add: Pathway vector store and retriever as LangChain component

[berkecanrizai]

Description

Adds Pathway Vector Store as a vector DB option.

Note that Pathway expects text documents for the retrieval, thus I added the embedder (both the local HF model and the HuggingFaceHubEmbeddings(model=tei_embedding_endpoint)) to the vector store initialization.

Another additional note, Pathway connects to the input sources from the start, and does not allow upsert type operations via client. For now, I set the data source as the local data folder. It is possible to use any other data source, check out here.

Issues

Type of change

List the type of change like below. Please delete options that are not relevant.

• New feature (non-breaking change which adds new functionality)
  Pathway vector DB and retriever.

Dependencies

Pathway is needed if server is being started outside of the Docker.
Retriever and the other services do not need any change in the requirements.

Tests

[ashahba]

Thanks @berkecanrizai for the contribution.
Most of my change requests are repeated within the docs, so it should be easy to make the changes. However if you need any help, please feel free to post here.

Thanks.

[berkecanrizai]

Thanks @berkecanrizai for the contribution. Most of my change requests are repeated within the docs, so it should be easy to make the changes. However if you need any help, please feel free to post here.

Thanks.

Hey, I think I resolved your comments above, thanks for the clear instructions!

[chensuyue]

Please click Resolve conversation when it's done.

[ashahba]

Thanks @berkecanrizai
Nice work and LGTM!

Please make sure all other reviewers' feedbacks are resolved before merging.