Skip to content

chore: datafactory install pins → PyPI floor (views-datafactory>=1.9.0) - #283

Open
Polichinel wants to merge 1 commit into
developmentfrom
chore/datafactory-pypi-pins
Open

chore: datafactory install pins → PyPI floor (views-datafactory>=1.9.0)#283
Polichinel wants to merge 1 commit into
developmentfrom
chore/datafactory-pypi-pins

Conversation

@Polichinel

Copy link
Copy Markdown
Collaborator

views-datafactory is public and on PyPI as of v1.9.0 — pip install views-datafactory.

  • 27 model requirements.txt: git+...@developmentviews-datafactory>=1.9.0 (matches the existing views-r2darts2>=0.1.0 house style in the same files). Fresh model envs no longer need GitHub access to resolve the dependency.
  • README: adds a datafactory-backend paragraph to the config_queryset.py section — the backend was documented in 12+ per-model READMEs but never in the front door — linking the new model consumer quickstart.
  • Register C-36 notes: remediation command updated to the PyPI pin.

Deliberately excluded: postprocessors/un_fao/* install lines — those files carry the in-flight land_gaul/feature_frame migration in the maintainer's working tree; fold the one-line pin change into that commit instead (requirements.txt:1, README.md:73, configs/config_queryset.py:21).

Part of the cross-repo tagged-release rollout (sibling PR: views-pipeline-core#315).

🤖 Generated with Claude Code

views-datafactory is public and on PyPI as of v1.9.0. House convention
is now release floor pins, never git+...@development:

- 27 model requirements.txt: git branch pin -> views-datafactory>=1.9.0
  (matches the existing views-r2darts2>=0.1.0 style in the same files)
- README: datafactory-backend paragraph in the config_queryset.py section
  (the backend existed in 12+ per-model READMEs but never in the front
  door) + link to the new model_consumer_quickstart guide
- register C-36 notes: remediation command updated to the PyPI pin

un_fao's install lines are deliberately NOT touched here — those files
carry an in-flight land_gaul/feature_frame migration in the maintainer's
working tree; the pin change should ride with that commit.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@Polichinel

Copy link
Copy Markdown
Collaborator Author

Review — sound, approve with one recommended change

Verified the claims against the repo:

  • views-datafactory 1.9.0 is live on PyPI (checked the PyPI API — latest release, present). Since CI and fresh envs install published packages, this is what makes the floor pin safe. Cleared.
  • Scope correct — 27 model requirements.txt (git+…@development → PyPI pin), README, register; un_fao deliberately excluded (confirmed — it still carries the git pin locally).
  • README front-door "Data backends" paragraph fills a real gap (datafactory backend was only ever documented in per-model READMEs).
  • Register update: the datafactory-readiness Fix: command correctly moves to the PyPI pin.

Recommended change before merge

The new pin is unbounded (views-datafactory>=1.9.0), unlike its siblings: views-hydranet>=0.1.0,<1.0.0 and views-r2darts2>=1.0.0,<2.0.0 both cap the major. Suggest views-datafactory>=1.9.0,<2.0.0 across the 27 files to match house style and prevent a future breaking 2.0 being pulled in silently. (Cosmetic follow-on: the PR body cites views-r2darts2>=0.1.0, but the actual r2darts pin is >=1.0.0,<2.0.0.)

The deferred un_fao pin is now handled

Per this PR's note, the un_fao datafactory pin (requirements.txt, config_queryset.py install string, README) has been folded into the in-flight land_gaul/feature_frame working-tree change, pinned to views-datafactory>=1.9.0,<2.0.0. So un_fao lands with the rest — just via that branch, not this PR.

Minor note

The diff edits 8 temporary_*/requirements.txt (rusty_bucket constituents / dwarf scaffolds). A requirements pin isn't the hp-config content parallel sessions tune, so fine — flagging as a conscious inclusion.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant