Support Long context for DocSum

[XinyaoWa]

Description

Support Long context for DocSum with five modes:
• Auto(Default mode): switch to stuff mode if input token < max_input_token, otherwise switch to "refine" mode
• Stuff (Default mode): input actual tokens, need to increase the max_input_token if want to use large context
• Truncate: truncate the tokens exceed the limitation
• Map_reduce: split the inputs into multiple chunks, map each document to an individual summary, then consolidate those summaries into a single global summary
• Refine: split the inputs into multiple chunks, generate summary for the first one, then combine with the second, loops over every remaining chunks to get the final summary

Related PR: opea-project/GenAIComps#981 opea-project/GenAIComps#1046

Issues

List the issue or RFC link this PR is working on. If there is no such link, please mark it as n/a.

Type of change

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

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds new functionality)
• Breaking change (fix or feature that would break existing design and interface)
• Others (enhancement, documentation, validation, etc.)

Dependencies

List the newly introduced 3rd party dependency if exists.

Tests

Describe the tests that you ran to verify your changes.

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues found.

Scanned Files

[eero-t]

@lvliang-intel Dockerfile check fails to a pre-existing "MultimodalQnA" docs issue, this PR does not touch them:

Missing Dockerfile: GenAIComps/comps/retrievers/multimodal/redis/langchain/Dockerfile (Referenced in GenAIExamples/./MultimodalQnA/docker_compose/intel/cpu/xeon/README.md:127)
Missing Dockerfile: GenAIComps/comps/retrievers/multimodal/redis/langchain/Dockerfile (Referenced in GenAIExamples/./MultimodalQnA/docker_compose/intel/hpu/gaudi/README.md:78)
Error: Process completed with exit code 1.

"DocSum, xeon" CI test failed to:

...
[ docsum-xeon-backend-server ] Content is as expected.
...
[ docsum-gaudi-backend-server ] Content is as expected.
...
[ docsum-gaudi-backend-server ] Content is as expected.
...
[ docsum-gaudi-backend-server ] Content does not match the expected result: 
Error response from daemon: No such container: docsum-gaudi-backend-server

Which is rather surprising when test is supposed to run on Xeon, and there's a separate test that runs same things.

"DocSum, gaudi" test fails due to backend error exit:

...
[ docsum-gaudi-backend-server ] Content is as expected.
...
[ docsum-gaudi-backend-server ] Content is as expected.
...
[ docsum-gaudi-backend-server ] Content does not match the expected result: 
 Error: Process completed with exit code 1.

"DocSum, rocm" test fails to something that looks like CI issue:
aiohttp.client_exceptions.ClientConnectorError: Cannot connect to host 0.0.0.0:7079 ssl:default [Connect call failed ('0.0.0.0', 7079)]

[XinyaoWa]

CICD pending for this PR: opea-project/GenAIComps#1046

[mkbhanda]

@XinyaoWa thank you for this PR! I am curious on a couple of issues. 1) what happens with map-reduce mode if the number of summaries totals an input token length greater than max-input-length .. hierarchical map-reduce or refine? 2) Also if possible would you explain the formulas used, 50 etc. 3) Are there speed/accuracy trade-offs to using smaller chunks?