building knowledge graph index error, return metadata (error type: dict_type)

[zhengwanbo]

when build a graph knowledge, use qwen-trubo LLM.

[2025-02-27 13:40:59,221: INFO/ForkPoolWorker-1] Start building knowledge graph index for chunk #b458114c-a339-44cf-a788-509ed730e2fb.
[2025-02-27 13:40:59,231: DEBUG/ForkPoolWorker-1] OpenAI Response Token Usage: 2819
[2025-02-27 13:40:59,234: DEBUG/ForkPoolWorker-1] OpenAI Response Token Usage: 727
[2025-02-27 13:40:59,237: DEBUG/ForkPoolWorker-1] OpenAI Response Token Usage: 3243
[2025-02-27 13:40:59,239: DEBUG/ForkPoolWorker-1] OpenAI Response Token Usage: 3553
[2025-02-27 13:40:59,241: ERROR/ForkPoolWorker-1] app.tasks.build_index.build_kg_index_for_chunk[f8caf1c8-8649-4144-bccc-4530d6d52c72]: Failed to build knowledge graph index for chunk #b458114c-a339-44cf-a788-509ed730e2fb
Traceback (most recent call last):
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/tasks/build_index.py", line 125, in build_kg_index_for_chunk
index_service.build_kg_index_for_chunk(index_session, db_chunk)
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/rag/build_index.py", line 100, in build_kg_index_for_chunk
graph_index.insert_nodes([node])
File "/Users/wanzheng/src/ai_code/autoflow/auto4/lib/python3.11/site-packages/llama_index/core/indices/base.py", line 203, in insert_nodes
self._insert(nodes, **insert_kwargs)
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/rag/knowledge_graph/base.py", line 205, in _insert
self._insert_nodes(nodes)
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/rag/knowledge_graph/base.py", line 135, in _insert_nodes
entities_df, rel_df = extractor.extract(
^^^^^^^^^^^^^^^^^^
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/rag/knowledge_graph/extractor.py", line 161, in extract
pred = self.extract_prog(text=text)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/wanzheng/src/ai_code/autoflow/auto4/lib/python3.11/site-packages/dspy/primitives/program.py", line 26, in call
return self.forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/wanzheng/src/ai_code/autoflow/backend/app/rag/knowledge_graph/extractor.py", line 123, in forward
pred_graph = self.prog_graph(
^^^^^^^^^^^^^^^^
File "/Users/wanzheng/src/ai_code/autoflow/auto4/lib/python3.11/site-packages/dspy/primitives/program.py", line 26, in call
return self.forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/wanzheng/src/ai_code/autoflow/auto4/lib/python3.11/site-packages/dspy/functional/functional.py", line 369, in forward
raise ValueError(
ValueError: ('Too many retries trying to get the correct output format. Try simplifying the requirements.', {'knowledge': 'Input should be an object: entities, 0, metadata (error type: dict_type); Input should be an object: entities, 1, metadata (error type: dict_type); Input should be an object: entities, 2, metadata (error type: dict_type); Input should be an object: entities, 3, metadata (error type: dict_type); Input should be an object: entities, 4, metadata (error type: dict_type); Input should be an object: entities, 5, metadata (error type: dict_type); Input should be an object: entities, 6, metadata (error type: dict_type); Input should be an object: entities, 7, metadata (error type: dict_type); Input should be an object: entities, 8, metadata (error type: dict_type); Input should be an object: entities, 9, metadata (error type: dict_type); Input should be an object: entities, 10, metadata (error type: dict_type); Input should be an object: entities, 11, metadata (error type: dict_type); Input should be an object: entities, 12, metadata (error type: dict_type); Input should be an object: entities, 13, metadata (error type: dict_type); Input should be an object: entities, 14, metadata (error type: dict_type); Input should be an object: entities, 15, metadata (error type: dict_type)'})

[zhengwanbo]

use v0.4.1, any one can help this. thanks.