[performance]: Use hard coded langchain-core version to avoid importlib import

[sydney-runkle]

This PR aims to reduce import time of langchain-core tools by removing the importlib.metadata import previously used in __init__.py. This is the first in a sequence of PRs to reduce import time delays for langchain-core features and structures 🚀.

Because we're now hard coding the version, we need to make sure version.py and pyproject.toml stay in sync, so I've added a new CI job that runs whenever either of those files are modified. This run demonstrates the failure that occurs whenever the version gets out of sync (thus blocking a PR).

Before, note the ~15% of time spent on the importlib.metadata /related imports

After (note, lack of importlib.metadata time sink):

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Skipped Deployment

Name	Status	Preview	Comments	Updated (UTC)
langchain	⬜️ Ignored (Inspect)	Visit Preview		Apr 10, 2025 6:10pm

[sydney-runkle]

Waiting on merging this until we have benchmarks from #30747 in place so that we can track progress.

[codspeed-hq]

CodSpeed Performance Report

Merging #30744 will improve performances by 27.77%

_{Comparing hard-coded-version (02e7498) with master (cd6a831)}

Summary

⚡ 1 improvements
✅ 11 untouched benchmarks

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⚡	test_import_time[InMemoryRateLimiter]	273 ms	213.6 ms	+27.77%

[eyurtsev]

We need to generalize this I think. How about we do this as part of the release pipeline so it'll apply to all libraries. We can put logic in workflow to only apply it to libraries that we fixed already (or even skip and force all libraries to include version which I would prefer!)

[sydney-runkle]

Yeah, I think generalizing this would be good... maybe in a separate PR where we automate some of the manual and redundant release tasks (like version bumping, etc).

[Viicos]

Why not leverage dynamic versioning? https://backend.pdm-project.org/metadata/#dynamic-project-version

[sydney-runkle]

Good idea @Viicos, thanks! Will do :)

[sydney-runkle]

Yeah, maybe I should just do the version stuff across all of our libs at once then :)

[sydney-runkle]

@Viicos, explored dynamic versioning via #30771. pdm-backend has relatively limited support, and it breaks our release CI, so going to stick with keeping the version in pyproject.toml for now.

Maybe if we move to hatch, that switch would make sense.