revert decorator io media

Author: hassiebp

langfuse/decorators/langfuse_decorator.py

@@ -5,6 +5,7 @@
 from functools import wraps
 import httpx
 import inspect
+import json
 import logging
 from typing import (
     Any,
@@ -37,6 +38,7 @@
     ScoreDataType,
     StateType,
 )
+from langfuse.serializer import EventSerializer
 from langfuse.types import ObservationParams, SpanLevel
 from langfuse.utils import _get_timestamp
 from langfuse.utils.langfuse_singleton import LangfuseSingleton
@@ -388,11 +390,15 @@ def _get_input_from_func_args(
     ) -> Any:
         # Remove implicitly passed "self" or "cls" argument for instance or class methods
         logged_args = func_args[1:] if is_method else func_args
-        return {
+        raw_input = {
             "args": logged_args,
             "kwargs": func_kwargs,
         }
 
+        # Serialize and deserialize to ensure proper JSON serialization.
+        # Objects are later serialized again so deserialization is necessary here to avoid unnecessary escaping of quotes.
+        return json.loads(json.dumps(raw_input, cls=EventSerializer))
+
     def _finalize_call(
         self,
         observation: Optional[
@@ -440,10 +446,13 @@ def _handle_call_result(
             )
 
             end_time = observation_params["end_time"] or _get_timestamp()
-            output = observation_params["output"] or (
+            raw_output = observation_params["output"] or (
                 result if result and capture_output else None
             )
 
+            # Serialize and deserialize to ensure proper JSON serialization.
+            # Objects are later serialized again so deserialization is necessary here to avoid unnecessary escaping of quotes.
+            output = json.loads(json.dumps(raw_output, cls=EventSerializer))
             observation_params.update(end_time=end_time, output=output)
 
             if isinstance(observation, (StatefulSpanClient, StatefulGenerationClient)):

tests/test_decorators.py

@@ -1462,19 +1462,21 @@ def main():
     assert len(child_observations) == 2
 
 
-def test_pdf_in_metadata():
+def test_media():
     mock_trace_id = create_uuid()
 
     with open("static/bitcoin.pdf", "rb") as pdf_file:
         pdf_bytes = pdf_file.read()
 
+    media = LangfuseMedia(content_bytes=pdf_bytes, content_type="application/pdf")
+
     @observe()
     def main():
         langfuse_context.update_current_trace(
             metadata={
-                "context": LangfuseMedia(
-                    content_bytes=pdf_bytes, content_type="application/pdf"
-                )
+                "context": {
+                    "nested": media,
+                },
             },
         )
 
@@ -1484,4 +1486,13 @@ def main():
 
     trace_data = get_api().trace.get(mock_trace_id)
 
-    assert "@@@langfuseMedia:type=application/pdf|id=" in trace_data.metadata["context"]
+    assert (
+        "@@@langfuseMedia:type=application/pdf|id="
+        in trace_data.metadata["context"]["nested"]
+    )
+    parsed_reference_string = LangfuseMedia.parse_reference_string(
+        trace_data.metadata["context"]["nested"]
+    )
+    assert parsed_reference_string["content_type"] == "application/pdf"
+    assert parsed_reference_string["media_id"] is not None
+    assert parsed_reference_string["source"] == "bytes"