vision-agent 0.2.205__tar.gz → 0.2.207__tar.gz

{vision_agent-0.2.205 → vision_agent-0.2.207}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.205
+Version: 0.2.207
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai

{vision_agent-0.2.205 → vision_agent-0.2.207}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "vision-agent"
-version = "0.2.205"
+version = "0.2.207"
 description = "Toolset for Vision Agent"
 authors = ["Landing AI <dev@landing.ai>"]
 readme = "README.md"

{vision_agent-0.2.205 → vision_agent-0.2.207}/vision_agent/agent/vision_agent_coder_v2.py

@@ -34,7 +34,7 @@ from vision_agent.utils.execute import (
     CodeInterpreterFactory,
     Execution,
 )
-from vision_agent.utils.sim import Sim, load_cached_sim
+from vision_agent.utils.sim import Sim
 
 _CONSOLE = Console()
 
@@ -316,7 +316,7 @@ class VisionAgentCoderV2(AgentCoder):
             elif isinstance(tool_recommender, Sim):
                 self.tool_recommender = tool_recommender
         else:
-            self.tool_recommender = load_cached_sim(T.TOOLS_DF)
+            self.tool_recommender = T.get_tool_recommender()
 
         self.verbose = verbose
         self.code_sandbox_runtime = code_sandbox_runtime

{vision_agent-0.2.205 → vision_agent-0.2.207}/vision_agent/tools/__init__.py

@@ -45,6 +45,7 @@ from .tools import (
     generate_pose_image,
     generate_soft_edge_image,
     get_tool_documentation,
+    get_tool_recommender,
     git_vqa_v2,
     gpt4o_image_vqa,
     gpt4o_video_vqa,

{vision_agent-0.2.205 → vision_agent-0.2.207}/vision_agent/tools/planner_tools.py

@@ -32,10 +32,8 @@ from vision_agent.utils.execute import (
     MimeType,
 )
 from vision_agent.utils.image_utils import convert_to_b64
-from vision_agent.utils.sim import load_cached_sim
 
 TOOL_FUNCTIONS = {tool.__name__: tool for tool in T.TOOLS}
-TOOL_RECOMMENDER = load_cached_sim(T.TOOLS_DF)
 
 _LOGGER = logging.getLogger(__name__)
 EXAMPLES = f"\n{TEST_TOOLS_EXAMPLE1}\n{TEST_TOOLS_EXAMPLE2}\n"
@@ -52,7 +50,7 @@ def format_tool_output(tool_thoughts: str, tool_docstring: str) -> str:
 
 
 def extract_tool_info(
-    tool_choice_context: Dict[str, Any]
+    tool_choice_context: Dict[str, Any],
 ) -> Tuple[Optional[Callable], str, str, str]:
     tool_thoughts = tool_choice_context.get("thoughts", "")
     tool_docstring = ""
@@ -124,7 +122,7 @@ def run_tool_testing(
             f"I need models from the {category.strip()} category of tools. {task}"
         )
 
-    tool_docs = TOOL_RECOMMENDER.top_k(category, k=10, thresh=0.2)
+    tool_docs = T.get_tool_recommender().top_k(category, k=10, thresh=0.2)
     if exclude_tools is not None and len(exclude_tools) > 0:
         cleaned_tool_docs = []
         for tool_doc in tool_docs:
@@ -246,7 +244,9 @@ def get_tool_for_task(
                 context=f"<code>\n{code}\n</code>\n<tool_output>\n{tool_output_str}\n</tool_output>",
                 previous_attempts=error_message,
             )
-            tool_choice_context_dict = extract_json(lmm.generate(prompt, media=image_paths))  # type: ignore
+            tool_choice_context_dict = extract_json(
+                lmm.generate(prompt, media=image_paths)  # type: ignore
+            )
             tool, tool_thoughts, tool_docstring, error_message = extract_tool_info(
                 tool_choice_context_dict
             )

{vision_agent-0.2.205 → vision_agent-0.2.207}/vision_agent/tools/tools.py

@@ -4,6 +4,7 @@ import logging
 import os
 import tempfile
 import urllib.request
+from functools import lru_cache
 from importlib import resources
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple, Union, cast
@@ -45,6 +46,7 @@ from vision_agent.utils.image_utils import (
     rle_decode,
     rle_decode_array,
 )
+from vision_agent.utils.sim import Sim, load_cached_sim
 from vision_agent.utils.video import (
     extract_frames_from_video,
     frames_to_bytes,
@@ -80,6 +82,11 @@ _OCR_URL = "https://app.landing.ai/ocr/v1/detect-text"
 _LOGGER = logging.getLogger(__name__)
 
 
+@lru_cache(maxsize=1)
+def get_tool_recommender() -> Sim:
+    return load_cached_sim(TOOLS_DF)
+
+
 def grounding_dino(
     prompt: str,
     image: np.ndarray,