Publish reproducible PharmaPy documentation

[bernalde]

Summary

• define a reusable docs dependency extra and install it from Read the Docs and GitHub Actions
• make the complete Sphinx site build reproducibly with warnings treated as errors, including notebooks and their Pandoc/image requirements
• preserve the original PharmaPy credits while adding the current developers, maintainers, and joint-project partners
• update package metadata, installation links, and the CI badge for PharmaPy-org/PharmaPy

Verification

• sphinx-build -E -a -W --keep-going -b html doc/online_docs /tmp/pharmapy-docs-131-site
• actionlint .github/workflows/docs.yml .github/workflows/ci.yml
• python -m pip check in a fresh Python 3.11 docs environment installed with .[docs]
• python -m compileall -q PharmaPy
• python -m pytest tests/ -m "not assimulo" (56 passed, 4 skipped, 6 deselected)

Read the Docs follow-up

This PR makes the repository build-ready, but it intentionally does not close
the issue. After merge, the preferred publication path is a separate Read the
Docs project connected to PharmaPy-org/PharmaPy, leaving the legacy
pharmapy.readthedocs.io site untouched. Issue #130 remains open until that
external project is registered and its live site, push builds, and pull-request
previews are verified.

Refs #130

[bernalde]

Review requested from the full current team. GitHub has formally notified the active non-author members.

@ZacharyHillman18 @Piyush1698 @QHuckaba — your PharmaPy-org and repository invitations are still pending, so GitHub cannot add you as formal requested reviewers yet. Please accept those invitations and review this PR; after acceptance, you can also be added to the formal reviewer list.

[bernalde]

Maintainer review of the documentation-publishing PR (17 files: docs infra, Sphinx config, [docs] extra, RTD/CI workflows, attribution, metadata). I authored this PR, so it is posted as a COMMENT — GitHub does not accept an author APPROVE/REQUEST_CHANGES. No blocking findings.

What I verified

• File set matches gh pr diff (17 files); merge-base diff confirmed.
• Python changes are docstring-only and accurate. Verified against the live signatures at head:
  • RxnKinetics: rename to stoich_matrix matches the real parameter (Kinetics.py:214); the old docstring's stoiciometric_matrix/stoichiometric_matrix spellings were wrong.
  • BatchReactor.solve_unit: removed timesim_limit (not a parameter — confirmed absent) and documented the real params_control/sundials_opts; converted to NumPy-style Parameters/Returns.
  • Evaporator/ContinuousEvaporator: **fig_kwargs → fig_kwargs is a numpydoc rendering fix; signatures unchanged.
• Assets referenced by conf.py are tracked (images/PharmaPy_logo.jpeg, images/purdue_logo.png, images/fda_logo.png, references.bib), so the broad .gitignore *.png/*.jpg rule does not strand them (git ignore does not affect already-tracked files).
• Workflow docs.yml uses least-privilege permissions: contents: read, concurrency with cancel-in-progress, builds to site/ (which the PR adds to .gitignore), and does not deploy — consistent with RTD owning publication. No generated HTML committed.
• .readthedocs.yaml installs .[docs], fail_on_warning: true matches the CI -W build.
• Attribution (doc/online_docs/index.rst) preserves Original developers, adds Current developers and maintainers, and the joint-project acknowledgement (CryPTSys, SECQUOIA, Can Li, Eli Lilly) — matches #130's requirement.
• Metadata URLs/maintainers moved to PharmaPy-org.

Linked issue (#130)

Correctly Refs #130, not Closes — the RTD reconnection/live-site verification is an external admin step the PR intentionally leaves open. All repository-side acceptance criteria (docs extra, RTD config, warning-as-error CI, attribution, metadata/URLs, README badge, no committed HTML) are satisfied.

Nonblocking

• Incomplete stoich_matrix rename in the same docstring. See inline.
• README downloads badge still points to CryPTSys/PharmaPy while the CI badge and all URLs moved to PharmaPy-org. If historical download counts on the upstream repo are intended, fine; otherwise it is inconsistent with the rest of the metadata move.
• Docs Python version differs between environments: .readthedocs.yaml builds on 3.10, .github/workflows/docs.yml on 3.11. Both build the same docs and CI is green; aligning them (or documenting the intent) would make the "reproducible build" claim tighter.
• .gitignore image foot-gun: the broad *.png/*.jpg ignore required the manual !doc/online_docs/examples/img/pfd_flowsheet.png exception, and every future committed doc image will need another. Consider un-ignoring the doc image directories (e.g. !doc/online_docs/**) instead of per-file exceptions. Not breaking today since existing assets are tracked.

Merge-readiness

No blocking findings. Live state at review time: state=OPEN, mergeStateStatus=CLEAN, reviewDecision empty. CI at head 41304fd: Core tests SUCCESS, Sphinx docs SUCCESS, Assimulo integration tests SUCCESS. Posting this review adds inline threads; since mergeStateStatus is CLEAN, required_conversation_resolution is not enforced, so they are not a merge gate. This author account cannot supply a formal APPROVE — GitHub currently reports no approval gate, but if branch protection requires one, another eligible maintainer must approve. Per #130, the issue should remain open until the external RTD connection and live-site check are done.

[bernalde]

Nonblocking: this rename to stoich_matrix is correct (it matches the real __init__ parameter), but the pass is incomplete — the partic_species description a few lines below still refers to the old name: "the order of the names in 'partic_species' is that of the columns of 'stoichiometric_matrix'." Since the parameter is now stoich_matrix, update that prose reference too. While here, the keq_params entry is still the autogenerated TYPE, optional / DESCRIPTION. placeholder — worth filling in given this docstring is already being cleaned up. Neither blocks merge.

[bernalde]

Addressed in 9129d81: the participating-species prose now references stoich_matrix, and keq_params is documented as the dimensionless per-reaction equilibrium constant at tref_hrxn that enables reverse rates. The warning-as-error Sphinx build and all current-head CI jobs pass.

[bernalde]

Addressed the review feedback on head 9129d8142238bbec6c566c8437b0e0d23f5029c0.

• Commit pushed: 9129d81 (docs: address publishing review feedback).
• Inline finding kinetics-stale-stoich-name: completed the stoich_matrix rename and documented keq_params as dimensionless equilibrium constants at tref_hrxn.
• Body findings: aligned RTD with docs CI on Python 3.11; generalized the documentation PNG/JPG ignore exceptions; retained the only existing CryPTSys release count but labeled and linked it explicitly as historical.
• Verification: warning-as-error Sphinx build passed in a fresh Python 3.11 environment; python -m pip check reported no broken requirements; python -m compileall -q PharmaPy passed; core pytest passed with 56 passed, 4 skipped, and 6 deselected; actionlint passed; post-push Core, Sphinx docs, and Assimulo CI all passed.
• Intentionally not addressed: none.
• Remaining state: reviewDecision is empty. The inline thread remains for a maintainer to resolve. The parallel Read the Docs project remains an authorized post-merge external step tracked by Publish PharmaPy-org documentation and refresh developer attribution #130.