I am working with the Graphrag system and encountered an issue where the create_final_nodes.parquet file is not being generated during the workflow execution. The process for create_base_text_units.parquet is successful, but the pipeline fails when it reaches the create_final_nodes step, citing a missing file error,or I miss something?
Reproduction steps
Go to the Graphrag project folder.
Set up a CSV file for indexing.
Execute the pipeline with the default workflows, including the create_base_text_units, create_base_entity_graph, and create_final_nodes.
The pipeline successfully creates the create_base_text_units.parquet, but fails to create the create_final_nodes.parquet with a FileNotFoundError.
Screenshots
![DESCRIPTION](LINK.png)
Logs
FileNotFoundError: [Errno 2] No such file or directory: 'F:\\kotaemon-app\\ktem_app_data\\user_data\\files\\graphrag\\output\\create_final_nodes.parquet'
16:30:55,553 graphrag.index.cli INFO Logging enabled at F:\kotaemon-app\ktem_app_data\user_data\files\graphrag\output\20240904-163055\reports\indexing-engine.log
...
16:30:55,857 graphrag.index.emit.parquet_table_emitter INFO emitting parquet table create_base_text_units.parquet
Browsers
No response
OS
Windows
Additional information
Embeddings and LLM models are configured correctly in GraphRagConfig.
No relevant workflows were skipped in the configuration.
The process was run using the default pipeline configuration, with all dependencies installed as required.
Description
I am working with the Graphrag system and encountered an issue where the create_final_nodes.parquet file is not being generated during the workflow execution. The process for create_base_text_units.parquet is successful, but the pipeline fails when it reaches the create_final_nodes step, citing a missing file error,or I miss something?
Reproduction steps
Screenshots
Logs
Browsers
No response
OS
Windows
Additional information
Embeddings and LLM models are configured correctly in GraphRagConfig. No relevant workflows were skipped in the configuration. The process was run using the default pipeline configuration, with all dependencies installed as required.