script failed, print an error message and exit
+  echo "Evaluation failed. Dashboard will not be started."
+  exit 1
+fi

evals/diff-edits/types.ts

@@ -61,7 +61,22 @@ export type ConstructSystemPromptFn = (
 
 export interface TestResult {
 	success: boolean
-	streamResult?: any
+	streamResult?: {
+		assistantMessage: string
+		reasoningMessage: string
+		usage: {
+			inputTokens: number
+			outputTokens: number
+			cacheWriteTokens: number
+			cacheReadTokens: number
+			totalCost: number
+		}
+		timing?: {
+			timeToFirstTokenMs: number
+			timeToFirstEditMs?: number
+			totalRoundTripMs: number
+		}
+	}
 	diffEdit?: string
 	toolCalls?: ExtractedToolCall[]
 	diffEditSuccess?: boolean

evals/diff_editing/TestRunner.ts

@@ -1,391 +0,0 @@
-import { runSingleEvaluation, TestInput, TestResult } from "./ClineWrapper"
-import { basicSystemPrompt } from "./prompts/basicSystemPrompt-06-06-25"
-import { claude4SystemPrompt } from "./prompts/claude4SystemPrompt-06-06-25"
-import { formatResponse } from "./helpers"
-import { Anthropic } from "@anthropic-ai/sdk"
-import * as fs from "fs"
-import * as path from "path"
-import { Command } from "commander"
-import { InputMessage, ProcessedTestCase, TestCase, TestConfig, SystemPromptDetails, ConstructSystemPromptFn } from "./types"
-
-function log(isVerbose: boolean, message: string) {
-	if (isVerbose) {
-		console.log(message)
-	}
-}
-
-const systemPromptGeneratorLookup: Record<string, ConstructSystemPromptFn> = {
-	basicSystemPrompt: basicSystemPrompt,
-	claude4SystemPrompt: claude4SystemPrompt,
-}
-
-type TestResultSet = { [test_id: string]: (TestResult & { test_id?: string })[] }
-
-class NodeTestRunner {
-	private apiKey: string | undefined
-
-	constructor(isReplay: boolean) {
-		if (!isReplay) {
-			this.apiKey = process.env.OPENROUTER_API_KEY
-			if (!this.apiKey) {
-				throw new Error("OPENROUTER_API_KEY environment variable not set for a non-replay run.")
-			}
-		}
-	}
-
-	/**
-	 * convert our messages array into a properly formatted Anthropic messages array
-	 */
-	transformMessages(messages: InputMessage[]): Anthropic.Messages.MessageParam[] {
-		return messages.map((msg) => {
-			// Use TextBlockParam here for constructing the input message
-			const content: (Anthropic.TextBlockParam | Anthropic.ImageBlockParam)[] = []
-
-			if (msg.text) {
-				// This object now correctly matches the TextBlockParam type
-				content.push({ type: "text", text: msg.text })
-			}
-
-			if (msg.images && Array.isArray(msg.images)) {
-				const imageBlocks = formatResponse.imageBlocks(msg.images)
-				content.push(...imageBlocks)
-			}
-
-			return {
-				role: msg.role,
-				content: content,
-			}
-		})
-	}
-
-	/**
-	 * Generate the system prompt on the fly
-	 */
-	constructSystemPrompt(systemPromptDetails: SystemPromptDetails, systemPromptName: string) {
-		const systemPromptGenerator = systemPromptGeneratorLookup[systemPromptName]
-
-		const { cwd_value, browser_use, width, height, os_value, shell_value, home_value, mcp_string, user_custom_instructions } =
-			systemPromptDetails
-
-		const systemPrompt = systemPromptGenerator(
-			cwd_value,
-			browser_use,
-			width,
-			height,
-			os_value,
-			shell_value,
-			home_value,
-			mcp_string,
-			user_custom_instructions,
-		)
-
-		return systemPrompt
-	}
-
-	/**
-	 * Loads our test cases from a directory of json files
-	 */
-	loadTestCases(testDirectoryPath: string): TestCase[] {
-		const testCasesArray: TestCase[] = []
-		const dirents = fs.readdirSync(testDirectoryPath, { withFileTypes: true })
-
-		for (const dirent of dirents) {
-			if (dirent.isFile() && dirent.name.endsWith(".json")) {
-				const testFilePath = path.join(testDirectoryPath, dirent.name)
-				const fileContent = fs.readFileSync(testFilePath, "utf8")
-				const testCase: TestCase = JSON.parse(fileContent)
-
-				// Use the filename (without extension) as the test_id if not provided
-				if (!testCase.test_id) {
-					testCase.test_id = path.parse(dirent.name).name
-				}
-				testCasesArray.push(testCase)
-			}
-		}
-
-		return testCasesArray
-	}
-
-	/**
-	 * Saves the test results to the specified output directory.
-	 */
-	saveTestResults(results: TestResultSet, outputPath: string) {
-		// Ensure output directory exists
-		if (!fs.existsSync(outputPath)) {
-			fs.mkdirSync(outputPath, { recursive: true })
-		}
-
-		// Write each test result to its own file
-		for (const testId in results) {
-			const outputFilePath = path.join(outputPath, `${testId}.json`)
-			const testResult = results[testId]
-			fs.writeFileSync(outputFilePath, JSON.stringify(testResult, null, 2))
-		}
-	}
-
-	/**
-	 * Run a single test example
-	 */
-	async runSingleTest(testCase: ProcessedTestCase, testConfig: TestConfig): Promise<TestResult> {
-		if (testConfig.replay && !testCase.original_diff_edit_tool_call_message) {
-			return {
-				success: false,
-				error: "missing_original_diff_edit_tool_call_message",
-				errorString: `Test case ${testCase.test_id} is missing 'original_diff_edit_tool_call_message' for replay.`,
-			}
-		}
-
-		const customSystemPrompt = this.constructSystemPrompt(testCase.system_prompt_details, testConfig.system_prompt_name)
-
-		// messages don't include system prompt and are everything up to the first replace_in_file tool call which results in a diff edit error
-		const input: TestInput = {
-			apiKey: this.apiKey,
-			systemPrompt: customSystemPrompt,
-			messages: testCase.messages,
-			modelId: testConfig.model_id,
-			originalFile: testCase.file_contents,
-			originalFilePath: testCase.file_path,
-			parsingFunction: testConfig.parsing_function,
-			diffEditFunction: testConfig.diff_edit_function,
-			thinkingBudgetTokens: testConfig.thinking_tokens_budget,
-			originalDiffEditToolCallMessage: testConfig.replay ? testCase.original_diff_edit_tool_call_message : undefined,
-		}
-
-		return await runSingleEvaluation(input)
-	}
-
-	/**
-	 * Runs all the text examples synchonously
-	 */
-	async runAllTests(testCases: ProcessedTestCase[], testConfig: TestConfig, isVerbose: boolean): Promise<TestResultSet> {
-		const results: TestResultSet = {}
-
-		for (const testCase of testCases) {
-			results[testCase.test_id] = []
-
-			log(isVerbose, `-Running test: ${testCase.test_id}`)
-			for (let i = 0; i < testConfig.number_of_runs; i++) {
-				const result = await this.runSingleTest(testCase, testConfig)
-				results[testCase.test_id].push(result)
-			}
-		}
-		return results
-	}
-
-	/**
-	 * Runs all of the text examples asynchronously, with concurrency limit
-	 */
-	async runAllTestsParallel(
-		testCases: ProcessedTestCase[],
-		testConfig: TestConfig,
-		isVerbose: boolean,
-		maxConcurrency: number = 20,
-	): Promise<TestResultSet> {
-		const results: TestResultSet = {}
-		testCases.forEach((tc) => {
-			results[tc.test_id] = []
-		})
-
-		// Create a flat list of all individual runs we need to execute
-		const allRuns = testCases.flatMap((testCase) =>
-			Array(testConfig.number_of_runs)
-				.fill(null)
-				.map(() => testCase),
-		)
-
-		for (let i = 0; i < allRuns.length; i += maxConcurrency) {
-			const batch = allRuns.slice(i, i + maxConcurrency)
-
-			const batchPromises = batch.map((testCase) =>
-				this.runSingleTest(testCase, testConfig).then((result) => ({
-					...result,
-					test_id: testCase.test_id,
-				})),
-			)
-
-			const batchResults = await Promise.all(batchPromises)
-
-			// Calculate the total cost for this batch
-			const batchCost = batchResults.reduce((total, result) => {
-				return total + (result.streamResult?.usage?.totalCost || 0)
-			}, 0)
-
-			// Populate the results dictionary
-			for (const result of batchResults) {
-				if (result.test_id) {
-					results[result.test_id].push(result)
-				}
-			}
-
-			const batchNumber = i / maxConcurrency + 1
-			const totalBatches = Math.ceil(allRuns.length / maxConcurrency)
-			log(isVerbose, `-Completed batch ${batchNumber} of ${totalBatches}... (Batch Cost: $${batchCost.toFixed(6)})`)
-		}
-
-		return results
-	}
-
-	/**
-	 * Print output of the tests
-	 */
-	printSummary(results: TestResultSet, isVerbose: boolean) {
-		let totalRuns = 0
-		let totalPasses = 0
-		let totalInputTokens = 0
-		let totalOutputTokens = 0
-		let totalCost = 0
-		let runsWithUsageData = 0
-		let totalDiffEditSuccesses = 0
-		let totalRunsWithToolCalls = 0
-		const testCaseIds = Object.keys(results)
-
-		log(isVerbose, "\n=== TEST SUMMARY ===")
-
-		for (const testId of testCaseIds) {
-			const testResults = results[testId]
-			const passedCount = testResults.filter((r) => r.success && r.diffEditSuccess).length
-			const runCount = testResults.length
-
-			totalRuns += runCount
-			totalPasses += passedCount
-
-			const runsWithToolCalls = testResults.filter((r) => r.success === true).length
-			const diffEditSuccesses = passedCount
-			totalRunsWithToolCalls += runsWithToolCalls
-			totalDiffEditSuccesses += diffEditSuccesses
-
-			// Accumulate token and cost data
-			for (const result of testResults) {
-				if (result.streamResult?.usage) {
-					totalInputTokens += result.streamResult.usage.inputTokens
-					totalOutputTokens += result.streamResult.usage.outputTokens
-					totalCost += result.streamResult.usage.totalCost
-					runsWithUsageData++
-				}
-			}
-
-			log(isVerbose, `\n--- Test Case: ${testId} ---`)
-			log(isVerbose, `  Runs: ${runCount}`)
-			log(isVerbose, `  Passed: ${passedCount}`)
-			log(isVerbose, `  Success Rate: ${runCount > 0 ? ((passedCount / runCount) * 100).toFixed(1) : "N/A"}%`)
-		}
-
-		log(isVerbose, "\n\n=== OVERALL SUMMARY ===")
-		log(isVerbose, `Total Test Cases: ${testCaseIds.length}`)
-		log(isVerbose, `Total Runs Executed: ${totalRuns}`)
-		log(isVerbose, `Overall Passed: ${totalPasses}`)
-		log(isVerbose, `Overall Failed: ${totalRuns - totalPasses}`)
-		log(isVerbose, `Overall Success Rate: ${totalRuns > 0 ? ((totalPasses / totalRuns) * 100).toFixed(1) : "N/A"}%`)
-
-		log(isVerbose, "\n\n=== OVERALL DIFF EDIT SUCCESS RATE ===")
-		if (totalRunsWithToolCalls > 0) {
-			const diffSuccessRate = (totalDiffEditSuccesses / totalRunsWithToolCalls) * 100
-			log(isVerbose, `Total Runs with Successful Tool Calls: ${totalRunsWithToolCalls}`)
-			log(isVerbose, `Total Runs with Successful Diff Edits: ${totalDiffEditSuccesses}`)
-			log(isVerbose, `Diff Edit Success Rate: ${diffSuccessRate.toFixed(1)}%`)
-		} else {
-			log(isVerbose, "No successful tool calls to analyze for diff edit success.")
-		}
-
-		log(isVerbose, "\n\n=== TOKEN & COST ANALYSIS ===")
-		if (runsWithUsageData > 0) {
-			log(isVerbose, `Total Input Tokens: ${totalInputTokens.toLocaleString()}`)
-			log(isVerbose, `Total Output Tokens: ${totalOutputTokens.toLocaleString()}`)
-			log(isVerbose, `Total Cost: $${totalCost.toFixed(6)}`)
-			log(isVerbose, "---")
-			log(
-				isVerbose,
-				`Avg Input Tokens / Run: ${(totalInputTokens / runsWithUsageData).toLocaleString(undefined, {
-					maximumFractionDigits: 0,
-				})}`,
-			)
-			log(
-				isVerbose,
-				`Avg Output Tokens / Run: ${(totalOutputTokens / runsWithUsageData).toLocaleString(undefined, {
-					maximumFractionDigits: 0,
-				})}`,
-			)
-			log(isVerbose, `Avg Cost / Run: $${(totalCost / runsWithUsageData).toFixed(6)}`)
-		} else {
-			log(isVerbose, "No usage data available to analyze.")
-		}
-	}
-}
-
-async function main() {
-	const program = new Command()
-
-	const defaultTestPath = path.join(__dirname, "test_cases")
-	const defaultOutputPath = path.join(__dirname, "test_outputs")
-
-	program
-		.name("TestRunner")
-		.description("Run evaluation tests for diff editing")
-		.version("1.0.0")
-		.option("--test-path <path>", "Path to the directory containing test case JSON files", defaultTestPath)
-		.option("--output-path <path>", "Path to the directory to save the test output JSON files", defaultOutputPath)
-		.option("--model-id <model_id>", "The model ID to use for the test")
-		.option("--system-prompt-name <name>", "The name of the system prompt to use", "basicSystemPrompt")
-		.option("-n, --number-of-runs <number>", "Number of times to run each test case", "1")
-		.option("--parsing-function <name>", "The parsing function to use", "parseAssistantMessageV2")
-		.option("--diff-edit-function <name>", "The diff editing function to use", "constructNewFileContentV2")
-		.option("--thinking-budget <tokens>", "Set the thinking tokens budget", "0")
-		.option("--parallel", "Run tests in parallel", false)
-		.option("--replay", "Run evaluation from a pre-recorded LLM output, skipping the API call", false)
-		.option("-v, --verbose", "Enable verbose logging", false)
-
-	program.parse(process.argv)
-
-	const options = program.opts()
-	const isVerbose = options.verbose
-	const testPath = options.testPath
-	const outputPath = options.outputPath
-
-	const testConfig: TestConfig = {
-		model_id: options.modelId,
-		system_prompt_name: options.systemPromptName,
-		number_of_runs: parseInt(options.numberOfRuns, 10),
-		parsing_function: options.parsingFunction,
-		diff_edit_function: options.diffEditFunction,
-		thinking_tokens_budget: parseInt(options.thinkingBudget, 10),
-		replay: options.replay,
-	}
-
-	try {
-		const startTime = Date.now()
-
-		const runner = new NodeTestRunner(testConfig.replay)
-		const testCases = runner.loadTestCases(testPath)
-
-		const processedTestCases: ProcessedTestCase[] = testCases.map((tc) => ({
-			...tc,
-			messages: runner.transformMessages(tc.messages),
-		}))
-
-		log(isVerbose, `-Loaded ${testCases.length} test cases.`)
-		log(isVerbose, `-Executing ${testConfig.number_of_runs} run(s) per test case.`)
-		if (testConfig.replay) {
-			log(isVerbose, `-Running in REPLAY mode. No API calls will be made.`)
-		}
-		log(isVerbose, "Starting tests...\n")
-
-		const results = options.parallel
-			? await runner.runAllTestsParallel(processedTestCases, testConfig, isVerbose)
-			: await runner.runAllTests(processedTestCases, testConfig, isVerbose)
-
-		runner.printSummary(results, isVerbose)
-
-		const endTime = Date.now()
-		const durationSeconds = ((endTime - startTime) / 1000).toFixed(2)
-		log(isVerbose, `\n-Total execution time: ${durationSeconds} seconds`)
-
-		runner.saveTestResults(results, outputPath)
-	} catch (error) {
-		console.error("\nError running tests:", error)
-		process.exit(1)
-	}
-}
-
-if (require.main === module) {
-	main()
-}

evals/package.json

@@ -0,0 +1,44 @@
+{
+  "name": "cline-evals",
+  "version": "0.1.0",
+  "description": "Evaluation scripts and tools for Cline",
+  "main": "cli/dist/index.js",
+  "scripts": {
+    "build:cli": "cd cli && tsc",
+    "start:cli": "cd cli && node dist/index.js",
+    "dev:cli": "cd cli && ts-node src/index.ts",
+    "diff-eval": "./diff-edits/run_and_open_dashboard.sh",
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "keywords": [
+    "cline",
+    "evaluation",
+    "benchmark",
+    "diff-edits"
+  ],
+  "author": "",
+  "license": "MIT",
+  "dependencies": {
+    "axios": "^1.8.2",
+    "better-sqlite3": "^11.10.0",
+    "chalk": "^4.1.2",
+    "dotenv": "^16.5.0",
+    "commander": "^9.4.1",
+    "execa": "^5.1.1",
+    "node-fetch": "^2.7.0",
+    "ora": "^5.4.1",
+    "sqlite": "^4.1.2",
+    "tiktoken": "^1.0.21",
+    "uuid": "^9.0.0",
+    "yargs": "^17.6.2"
+  },
+  "devDependencies": {
+    "@types/better-sqlite3": "^7.6.3",
+    "@types/node": "^18.11.18",
+    "@types/node-fetch": "^2.6.12",
+    "@types/uuid": "^9.0.0",
+    "@types/yargs": "^17.0.19",
+    "ts-node": "^10.9.1",
+    "typescript": "^4.9.4"
+  }
+}

evals/tsconfig.json

@@ -0,0 +1,6 @@
+{
+	"extends": "../tsconfig.json",
+	"compilerOptions": {
+		"baseUrl": ".."
+	}
+}