[Enhancement] TigerRAG: support additional data format and VectorDB.

[lilinwang]

Currently, only csv is supported in TigerRAG:

tiger/TigerRag/tigerrag/demo/movie_recs/utils.py

Line 20 in 8ad8d38

labels_df = pd.read_csv(labels_path)

. It could be useful to support other data format, like txt, jsonl, connect with VectorDB, etc.

[tigerlab-ai]

Thoughts about pinecone integration: Create a PineconeVectorStore class, which supports 4 operations:

1. Initialize and create an instance of PineconeVectorStore, with API key, environment, index_name, optional dimension, optional similarity search metric:
   pinecone.init(api_key="YOUR_API_KEY", environment="YOUR_ENVIRONMENT")
   pinecone.create_index("quickstart", dimension=8, metric="euclidean")
   index = pinecone.Index("quickstart")
2. Insert data: https://docs.pinecone.io/docs/quickstart#5-insert-vectors
3. Query: https://docs.pinecone.io/docs/quickstart#7-run-a-similarity-search, which is similar to https://github.com/tigerlab-ai/tiger/blob/main/TigerRag/demos/movie_recs/demo_rag.py#L39C37-L39C37.