[BugFix] Fix issues related to data retrieval logic, parameter validation, and result serialization in both online and offline interfaces.

[qwes5s5]

Motivation

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick，PR标题需遵循格式，在最开始加上[Cherry-Pick]标签，以及最后面加上原PR ID，例如[Cherry-Pick][CI] Add check trigger and logic(#5191)
在pr https://github.com/PaddlePaddle/FastDeploy/pull/5089合入之后，仍然存在部分问题亟待解决：

• When the environment variable FD_USE_GET_SAVE_OUTPUT_V1 is enabled and the service does not support enable-logprob: The online service lacks specific error interception for cases where prompt_logprobs and logprobs are zero, only returning null. The offline service, however, has no specific interception for either prompt_logprobs or logprobs.
• When the environment variable FD_USE_GET_SAVE_OUTPUT_V1 is disabled: The error interception for logprobs and top_logprobs based on max_logprobs is incomplete and the suggested range in the error message is inappropriate.
• In the relevant error messages for online inference, logprobs is used as the subject, whereas the chat interface utilizes top_logprobs.
• During offline inference, the results when logprobs=0 are empty/null, and there is an issue with the data retrieval logic.
• The output format for prompt_logprobs in both offline and online inference is a list instead of a dictionary (dict).

Modifications

• Modified the parameter validation logic when FD_USE_GET_SAVE_OUTPUT_V1 is enabled and disabled, and added the missing error interceptions.
• Updated the subject of the error message based on the interface type (e.g., changing from logprobs to top_logprobs where appropriate).
• Changed the construction method for the Logprob type: it no longer inherits from NamedTuple but now uses @DataClass, allowing it to be correctly serialized as a dictionary (dict).
• Modified the data retrieval logic for offline inference, so that the result for logprobs=0 now displays correctly.

Usage or Command

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 17.14286% with 29 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@a4bb3e9). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/entrypoints/engine_client.py	4.16%	20 Missing and 3 partials ⚠️
fastdeploy/entrypoints/llm.py	25.00%	2 Missing and 4 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #5335   +/-   ##
==========================================
  Coverage           ?   59.31%           
==========================================
  Files              ?      324           
  Lines              ?    40066           
  Branches           ?     6055           
==========================================
  Hits               ?    23767           
  Misses             ?    14409           
  Partials           ?     1890           

Flag	Coverage Δ
GPU	59.31% <17.14%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[paddle-bot]

Thanks for your contribution!