only can you share exact .txt file you used

[Sandy4321]

great code thanks only can you share exact .txt file you used in with open("./ragtest/input/ft_guide.txt", 'r') as file:

to reproduce your results ?

[Sandy4321]

also entities = pd.read_parquet('./ragtest/output/create_final_entities.parquet')

[sandman]

@Sandy4321 Just rename the parquet files in the output directory: create_final_entities.parquet --> entities.parquet create_final_relationships.parquet --> relationships.parquet

This is assuming you have run the graphrag tool on the text file. Also, if you are missing some of the create_XXX files, it means the graphrag tool didn't generate the full output due to token rate limits. One solution is to use a self-hosted LLM if you have a beefy machine to run graphrag.