reorganize folder structure

This commit is contained in:
alckasoc
2025-03-11 15:19:34 -07:00
parent c011a9e654
commit 6db793fcf7
81 changed files with 85 additions and 70 deletions
@@ -7,8 +7,8 @@ from typing import Dict, List, Optional, Tuple, Any, Sequence
import numpy as np
import requests
from gui_agents.v1.aci.ACI import ACI
from gui_agents.v1.utils.common_utils import box_iou
from gui_agents.s1.aci.ACI import ACI
from gui_agents.s1.utils.common_utils import box_iou
import platform
@@ -5,7 +5,7 @@ from typing import Any, Dict, List, Tuple
import numpy as np
import requests
import platform
from gui_agents.v1.utils.common_utils import box_iou
from gui_agents.s1.utils.common_utils import box_iou
if platform.system() == "Darwin":
from AppKit import *
@@ -15,7 +15,7 @@ if platform.system() == "Darwin":
AXUIElementCreateSystemWide,
)
from gui_agents.v1.aci.ACI import ACI, agent_action
from gui_agents.s1.aci.ACI import ACI, agent_action
def _normalize_key(key: str) -> str:
@@ -6,7 +6,7 @@ from typing import Any, Dict, List, Tuple
import numpy as np
import psutil
import requests
from gui_agents.v1.utils.common_utils import box_iou
from gui_agents.s1.utils.common_utils import box_iou
if platform.system() == "Windows":
import pywinauto
@@ -14,7 +14,7 @@ if platform.system() == "Windows":
import win32gui
import win32process
from gui_agents.v1.aci.ACI import ACI, agent_action
from gui_agents.s1.aci.ACI import ACI, agent_action
# Helper functions
@@ -6,7 +6,7 @@ import xml.etree.ElementTree as ET
from typing import Dict, List, Tuple
import numpy as np
import requests
from gui_agents.v1.utils.common_utils import box_iou
from gui_agents.s1.utils.common_utils import box_iou
logger = logging.getLogger("desktopenv.agent")
@@ -11,17 +11,17 @@ import pyautogui
if platform.system() == "Darwin":
current_platform = "macos"
from gui_agents.v1.aci.MacOSACI import MacOSACI, UIElement
from gui_agents.s1.aci.MacOSACI import MacOSACI, UIElement
elif platform.system() == "Windows":
current_platform = "windows"
from gui_agents.v1.aci.WindowsOSACI import WindowsACI, UIElement
from gui_agents.s1.aci.WindowsOSACI import WindowsACI, UIElement
elif platform.system() == "Linux":
current_platform = "ubuntu"
from gui_agents.v1.aci.LinuxOSACI import LinuxACI, UIElement
from gui_agents.s1.aci.LinuxOSACI import LinuxACI, UIElement
else:
raise ValueError("Unsupported platform")
from gui_agents.v1.core.AgentS import GraphSearchAgent, UIAgent
from gui_agents.s1.core.AgentS import GraphSearchAgent, UIAgent
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)
@@ -6,10 +6,10 @@ import time
from dataclasses import dataclass
from typing import Any, Dict, List, Optional, Tuple
from gui_agents.v1.aci.ACI import ACI
from gui_agents.v1.core.Manager import Manager
from gui_agents.v1.core.Worker import Worker
from gui_agents.v1.utils.common_utils import Node
from gui_agents.s1.aci.ACI import ACI
from gui_agents.s1.core.Manager import Manager
from gui_agents.s1.core.Worker import Worker
from gui_agents.s1.utils.common_utils import Node
logger = logging.getLogger("desktopenv.agent")
working_dir = os.path.dirname(os.path.abspath(__file__))
@@ -1,6 +1,6 @@
from typing import Dict, Optional
from gui_agents.v1.mllm.MultimodalAgent import LMMAgent
from gui_agents.s1.mllm.MultimodalAgent import LMMAgent
class BaseModule:
@@ -5,15 +5,15 @@ from typing import Dict, Tuple
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
from gui_agents.v1.core.BaseModule import BaseModule
from gui_agents.v1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.v1.mllm.MultimodalEngine import OpenAIEmbeddingEngine
from gui_agents.v1.utils.common_utils import (
from gui_agents.s1.core.BaseModule import BaseModule
from gui_agents.s1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.s1.mllm.MultimodalEngine import OpenAIEmbeddingEngine
from gui_agents.s1.utils.common_utils import (
load_embeddings,
load_knowledge_base,
save_embeddings,
)
from gui_agents.v1.utils.query_perplexica import query_to_perplexica
from gui_agents.s1.utils.query_perplexica import query_to_perplexica
working_dir = os.path.dirname(os.path.abspath(__file__))
@@ -3,11 +3,11 @@ import os
from collections import defaultdict
from typing import Dict, List, Optional, Tuple
from gui_agents.v1.aci.ACI import ACI
from gui_agents.v1.core.BaseModule import BaseModule
from gui_agents.v1.core.Knowledge import KnowledgeBase
from gui_agents.v1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.v1.utils.common_utils import (
from gui_agents.s1.aci.ACI import ACI
from gui_agents.s1.core.BaseModule import BaseModule
from gui_agents.s1.core.Knowledge import KnowledgeBase
from gui_agents.s1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.s1.utils.common_utils import (
Dag,
Node,
calculate_tokens,
@@ -4,13 +4,13 @@ import os
import re
from typing import Dict, List, Tuple
from gui_agents.v1.aci.ACI import ACI
from gui_agents.v1.core.BaseModule import BaseModule
from gui_agents.v1.core.Knowledge import KnowledgeBase
from gui_agents.v1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.v1.mllm.MultimodalEngine import OpenAIEmbeddingEngine
from gui_agents.v1.utils import common_utils
from gui_agents.v1.utils.common_utils import Node, calculate_tokens, call_llm_safe
from gui_agents.s1.aci.ACI import ACI
from gui_agents.s1.core.BaseModule import BaseModule
from gui_agents.s1.core.Knowledge import KnowledgeBase
from gui_agents.s1.core.ProceduralMemory import PROCEDURAL_MEMORY
from gui_agents.s1.mllm.MultimodalEngine import OpenAIEmbeddingEngine
from gui_agents.s1.utils import common_utils
from gui_agents.s1.utils.common_utils import Node, calculate_tokens, call_llm_safe
logger = logging.getLogger("desktopenv.agent")
working_dir = os.path.dirname(os.path.abspath(__file__))
@@ -5,7 +5,7 @@
import base64
import re
from gui_agents.v1.mllm.MultimodalEngine import (
from gui_agents.s1.mllm.MultimodalEngine import (
LMMEngineAnthropic,
LMMEngineAzureOpenAI,
LMMEngineOpenAI,
@@ -4,10 +4,10 @@ import os
import shutil
from typing import Any, Dict, List, Optional, Tuple
from gui_agents.v2.core.grounding import ACI
from gui_agents.v2.core.worker import Worker
from gui_agents.v2.core.manager import Manager
from gui_agents.v2.utils.common_utils import Node
from gui_agents.s2.agents.grounding import ACI
from gui_agents.s2.agents.worker import Worker
from gui_agents.s2.agents.manager import Manager
from gui_agents.s2.utils.common_utils import Node
logger = logging.getLogger("desktopenv.agent")
working_dir = os.path.dirname(os.path.abspath(__file__))
@@ -181,7 +181,7 @@ class GraphSearchAgent(UIAgent):
self.step_count = 0
def predict(
self, instruction: str, observation: Dict, info: Dict[str, Any]
self, instruction: str, observation: Dict
) -> Tuple[Dict, List[str]]:
# Initialize the three info dictionaries
planner_info = {}
@@ -15,9 +15,9 @@ from openai import OpenAI
from PIL import Image
from pytesseract import Output
from gui_agents.v2.core.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.v2.mllm.mllm import LMMAgent
from gui_agents.v2.utils.common_utils import (
from gui_agents.s2.memory.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.s2.core.mllm import LMMAgent
from gui_agents.s2.utils.common_utils import (
call_llm_safe,
parse_single_code_from_string,
)
@@ -3,11 +3,11 @@ import re
from collections import defaultdict
from typing import Dict, List, Optional, Tuple
from gui_agents.v2.core.grounding import ACI
from gui_agents.v2.core.base_module import BaseModule
from gui_agents.v2.core.knowledge import KnowledgeBase
from gui_agents.v2.core.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.v2.utils.common_utils import (
from gui_agents.s2.agents.grounding import ACI
from gui_agents.s2.core.module import BaseModule
from gui_agents.s2.core.knowledge import KnowledgeBase
from gui_agents.s2.memory.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.s2.utils.common_utils import (
Dag,
Node,
calculate_tokens,
@@ -3,12 +3,12 @@ import re
import textwrap
from typing import Dict, List, Tuple
from gui_agents.v2.core.grounding import ACI
from gui_agents.v2.core.base_module import BaseModule
from gui_agents.v2.core.knowledge import KnowledgeBase
from gui_agents.v2.core.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.v2.utils import common_utils
from gui_agents.v2.utils.common_utils import Node, calculate_tokens, call_llm_safe
from gui_agents.s2.agents.grounding import ACI
from gui_agents.s2.core.module import BaseModule
from gui_agents.s2.core.knowledge import KnowledgeBase
from gui_agents.s2.memory.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.s2.utils import common_utils
from gui_agents.s2.utils.common_utils import Node, calculate_tokens, call_llm_safe
logger = logging.getLogger("desktopenv.agent")
@@ -18,8 +18,8 @@ elif platform.system() == "Windows":
else:
raise ValueError("Unsupported platform")
from gui_agents.v2.core.grounding import OSWorldACI
from gui_agents.v2.core.agent_s import GraphSearchAgent
from gui_agents.s2.agents.grounding import OSWorldACI
from gui_agents.s2.agents.agent_s import GraphSearchAgent
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)
@@ -148,9 +148,24 @@ def main():
default="gpt-4o-mini",
help="Specify the model to use (e.g., gpt-4o)",
)
parser.add_argument(
"--endpoint_provider",
type=str,
default="huggingface",
help="Specify the endpoint provider (e.g., huggingface)",
)
parser.add_argument(
"--endpoint_url",
type=str,
default="",
help="Specify the endpoint URL to your HuggingFace Inference Endpoint.",
)
args = parser.parse_args()
grounding_agent = OSWorldACI()
grounding_agent = OSWorldACI(
endpoint_provider=args.endpoint_provider,
endpoint_url=args.endpoint_url
)
while True:
query = input("Query: ")
@@ -5,16 +5,16 @@ from typing import Dict, Tuple
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
from gui_agents.v2.core.base_module import BaseModule
from gui_agents.v2.core.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.v2.mllm.engine import OpenAIEmbeddingEngine
from gui_agents.v2.utils.common_utils import (
from gui_agents.s2.core.module import BaseModule
from gui_agents.s2.memory.procedural_memory import PROCEDURAL_MEMORY
from gui_agents.s2.core.engine import OpenAIEmbeddingEngine
from gui_agents.s2.utils.common_utils import (
call_llm_safe,
load_embeddings,
load_knowledge_base,
save_embeddings,
)
from gui_agents.v2.utils.query_perplexica import query_to_perplexica
from gui_agents.s2.utils.query_perplexica import query_to_perplexica
class KnowledgeBase(BaseModule):
@@ -2,7 +2,7 @@ import base64
import numpy as np
from gui_agents.v2.mllm.engine import (
from gui_agents.s2.core.engine import (
LMMEngineAnthropic,
LMMEngineAzureOpenAI,
LMMEngineOpenAI,
@@ -1,5 +1,5 @@
from typing import Dict, Optional
from gui_agents.v2.mllm.mllm import LMMAgent
from gui_agents.s2.core.mllm import LMMAgent
class BaseModule:
View File
+4 -4
View File
@@ -4,7 +4,7 @@ import platform
from fastapi import FastAPI, HTTPException
from fastapi.responses import StreamingResponse
from pydantic import BaseModel
from gui_agents.v1.core.AgentS import GraphSearchAgent
from gui_agents.s1.core.AgentS import GraphSearchAgent
import io
import pyautogui
import time
@@ -13,15 +13,15 @@ from threading import Event, Lock
# Determine the operating system and select appropriate ACI
os_name = platform.system().lower()
if os_name == "linux":
from gui_agents.v1.aci.LinuxOSACI import LinuxACI, UIElement
from gui_agents.s1.aci.LinuxOSACI import LinuxACI, UIElement
grounding_agent = LinuxACI()
platform_name = "ubuntu"
elif os_name == "darwin":
from gui_agents.v1.aci.MacOSACI import MacOSACI, UIElement
from gui_agents.s1.aci.MacOSACI import MacOSACI, UIElement
grounding_agent = MacOSACI()
platform_name = "macos"
elif os_name == "windows":
from gui_agents.v1.aci.WindowsOSACI import WindowsACI, UIElement
from gui_agents.s1.aci.WindowsOSACI import WindowsACI, UIElement
grounding_agent = WindowsACI()
platform_name = "windows"
else:
+1 -1
View File
@@ -2,7 +2,7 @@ from unittest.mock import Mock, patch
import pytest
from gui_agents.v1.aci.ACI import ACI, _normalize_key
from gui_agents.s1.aci.ACI import ACI, _normalize_key
@pytest.fixture
+1 -1
View File
@@ -3,7 +3,7 @@ import time
import pyautogui
from AppKit import NSWorkspace
from gui_agents.v1.aci.MacOSACI import MacOSACI
from gui_agents.s1.aci.MacOSACI import MacOSACI
agent = MacOSACI()
+1 -1
View File
@@ -3,7 +3,7 @@ from unittest.mock import Mock, patch
import pytest
from gui_agents.v1.aci.MacOSACI import UIElement
from gui_agents.s1.aci.MacOSACI import UIElement
@pytest.fixture