trim input to TGI, moved clustering and summarization to dataprep and store in DB

[rbrugaro]

Several enhancements:

1. trim input to TGI in cases where all community partial answers do not fit in the input context of the final answer generation -> this was causing error
2. In previous implementation clustering and summary extraction was done at query time resulting in slow time to fist token. Moved clustering a full dataset summariization to the dataprep step. In addition to storing the graph in Neo4j now we also store the entity_info and the community_summaries for retrieval with cypher queries in retriever code
3. fix gateway input

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
comps/cores/mega/gateway.py	0.00%	1 Missing ⚠️

Files with missing lines	Coverage Δ
comps/cores/mega/gateway.py	29.82% <0.00%> (ø)