vision-agent 0.2.237__tar.gz → 0.2.238__tar.gz

{vision_agent-0.2.237 → vision_agent-0.2.238}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.237
+Version: 0.2.238
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai
@@ -15,7 +15,6 @@ Requires-Dist: e2b (>=0.17.2a50,<0.18.0)
 Requires-Dist: e2b-code-interpreter (==0.0.11a37)
 Requires-Dist: flake8 (>=7.0.0,<8.0.0)
 Requires-Dist: ipykernel (>=6.29.4,<7.0.0)
-Requires-Dist: langsmith (>=0.1.58,<0.2.0)
 Requires-Dist: libcst (>=1.5.0,<2.0.0)
 Requires-Dist: matplotlib (>=3.9.2,<4.0.0)
 Requires-Dist: nbclient (>=0.10.0,<0.11.0)

{vision_agent-0.2.237 → vision_agent-0.2.238}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "vision-agent"
-version = "0.2.237"
+version = "0.2.238"
 description = "Toolset for Vision Agent"
 authors = ["Landing AI <dev@landing.ai>"]
 readme = "README.md"
@@ -34,7 +34,6 @@ scipy = "1.13.*"
 nbclient = "^0.10.0"
 nbformat = "^5.10.4"
 rich = "^13.7.1"
-langsmith = "^0.1.58"
 ipykernel = "^6.29.4"
 e2b = "^0.17.2a50"
 e2b-code-interpreter = "0.0.11a37"

vision_agent-0.2.237/vision_agent/configs/anthropic_openai_config.py → vision_agent-0.2.238/vision_agent/configs/config.py

@@ -96,13 +96,24 @@ class Config(BaseModel):
         }
     )
 
+    # for get_tool_for_task
+    od_judge: Type[LMM] = Field(default=AnthropicLMM)
+    od_judge_kwargs: dict = Field(
+        default_factory=lambda: {
+            "model_name": "claude-3-5-sonnet-20241022",
+            "temperature": 0.0,
+            "image_size": 512,
+        }
+    )
+
     # for suggestions module
-    suggester: Type[LMM] = Field(default=AnthropicLMM)
+    suggester: Type[LMM] = Field(default=OpenAILMM)
     suggester_kwargs: dict = Field(
         default_factory=lambda: {
-            "model_name": "claude-3-5-sonnet-20241022",
+            "model_name": "o1",
             "temperature": 1.0,
-            "image_size": 768,
+            "image_detail": "high",
+            "image_size": 1024,
         }
     )
 
@@ -143,6 +154,9 @@ class Config(BaseModel):
     def create_tool_chooser(self) -> LMM:
         return self.tool_chooser(**self.tool_chooser_kwargs)
 
+    def create_od_judge(self) -> LMM:
+        return self.od_judge(**self.od_judge_kwargs)
+
     def create_suggester(self) -> LMM:
         return self.suggester(**self.suggester_kwargs)
 

{vision_agent-0.2.237 → vision_agent-0.2.238}/vision_agent/sim/__init__.py

@@ -2,6 +2,7 @@ from .sim import (
     AzureSim,
     OllamaSim,
     Sim,
+    StellaSim,
     get_tool_recommender,
     load_cached_sim,
     load_sim,

{vision_agent-0.2.237 → vision_agent-0.2.238}/vision_agent/tools/planner_tools.py

@@ -368,6 +368,15 @@ def get_tool_for_task(
     tool_tester = CONFIG.create_tool_tester()
     tool_chooser = CONFIG.create_tool_chooser()
 
+    if isinstance(images, list):
+        if len(images) > 0 and isinstance(images[0], dict):
+            if all(["frame" in image for image in images]):
+                images = [image["frame"] for image in images]
+            else:
+                raise ValueError(
+                    f"Expected a list of numpy arrays or a dictionary of strings to lists of numpy arrays, got a list of dictionaries instead: {images}"
+                )
+
     if isinstance(images, list):
         images = {"image": images}
 
@@ -410,6 +419,15 @@ def get_tool_for_task_human_reviewer(
     # NOTE: this will have the same documentation as get_tool_for_task
     tool_tester = CONFIG.create_tool_tester()
 
+    if isinstance(images, list):
+        if len(images) > 0 and isinstance(images[0], dict):
+            if all(["frame" in image for image in images]):
+                images = [image["frame"] for image in images]
+            else:
+                raise ValueError(
+                    f"Expected a list of numpy arrays or a dictionary of strings to lists of numpy arrays, got a list of dictionaries instead: {images}"
+                )
+
     if isinstance(images, list):
         images = {"image": images}
 
@@ -424,6 +442,9 @@ def get_tool_for_task_human_reviewer(
                 Image.fromarray(image).save(image_path)
                 image_paths.append(image_path)
 
+        # run no more than 3 images or else it overloads the LLM
+        image_paths = image_paths[:3]
+
         tools = [
             t.__name__
             for t in get_tools()