vision-agent 0.2.159__tar.gz → 0.2.161__tar.gz

{vision_agent-0.2.159 → vision_agent-0.2.161}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vision-agent
-Version: 0.2.159
+Version: 0.2.161
 Summary: Toolset for Vision Agent
 Author: Landing AI
 Author-email: dev@landing.ai

{vision_agent-0.2.159 → vision_agent-0.2.161}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
 name = "vision-agent"
-version = "0.2.159"
+version = "0.2.161"
 description = "Toolset for Vision Agent"
 authors = ["Landing AI <dev@landing.ai>"]
 readme = "README.md"

{vision_agent-0.2.159 → vision_agent-0.2.161}/vision_agent/agent/vision_agent.py

@@ -197,6 +197,7 @@ class VisionAgent(Agent):
         local_artifacts_path: Optional[Union[str, Path]] = None,
         code_sandbox_runtime: Optional[str] = None,
         callback_message: Optional[Callable[[Dict[str, Any]], None]] = None,
+        code_interpreter: Optional[CodeInterpreter] = None,
     ) -> None:
         """Initialize the VisionAgent.
 
@@ -207,12 +208,14 @@ class VisionAgent(Agent):
             local_artifacts_path (Optional[Union[str, Path]]): The path to the local
                 artifacts file.
             code_sandbox_runtime (Optional[str]): The code sandbox runtime to use.
+            code_interpreter (Optional[CodeInterpreter]): if not None, use this CodeInterpreter
         """
 
         self.agent = AnthropicLMM(temperature=0.0) if agent is None else agent
         self.max_iterations = 12
         self.verbosity = verbosity
         self.code_sandbox_runtime = code_sandbox_runtime
+        self.code_interpreter = code_interpreter
         self.callback_message = callback_message
         if self.verbosity >= 1:
             _LOGGER.setLevel(logging.INFO)
@@ -284,9 +287,14 @@ class VisionAgent(Agent):
             # this is setting remote artifacts path
             artifacts = Artifacts(WORKSPACE / "artifacts.pkl")
 
-        with CodeInterpreterFactory.new_instance(
-            code_sandbox_runtime=self.code_sandbox_runtime,
-        ) as code_interpreter:
+        code_interpreter = (
+            self.code_interpreter
+            if self.code_interpreter is not None
+            else CodeInterpreterFactory.new_instance(
+                code_sandbox_runtime=self.code_sandbox_runtime,
+            )
+        )
+        with code_interpreter:
             orig_chat = copy.deepcopy(chat)
             int_chat = copy.deepcopy(chat)
             last_user_message = chat[-1]

{vision_agent-0.2.159 → vision_agent-0.2.161}/vision_agent/tools/meta_tools.py

@@ -1,3 +1,4 @@
+import base64
 import difflib
 import json
 import os
@@ -27,6 +28,7 @@ CURRENT_FILE = None
 CURRENT_LINE = 0
 DEFAULT_WINDOW_SIZE = 100
 ZMQ_PORT = os.environ.get("ZMQ_PORT", None)
+VERBOSITY = os.environ.get("VERBOSITY", 0)
 
 
 def report_progress_callback(port: int, inp: Dict[str, Any]) -> None:
@@ -374,7 +376,7 @@ def generate_vision_code(
             )
         )
     else:
-        agent = va.agent.VisionAgentCoder()
+        agent = va.agent.VisionAgentCoder(verbosity=int(VERBOSITY))
 
     fixed_chat: List[Message] = [{"role": "user", "content": chat, "media": media}]
     response = agent.chat_with_workflow(
@@ -437,7 +439,7 @@ def edit_vision_code(
             return dogs
     """
 
-    agent = va.agent.VisionAgentCoder()
+    agent = va.agent.VisionAgentCoder(verbosity=int(VERBOSITY))
     if name not in artifacts:
         print(f"[Artifact {name} does not exist]")
         return f"[Artifact {name} does not exist]"
@@ -510,6 +512,19 @@ def write_media_artifact(
         return f"[Invalid media type {type(media)}]"
     artifacts[name] = media_bytes
     print(f"[Media {name} saved]")
+    display(
+        {
+            MimeType.APPLICATION_ARTIFACT: json.dumps(
+                {
+                    "name": name,
+                    "action": "create",
+                    "content": base64.b64encode(media_bytes).decode("utf-8"),
+                    "contentType": "media_output",
+                }
+            )
+        },
+        raw=True,
+    )
     return f"[Media {name} saved]"