[langchain-mongodb-deepagents-vfs] First release v0.1.0

XMLWordPrintableJSON

    • None
    • Python Drivers
    • None
    • None
    • None
    • None
    • None
    • None

      Goal

      First public release of langchain-mongodb-deepagents-vfs (version 0.1.0) to PyPI, following RELEASE.md.

      This is the monorepo's fourth library and its first release of a brand-new PyPI project name, so two prerequisites apply that RELEASE.md does not cover. See "First-release prerequisites" below.

      Blocked by

      • PR #415 — adds the package. Must merge first.

      Current state (verified on the PR merge commit)

      Item Status
      version in pyproject.toml already 0.1.0
      CHANGELOG.md 0.1.0 entry present, but no release date
      _release.yml dropdown entry added by #415
      check_diff.py LIB_DIRS added by #415
      Root pyproject.toml [tool.uv.sources] added by #415
      PyPI trusted publisher not configured — new project name
      JIRA release version does not exist yet
      Python 3.10 CI matrix unresolved, see #415

      First-release prerequisites (not in RELEASE.md)

      1. PyPI + TestPyPI pending publisher. The release workflow publishes via OIDC trusted publishing (id-token: write, pypa/gh-action-pypi-publish). For an existing project the publisher is already registered, but langchain-mongodb-deepagents-vfs does not exist on either index yet. A pending publisher must be created manually on both PyPI and TestPyPI before the workflow runs, or the publish step fails after a successful build. Values: owner langchain-ai, repo langchain-mongodb, workflow _release.yml.
      2. Create the JIRA release version. RELEASE.md step 1 assumes the version already exists on the INTPYTHON release page. Create it first.

      Steps

      1. Resolve the Python 3.10 CI matrix question (see Decisions) and merge PR #415.
      2. Create the JIRA release version on the INTPYTHON release page, and confirm no unfinished tickets are assigned to it.
      3. Configure the pending publisher on PyPI and TestPyPI (above).
      4. Open a PR adding the release date to the 0.1.0 heading in libs/langchain-mongodb-deepagents-vfs/CHANGELOG.md. No version bump needed — pyproject.toml is already at 0.1.0. # Merge that PR.
      5. Run the release workflow, selecting libs/langchain-mongodb-deepagents-vfs. It builds, publishes to TestPyPI, runs pre-release checks, publishes to PyPI, tags, and creates the GitHub Release.
      6. Mark the JIRA release as Released with today's date. # Verify pip install langchain-mongodb-deepagents-vfs works from a clean venv on Python 3.11 and 3.13.
      7. Announce on Slack, e.g. "ANN: langchain-mongodb-deepagents-vfs 0.1.0 — MongoDB Atlas-backed virtual filesystem for DeepAgents."

        Acceptance criteria

      • langchain-mongodb-deepagents-vfs 0.1.0 is installable from PyPI. * Git tag and GitHub Release exist.
      • CHANGELOG.md 0.1.0 entry carries the release date. * The INTPYTHON release version is marked Released.
      • Release announced on Slack.

        Decisions needed

      1. Python 3.10 CI matrix. This package requires >=3.11 because deepagents does; the shared matrix in _test.yml / _lint.yml is hardcoded to ["3.10", "3.13"] for every lib, so the 3.10 job fails at install. Options: (a) per-library matrix, testing this lib on 3.11+ only; (b) drop 3.10 repo-wide. Repo-wide policy call — must be
      2. JIRA release version name. Sibling libraries share the INTPYTHON project. Options: (a) per-library versions, e.g. langchain-mongodb-deepagents-vfs 0.1.0; (b) fold into an existing dated release version. Pick whichever matches current practice.
      3. CHANGELOG heading style. This package uses ## 0.1.0; siblings use ## Changes in version X (date). Align now or leave as-is.
      4. Ship py.typed? Absent, matching langgraph-store-mongodb. The package is mypy --strict cl and would export type info to consumers. Include in 0.1.0 or defer.

      Notes

      • requires-python = ">=3.11", unlike the other three libraries.
      • First release of a package whose search layer depends on Atlas Search and Vector Search; non-Atlas deployments fall back to regex. Worth stating in release notes so users don't expect hybrid search on a community server.

            Assignee:
            Unassigned
            Reporter:
            Casey Clements
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated: