fix(sdk): make the 0.2.0 SDK release safe to publish (#6616)

The v2 SDK migration (#5273, #6564) shipped five breaking changes in both
SDKs but got the release mechanics wrong in three separate ways, and left
one of the two rewrites unable to complete a single successful call.

Versions. packages/ts-sdk/package.json read 0.1.3 -- a patch digit added
inside an unrelated compatibility commit, never deliberated. npm expands
^0.1.2 to >=0.1.2 <0.2.0, so every existing consumer would have picked the
break up on a lockfile refresh: AsyncExecutionResult.jobId renamed to
runId, executionId dropped from that interface, a failed sync run now
throwing instead of resolving {success:false}, the request body reshaped,
and the endpoint moved to /api/v2 with no fallback. 0.2.0 excludes every
existing range, so the upgrade becomes opt-in. packages/python-sdk carries
the identical break and was never bumped at all, so its publish job would
have skipped green at the "version already exists" gate and left the repo
and PyPI silently divergent; it moves 0.1.2 -> 0.2.0 in lockstep, along
with the __version__ string in simstudio/__init__.py, which tracks
pyproject and would otherwise have started lying. setup.py is left at
0.1.1: it is unchanged from main and demonstrably unread (0.1.2 published
from pyproject while setup.py already said 0.1.1). It wants deleting, in
its own commit.

A 404 fallback was considered and rejected. The legacy 202 body's statusUrl
points at /api/jobs/{jobId}, so mapping jobId onto runId would hand the
caller an id that getWorkflowRun cannot resolve against that same old
server -- a successful execute followed by an inexplicable failure on the
next call is a worse contract than a clean 404. Both READMEs instead state
the minimum server version and name the endpoint to check for.

Cancelled runs. packages/python-sdk computed success as status != 'failed',
so a run cancelled out of band reported success=True. The TypeScript SDK
uses a closed whitelist and reports False, and before the migration both
SDKs read the server's own value, which was False -- so this was a Python
regression, not merely an inconsistency. Fixed by mirroring the whitelist.
The v2 contract enumerates exactly completed|failed|paused|cancelled, so
narrowing the blacklist to a whitelist cannot drop a live value, and a
status added later now defaults to "not successful" rather than silently
reporting True. WorkflowExecutionResult gains a status field because
Python, unlike TypeScript, does not throw on 'failed' -- so success=False
alone is ambiguous there in a way it is not in the TypeScript SDK, which
is why status is not added to both.

Rate-limit header. Found while auditing the two SDKs for further
divergence, and the reason the Python bump could not have shipped as it
stood: every authenticated v2 response now carries X-RateLimit-Reset as an
ISO 8601 timestamp (recorded by v2RateLimits.publicApi, stamped by
withRouteHandler). The Python SDK parsed it with int(), raising a bare
ValueError that no handler in execute_workflow catches -- so every
successful v2 execution raised instead of returning. None of the legacy
endpoints the SDK previously called record a rate-limit snapshot, which is
why the latent int() survived until the v2 move. The TypeScript SDK
already branches on the format; _parse_reset_header mirrors it, including
degrading an unrecognised value to 0, because a quota hint must not take
down the call it rode in on.

Timing metadata. The v2 rewrite stopped forwarding startedAt/endedAt, which
main passed through and the TypeScript SDK still reports; restored under
the same startTime/endTime keys the TypeScript SDK uses.

Tests: cancelled/failed/paused status coverage, the ISO reset header, and
the restored metadata keys, each verified red against the unfixed line
first. The TypeScript suite gains matching cancelled/paused and ISO-reset
pins -- they pass against today's source by design, and were confirmed to
fail against a deliberately degraded copy so they are not toothless.

Deliberately not included: a CI guard failing a PR that changes SDK source
without a version bump. It would have caught this twice over, but it is a
new script and workflow rather than a fix to the defect at hand.

Review revision. bun.lock recorded packages/ts-sdk at 0.1.3 and was left
stale by the first pass, so the repo asserted two versions for the same
workspace package -- in a change whose whole thesis is that the version
strings had diverged. It does not break CI (bun 1.3.14 accepts the
mismatch under --frozen-lockfile, confirmed here), but 092311ea68 bumped
the lock in lockstep with package.json, and the next unfrozen install
would otherwise drop the line into an unrelated PR.

_parse_reset_header gated the numeric branch on str.isdigit(), which
accepts characters int() rejects ('²'.isdigit() is True, int('²')
raises) -- and that int() sits outside the try, so the one function added
to stop a quota hint raising could still raise, contradicting its own
docstring. str.isdecimal() is exactly the set int() accepts. The
tolerates-unparseable test is parametrized over both forms and was
confirmed red on '²' against isdigit.

Docs and docstrings: apps/docs api-reference/python.mdx mirrors the
README's dataclass block and was the only copy left without the new
status field. RateLimitInfo now names its units, because reset is epoch
seconds for the legacy integer and milliseconds for the ISO form that v2
sends. execute_workflow's Args entry still described the pre-v2 body
shape ("spread at root level"); every input is nested under input now,
and this is the commit that ships that help() text to PyPI. The
"declared last so positional construction keeps working" sentence was a
maintainer's note that belongs in this message, not in every user's
help(WorkflowExecutionResult).
This commit is contained in:
Waleed
2026-08-12 01:48:37 -07:00
committed by GitHub
parent 5a63eb8567
commit 34d65df7d6
9 changed files with 248 additions and 18 deletions
@@ -275,8 +275,11 @@ class WorkflowExecutionResult:
metadata: Optional[Dict[str, Any]] = None
trace_spans: Optional[List[Any]] = None
total_duration: Optional[float] = None
status: Optional[str] = None
```
`success` is `True` only for the `completed` and `paused` statuses. `status` carries the server's terminal status verbatim, so a cancelled run (`success=False`, `error=None`) is distinguishable from a failed one.
### AsyncExecutionResult
```python
+1 -1
View File
@@ -611,7 +611,7 @@
},
"packages/ts-sdk": {
"name": "simstudio-ts-sdk",
"version": "0.1.3",
"version": "0.2.0",
"devDependencies": {
"@sim/tsconfig": "workspace:*",
"@types/node": "24.2.1",
+19
View File
@@ -2,6 +2,22 @@
The official Python SDK for [Sim](https://sim.ai), allowing you to execute workflows programmatically from your Python applications.
## Server compatibility
`0.2.x` talks to the v2 API and has no fallback to the older endpoints, so it requires a Sim deployment that serves `POST /api/v2/workflows/{id}/execute`. That surface is newer than the endpoints `0.1.x` used, and a deployment can also have it switched off — a self-hosted build serves `/api/v2` only when the operator enables `V2_API`. Where it is unavailable every v2 route answers 404, so `execute_workflow` raises `SimStudioError('HTTP 404: Not Found')` — enable or upgrade the v2 API on the server, or pin `simstudio-sdk<0.2`, which keeps using `/api/workflows/{id}/execute` and `/api/jobs/{id}`.
## Upgrading from 0.1.x to 0.2.0
`0.2.0` is a breaking release.
- **Requests move to `/api/v2`.** `execute_workflow` posts to `/api/v2/workflows/{workflow_id}/execute`, sends the workflow input nested under `input`, and carries `async` / `executionTimeoutSeconds` in the body instead of the `X-Execution-Mode` and `X-Execution-Timeout-Seconds` headers.
- **`AsyncExecutionResult.job_id` is now `run_id`,** and `execution_id` has been removed from that dataclass. Replace `result.job_id` with `result.run_id`.
- **`get_job_status(job_id)` is legacy.** It still calls `/api/jobs/{job_id}` and only resolves IDs from a `0.1.x` async execution. For runs started by `0.2.x`, use `get_workflow_run(workflow_id, run_id)`, which reads `/api/v2/workflows/{workflow_id}/runs/{run_id}`.
- **`WorkflowExecutionResult.success` is derived from the run status** rather than read from the response body, and is `True` only for `completed` and `paused` runs — so a run cancelled while it was in flight now reports `success=False`, as it did before the v2 migration. The new `WorkflowExecutionResult.status` field carries the server's terminal status (`'completed'`, `'failed'`, `'paused'` or `'cancelled'`), which is how you tell a cancelled run from a failed one.
- **`metadata` is now built by the SDK,** with the keys `duration`, `runId`, `startTime` and `endTime`. The v2 response carries no execution logs or trace spans, so `logs` and `trace_spans` are always `None`; the pre-v2 `metadata['executionId']` is now `metadata['runId']`.
Note one deliberate difference from the TypeScript SDK: a failed synchronous run *throws* there, but here it returns normally with `error` set and `status='failed'`.
## Installation
```bash
@@ -245,8 +261,11 @@ class WorkflowExecutionResult:
metadata: Optional[Dict[str, Any]] = None
trace_spans: Optional[list] = None
total_duration: Optional[float] = None
status: Optional[str] = None
```
`success` is `True` only for the `completed` and `paused` statuses. `status` carries the server's terminal status verbatim, so a cancelled run (`success=False`, `error=None`) is distinguishable from a failed one.
### WorkflowStatus
```python
+1 -1
View File
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
[project]
name = "simstudio-sdk"
version = "0.1.2"
version = "0.2.0"
authors = [
{name = "Sim", email = "help@sim.ai"},
]
+53 -10
View File
@@ -6,6 +6,7 @@ Official Python SDK for Sim, allowing you to execute workflows programmatically.
from typing import Any, Dict, Optional, Union
from dataclasses import dataclass
from datetime import datetime
import time
import random
import os
@@ -14,7 +15,12 @@ import requests
MAX_EXECUTION_TIMEOUT_SECONDS = 604_800
__version__ = "0.1.2"
# Run statuses that count as a successful synchronous execution. Deliberately a
# whitelist: a status later added to the API then defaults to "not successful"
# rather than silently reporting True.
_SUCCESSFUL_RUN_STATUSES = ('completed', 'paused')
__version__ = "0.2.0"
__all__ = [
"SimStudioClient",
"SimStudioError",
@@ -28,7 +34,14 @@ __all__ = [
@dataclass
class WorkflowExecutionResult:
"""Result of a workflow execution."""
"""
Result of a workflow execution.
``success`` is True only for the 'completed' and 'paused' statuses, so a
run the server cancels and a run that fails both report False. Read
``status`` to tell those apart -- it carries the server's own terminal
status ('completed', 'failed', 'paused' or 'cancelled').
"""
success: bool
output: Optional[Any] = None
error: Optional[str] = None
@@ -36,6 +49,7 @@ class WorkflowExecutionResult:
metadata: Optional[Dict[str, Any]] = None
trace_spans: Optional[list] = None
total_duration: Optional[float] = None
status: Optional[str] = None
@dataclass
@@ -58,7 +72,13 @@ class AsyncExecutionResult:
@dataclass
class RateLimitInfo:
"""Rate limit information from API response headers."""
"""
Rate limit information from API response headers.
``reset`` is epoch milliseconds when the server sends the ISO 8601
``X-RateLimit-Reset`` the v2 API uses, and epoch seconds for the bare
integer older endpoints sent. ``retry_after`` is milliseconds.
"""
limit: int
remaining: int
reset: int
@@ -82,6 +102,23 @@ class SimStudioError(Exception):
self.status = status
def _parse_reset_header(value: str) -> int:
"""
Parse the ``X-RateLimit-Reset`` header, matching the TypeScript SDK.
The v2 API sends an ISO 8601 timestamp, which yields epoch milliseconds;
older endpoints sent an epoch integer, which is kept as-is. A quota hint
must never take down the call it rode in on, so an unrecognised value
degrades to 0 rather than raising.
"""
if value.isdecimal():
return int(value)
try:
return int(datetime.fromisoformat(value.replace('Z', '+00:00')).timestamp() * 1000)
except ValueError:
return 0
class SimStudioClient:
"""
Sim API client for executing workflows programmatically.
@@ -166,9 +203,11 @@ class SimStudioClient:
Args:
workflow_id: The ID of the workflow to execute
input: Input data to pass to the workflow. Can be a dict (spread at root level),
primitive value (string, number, bool), or list (wrapped in 'input' field).
File-like objects within dicts are automatically converted to base64.
input: Input data to pass to the workflow, sent nested under the request
body's 'input' field. A dict becomes the workflow input as-is; any
other value (string, number, bool, list) is wrapped as
{'input': value}. File-like objects within it are automatically
converted to base64.
timeout: Timeout in seconds (default: 30.0)
stream: Enable streaming responses (default: None)
selected_outputs: Block outputs to stream (e.g., ["agent1.content"])
@@ -270,15 +309,19 @@ class SimStudioClient:
)
execution_error = result_data.get('error')
status = result_data.get('status')
return WorkflowExecutionResult(
success=result_data.get('status') != 'failed',
success=status in _SUCCESSFUL_RUN_STATUSES,
output=result_data.get('output'),
error=execution_error.get('message') if execution_error else None,
metadata={
'duration': result_data.get('durationMs'),
'runId': result_data['runId']
'runId': result_data['runId'],
'startTime': result_data.get('startedAt'),
'endTime': result_data.get('endedAt')
},
total_duration=result_data.get('durationMs')
total_duration=result_data.get('durationMs'),
status=status
)
except requests.Timeout:
@@ -593,7 +636,7 @@ class SimStudioClient:
self._rate_limit_info = RateLimitInfo(
limit=int(limit) if limit else 0,
remaining=int(remaining) if remaining else 0,
reset=int(reset) if reset else 0,
reset=_parse_reset_header(reset) if reset else 0,
retry_after=int(retry_after) * 1000 if retry_after else None
)
+106 -3
View File
@@ -7,19 +7,32 @@ from unittest.mock import Mock, patch
from simstudio import SimStudioClient, SimStudioError, WorkflowExecutionResult, WorkflowStatus
def v2_execution_response(output=None):
def v2_execution_response(output=None, status="completed", error=None):
return {
"data": {
"runId": "execution-123",
"workflowId": "workflow-id",
"status": "completed",
"status": status,
"output": {} if output is None else output,
"error": None,
"error": error,
"startedAt": "2026-08-11T12:00:00.000Z",
"endedAt": "2026-08-11T12:00:00.010Z",
"durationMs": 10
}
}
def mock_execution_post(mock_post, status="completed", error=None, headers=None):
"""Wire a mocked 200 v2 execution response with the given terminal status."""
mock_response = Mock()
mock_response.ok = True
mock_response.status_code = 200
mock_response.json.return_value = v2_execution_response(status=status, error=error)
mock_response.headers.get.side_effect = lambda h: (headers or {}).get(h)
mock_post.return_value = mock_response
return mock_response
def test_simstudio_client_initialization():
"""Test SimStudioClient initialization."""
client = SimStudioClient(api_key="test-api-key", base_url="https://test.sim.ai")
@@ -161,10 +174,58 @@ def test_sync_execution_returns_result(mock_post):
)
assert result.success is True
assert result.status == "completed"
assert result.output == {"result": "completed"}
assert result.metadata == {
"duration": 10,
"runId": "execution-123",
"startTime": "2026-08-11T12:00:00.000Z",
"endTime": "2026-08-11T12:00:00.010Z"
}
assert not hasattr(result, 'task_id')
@patch('simstudio.requests.Session.post')
def test_sync_execution_cancelled_is_not_success(mock_post):
"""A run cancelled out of band is not a success, matching the TypeScript SDK."""
mock_execution_post(mock_post, status="cancelled")
client = SimStudioClient(api_key="test-api-key")
result = client.execute_workflow("workflow-id", {})
assert result.success is False
assert result.status == "cancelled"
@patch('simstudio.requests.Session.post')
def test_sync_execution_failed_is_not_success(mock_post):
"""A failed run is not a success and surfaces the server's error message."""
mock_execution_post(
mock_post,
status="failed",
error={"code": "BLOCK_EXECUTION_FAILED", "message": "Invalid credentials"}
)
client = SimStudioClient(api_key="test-api-key")
result = client.execute_workflow("workflow-id", {})
assert result.success is False
assert result.status == "failed"
assert result.error == "Invalid credentials"
@patch('simstudio.requests.Session.post')
def test_sync_execution_paused_is_success(mock_post):
"""A paused run is still a success -- it is waiting, not broken."""
mock_execution_post(mock_post, status="paused")
client = SimStudioClient(api_key="test-api-key")
result = client.execute_workflow("workflow-id", {})
assert result.success is True
assert result.status == "paused"
@patch('simstudio.requests.Session.post')
def test_async_header_not_set_when_false(mock_post):
"""Test X-Execution-Mode header is not set when async_execution is None."""
@@ -497,6 +558,48 @@ def test_get_rate_limit_info_after_api_call(mock_post):
assert info.reset == 1704067200
@patch('simstudio.requests.Session.post')
def test_rate_limit_reset_accepts_iso_timestamp(mock_post):
"""The v2 API sends X-RateLimit-Reset as an ISO 8601 timestamp, not an epoch int."""
mock_execution_post(mock_post, headers={
'x-ratelimit-limit': '100',
'x-ratelimit-remaining': '99',
'x-ratelimit-reset': '2024-01-01T00:00:00.000Z'
})
client = SimStudioClient(api_key="test-api-key")
result = client.execute_workflow("workflow-id", {})
assert result.success is True
info = client.get_rate_limit_info()
assert info is not None
assert info.reset == 1704067200000
@pytest.mark.parametrize('reset_header', ['not-a-timestamp', '²'])
@patch('simstudio.requests.Session.post')
def test_rate_limit_reset_tolerates_unparseable_value(mock_post, reset_header):
"""
An unrecognised quota hint reports 0 rather than failing the execution.
'²' covers the digit-like characters str.isdigit() accepts but int()
rejects.
"""
mock_execution_post(mock_post, headers={
'x-ratelimit-limit': '100',
'x-ratelimit-remaining': '99',
'x-ratelimit-reset': reset_header
})
client = SimStudioClient(api_key="test-api-key")
result = client.execute_workflow("workflow-id", {})
assert result.success is True
info = client.get_rate_limit_info()
assert info is not None
assert info.reset == 0
@patch('simstudio.requests.Session.get')
def test_get_usage_limits_success(mock_get):
"""Test getting usage limits."""
+14
View File
@@ -2,6 +2,20 @@
The official TypeScript/JavaScript SDK for [Sim](https://sim.ai), allowing you to execute workflows programmatically from your applications.
## Server compatibility
`0.2.x` talks to the v2 API and has no fallback to the older endpoints, so it requires a Sim deployment that serves `POST /api/v2/workflows/{id}/execute`. That surface is newer than the endpoints `0.1.x` used, and a deployment can also have it switched off — a self-hosted build serves `/api/v2` only when the operator enables `V2_API`. Where it is unavailable every v2 route answers 404, so `executeWorkflow` fails with `HTTP 404: Not Found` — enable or upgrade the v2 API on the server, or stay on `simstudio-ts-sdk@0.1.x`, which keeps using `/api/workflows/{id}/execute` and `/api/jobs/{id}`.
## Upgrading from 0.1.x to 0.2.0
`0.2.0` is a breaking release. It is a minor bump rather than a patch precisely so that `^0.1.2` does not pick it up — you upgrade when you choose to.
- **Requests move to `/api/v2`.** `executeWorkflow` posts to `/api/v2/workflows/{id}/execute`, sends the workflow input nested under `input`, and carries `async` / `executionTimeoutSeconds` in the body instead of the `X-Execution-Mode` and `X-Execution-Timeout-Seconds` headers.
- **`AsyncExecutionResult.jobId` is now `runId`,** and `executionId` has been removed from that interface. Replace `result.jobId` with `result.runId`.
- **`getJobStatus(taskId)` is legacy.** It still calls `/api/jobs/{taskId}` and only resolves IDs from a `0.1.x` async execution. For runs started by `0.2.x`, use `getWorkflowRun(workflowId, runId)`, which reads `/api/v2/workflows/{id}/runs/{runId}` and returns a typed `WorkflowRunStatus`.
- **A failed synchronous run now throws.** Previously it resolved with `{ success: false }`; it now rejects with a `SimStudioError` carrying the server's `error.code` and `error.message`. Any `if (!result.success)` branch that handled failures must move into a `catch`.
- **`success` is derived from the run status.** It is `true` for `completed` and `paused` runs only, so a run cancelled while it was in flight resolves with `success: false` rather than throwing. Combined with the point above: a rejection means the run failed, and a resolved `success: false` means it was cancelled.
## Installation
```bash
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "simstudio-ts-sdk",
"version": "0.1.3",
"version": "0.2.0",
"description": "Sim SDK - Execute workflows programmatically",
"type": "module",
"exports": {
+50 -2
View File
@@ -4,12 +4,12 @@ import { SimStudioClient, SimStudioError } from './index'
const mockFetch = vi.fn()
vi.stubGlobal('fetch', mockFetch)
function v2ExecutionResponse(output: unknown = {}) {
function v2ExecutionResponse(output: unknown = {}, status = 'completed') {
return {
data: {
runId: 'execution-123',
workflowId: 'workflow-id',
status: 'completed',
status,
output,
error: null,
startedAt: '2026-08-11T12:00:00.000Z',
@@ -195,6 +195,32 @@ describe('SimStudioClient', () => {
})
})
it('reports a cancelled sync run as unsuccessful', async () => {
vi.mocked(mockFetch).mockResolvedValue({
ok: true,
status: 200,
json: vi.fn().mockResolvedValue(v2ExecutionResponse({}, 'cancelled')),
headers: { get: vi.fn().mockReturnValue(null) },
})
const result = await client.executeWorkflow('workflow-id', {})
expect(result).toHaveProperty('success', false)
})
it('reports a paused sync run as successful', async () => {
vi.mocked(mockFetch).mockResolvedValue({
ok: true,
status: 200,
json: vi.fn().mockResolvedValue(v2ExecutionResponse({}, 'paused')),
headers: { get: vi.fn().mockReturnValue(null) },
})
const result = await client.executeWorkflow('workflow-id', {})
expect(result).toHaveProperty('success', true)
})
it('should not set X-Execution-Mode header when async is undefined', async () => {
const mockResponse = {
ok: true,
@@ -512,6 +538,28 @@ describe('SimStudioClient', () => {
expect(info?.remaining).toBe(95)
expect(info?.reset).toBe(1704067200)
})
it('parses an ISO x-ratelimit-reset, the format the v2 API sends', async () => {
const mockResponse = {
ok: true,
status: 200,
json: vi.fn().mockResolvedValue(v2ExecutionResponse()),
headers: {
get: vi.fn((header: string) => {
if (header === 'x-ratelimit-limit') return '100'
if (header === 'x-ratelimit-remaining') return '99'
if (header === 'x-ratelimit-reset') return '2024-01-01T00:00:00.000Z'
return null
}),
},
}
vi.mocked(mockFetch).mockResolvedValue(mockResponse as any)
await client.executeWorkflow('workflow-id', {})
expect(client.getRateLimitInfo()?.reset).toBe(1704067200000)
})
})
describe('getUsageLimits', () => {