fix(core)!: flatten generations for LangChainTracer

[hntrl]

(breaking - do not merge)

ref langchain-ai/langchainjs#9688

• Extracts and aggregates usage_metadata from generations, storing it in run.extra["metadata"]
• Flattens output structure when there's a single generation with a message, simplifying traced output

Example

Before (nested structure):

run.outputs = {
    "generations": [
        [
            {
                "text": "Hello!",
                "message": {
                    "content": "Hello!",
                    "usage_metadata": {"input_tokens": 10, "output_tokens": 5}
                }
            }
        ]
    ]
}

After (flattened):

run.outputs = {
    "content": "Hello!",
    "usage_metadata": {"input_tokens": 10, "output_tokens": 5}
}

run.extra["metadata"]["usage_metadata"] = {"input_tokens": 10, "output_tokens": 5}

Flattening only occurs for single-generation outputs; multiple generations/batches retain the original nested structure.

[codspeed-hq]

CodSpeed Performance Report

Merging #34415 will improve performances by 14.91%

_{Comparing hunter/flatten-lct-generations (8178e9b) with hunter/usage-metadata-extras (9511665)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

⚡ 5 improvements
✅ 8 untouched
⏩ 21 skipped¹

Benchmarks breakdown

	Mode	Benchmark	BASE	HEAD	Change
⚡	WallTime	test_import_time[PydanticOutputParser]	597.5 ms	520 ms	+14.91%
⚡	WallTime	test_import_time[RunnableLambda]	545.1 ms	481.7 ms	+13.17%
⚡	WallTime	test_import_time[Runnable]	538.9 ms	486.6 ms	+10.77%
⚡	WallTime	test_import_time[CallbackManager]	513.1 ms	453 ms	+13.27%
⚡	WallTime	test_import_time[tool]	574.3 ms	509.6 ms	+12.7%

Footnotes

1. 21 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[hntrl]

Punting to v2

The base branch was changed.