Support serialization when inputs/outputs contain generators (#17338)

Pydantic's `dict()` function raises an error here if you pass in a
generator. We have a more robust serialization function in lagnsmith
that we will use instead.

libs/core/langchain_core/tracers/langchain.py

@@ -1,4 +1,5 @@
 """A Tracer implementation that records to LangChain endpoint."""
+
 from __future__ import annotations
 
 import logging
@@ -62,6 +63,14 @@ def _get_executor() -> ThreadPoolExecutor:
     return _EXECUTOR
 
 
+def _run_to_dict(run: Run) -> dict:
+    return {
+        **run.dict(exclude={"child_runs", "inputs", "outputs"}),
+        "inputs": run.inputs.copy() if run.inputs is not None else None,
+        "outputs": run.outputs.copy() if run.outputs is not None else None,
+    }
+
+
 class LangChainTracer(BaseTracer):
     """Implementation of the SharedTracer that POSTS to the LangChain endpoint."""
 
@@ -150,7 +159,7 @@ class LangChainTracer(BaseTracer):
 
     def _persist_run_single(self, run: Run) -> None:
         """Persist a run."""
-        run_dict = run.dict(exclude={"child_runs"})
+        run_dict = _run_to_dict(run)
         run_dict["tags"] = self._get_tags(run)
         extra = run_dict.get("extra", {})
         extra["runtime"] = get_runtime_environment()
@@ -165,7 +174,7 @@ class LangChainTracer(BaseTracer):
     def _update_run_single(self, run: Run) -> None:
         """Update a run."""
         try:
-            run_dict = run.dict()
+            run_dict = _run_to_dict(run)
             run_dict["tags"] = self._get_tags(run)
             self.client.update_run(run.id, **run_dict)
         except Exception as e:

libs/core/tests/unit_tests/tracers/test_langchain.py

@@ -22,6 +22,7 @@ def test_example_id_assignment_threadsafe() -> None:
         return unittest.mock.MagicMock()
 
     client = unittest.mock.MagicMock(spec=Client)
+    client.tracing_queue = None
     client.create_run = mock_create_run
     tracer = LangChainTracer(client=client)
     old_persist_run_single = tracer._persist_run_single
@@ -35,6 +36,7 @@ def test_example_id_assignment_threadsafe() -> None:
     ):
         run_id_1 = UUID("9d878ab3-e5ca-4218-aef6-44cbdc90160a")
         run_id_2 = UUID("f1f9fa53-8b2f-4742-bdbc-38215f7bd1e1")
+        run_id_3 = UUID("f1f9fa53-8b2f-4742-bdbc-38215f7cd1e1")
         example_id_1 = UUID("57e42c57-8c79-4d9f-8765-bf6cd3a98055")
         tracer.example_id = example_id_1
         tracer.on_llm_start({"name": "example_1"}, ["foo"], run_id=run_id_1)
@@ -44,9 +46,14 @@ def test_example_id_assignment_threadsafe() -> None:
         tracer.on_llm_start({"name": "example_2"}, ["foo"], run_id=run_id_2)
         tracer.on_llm_end(LLMResult(generations=[], llm_output={}), run_id=run_id_2)
         tracer.example_id = None
+        tracer.on_chain_start(
+            {"name": "no_examples"}, {"inputs": (i for i in range(10))}, run_id=run_id_3
+        )
+        tracer.on_chain_error(ValueError("Foo bar"), run_id=run_id_3)
         expected_example_ids = {
             run_id_1: example_id_1,
             run_id_2: example_id_2,
+            run_id_3: None,
         }
         tracer.wait_for_futures()
         assert example_ids == expected_example_ids