vision-agent 0.2.150__tar.gz → 0.2.152__tar.gz

{vision_agent-0.2.150 → vision_agent-0.2.152}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.150
+Version: 0.2.152
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai

{vision_agent-0.2.150 → vision_agent-0.2.152}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "vision-agent"
-version = "0.2.150"
+version = "0.2.152"
 description = "Toolset for Vision Agent"
 authors = ["Landing AI <dev@landing.ai>"]
 readme = "README.md"

{vision_agent-0.2.150 → vision_agent-0.2.152}/vision_agent/agent/vision_agent.py

@@ -307,7 +307,7 @@ class VisionAgent(Agent):
             remote_artifacts_path = code_interpreter.upload_file(
                 self.local_artifacts_path
             )
-            artifacts_loaded = artifacts.show()
+            artifacts_loaded = artifacts.show(code_interpreter.remote_path)
             int_chat.append({"role": "observation", "content": artifacts_loaded})
             orig_chat.append({"role": "observation", "content": artifacts_loaded})
             self.streaming_message({"role": "observation", "content": artifacts_loaded})

{vision_agent-0.2.150 → vision_agent-0.2.152}/vision_agent/tools/meta_tools.py

@@ -107,13 +107,16 @@ class Artifacts:
                 with open(self.remote_save_path.parent / k, mode) as f:
                     f.write(v)
 
-    def show(self) -> str:
+    def show(self, uploaded_file_path: Optional[Union[str, Path]] = None) -> str:
         """Shows the artifacts that have been loaded and their remote save paths."""
+        loaded_path = (
+            Path(uploaded_file_path)
+            if uploaded_file_path is not None
+            else self.remote_save_path
+        )
         output_str = "[Artifacts loaded]\n"
         for k in self.artifacts.keys():
-            output_str += (
-                f"Artifact {k} loaded to {str(self.remote_save_path.parent / k)}\n"
-            )
+            output_str += f"Artifact {k} loaded to {str(loaded_path / k)}\n"
         output_str += "[End of artifacts]\n"
         print(output_str)
         return output_str

{vision_agent-0.2.150 → vision_agent-0.2.152}/vision_agent/tools/tools.py

@@ -1603,7 +1603,7 @@ def extract_frames_and_timestamps(
     """
 
     def reformat(
-        frames_and_timestamps: List[Tuple[np.ndarray, float]]
+        frames_and_timestamps: List[Tuple[np.ndarray, float]],
     ) -> List[Dict[str, Union[np.ndarray, float]]]:
         return [
             {"frame": frame, "timestamp": timestamp}
@@ -2017,7 +2017,7 @@ def overlay_counting_results(
         fontsize,
     )
 
-    for i, elt in enumerate(instances):
+    for i, elt in enumerate(instances, 1):
         label = f"{i}"
         box = elt["bbox"]
 

{vision_agent-0.2.150 → vision_agent-0.2.152}/vision_agent/utils/execute.py

@@ -464,7 +464,9 @@ print(f"Vision Agent version: {va_version}")"""
         _LOGGER.info(
             f"E2BCodeInterpreter (sandbox id: {self.interpreter.sandbox_id}) initialized:\n{sys_versions}"
         )
-        self.remote_path = Path(remote_path if remote_path is not None else WORKSPACE)
+        self.remote_path = Path(
+            remote_path if remote_path is not None else "/home/user"
+        )
 
     def close(self, *args: Any, **kwargs: Any) -> None:
         try:
@@ -713,7 +715,7 @@ def _get_e2b_env() -> Union[Dict[str, str], None]:
     openai_api_key = os.getenv("OPENAI_API_KEY", "")
     anthropic_api_key = os.getenv("ANTHROPIC_API_KEY", "")
     if openai_api_key or anthropic_api_key:
-        envs = {}
+        envs = {"WORKSPACE": os.getenv("WORKSPACE", "/home/user")}
         if openai_api_key:
             envs["OPENAI_API_KEY"] = openai_api_key
         if anthropic_api_key: