[FEATURE] Add community level as additional parameter for global search

[rm-1997]

In the file backend/src/api/query.py, the community level, which is used behind the scences for the global search, is hardcoded to 1 (same for query_streaming.py):

# current investigations show that community level 1 is the most useful for global search
COMMUNITY_LEVEL = 1

I think it would be of great benefit if one could parse the desired community level for global search as additional parameter in the API (i.e. the global_search function from the notebooks respectively). Acknowledging that a community level of 1 might be the most promising level for many applications, I think this feature could really increase my understanding of how the community summarization works and also potentially increase the power of the global search for my given application.

[rm-1997]

Interesting enough, so far it is also not possible to pass a claim extraction prompt to the API; or at least the build_index function in the starter notebooks does not provide any functionality here. Is this done deliberately and if so, why? Is claim extraction prompting more vulnerable than the others perhaps?

[rm-1997]

Also note that the claim extraction prompt from the graphrag library uses some default parameters entity_specs which evaluates to ['organization', 'person', 'geo', 'event'] in this repo. Extract from the claim extraction cache:

The company is owned by Person C who was suspected of engaging in corruption activities in 2015)\n<|COMPLETE|>\n\n-Real Data-\nUse the following input for your answer.\nEntity specification: ['organization', 'person', 'geo', 'event']\nClaim description: Any claims or facts that could be relevant to information discovery.\nText:

Am I right in assuming that I can only configure this parameter in the initial deployment? This would limit of course its applicability since different indexes might aim at different entity types.