feat: delegate ecModel YAML I/O to raven-python#19
Merged
Conversation
…p SBML Mirrors MATLAB GECKO + RAVEN, where RAVEN's readYAMLmodel.m / writeYAMLmodel.m populate and serialise the model.ec struct directly and GECKO is just a consumer. The schema for ec-rxns / ec-enzymes / gecko_light, the parsing, the legacy-format normalisations (top-level smiles -> annotation, reverse-direction usage_prot_*, bare-`-` document root) all live in raven-python now (see raven_python.io.ec_data); geckopy keeps the algorithms that operate on top. geckopy.EcData is now a re-export of raven_python.io.EcData. Existing `from geckopy import EcData` keeps working unchanged; the dataclass gained EcData.empty (was empty_for_reactions) and kept .validate. save_ec_model / load_ec_model become thin wrappers (~30 LOC each) that add three application-level concerns on top of raven-python: 1. File-extension validation (YAML only; .xml/.sbml raises a helpful "SBML ecModel I/O has been removed" message). 2. Adapter-aware path resolution under <adapter.params.path>/models/. 3. Provenance + diagnostics: metaData injection on save (date, geckopy version, MATLAB-compat description); fast-fail messages on empty ec or non-ecModel YAML. Removals: - src/geckopy/io/sbml.py (437 LOC) — entire ecModel SBML I/O path. Per `cobra.io.read_sbml_model` is still used for loading conventional starting GEMs; only the ecModel-specific encode/decode is gone. - src/geckopy/io/ (whole package) — only contained sbml.py. - tests/test_sbml_io.py — covered the removed surface. - tests/test_ec_data.py — EcData validation tests now live in raven-python's test_io_yaml_ec_data.py (covered by the upstream PR #12). - The big libsbml-appendNotes-silently-fails gotcha in docs/future_improvements.md (no longer relevant; no libsbml use in geckopy proper). The Windows/Mac `e-005` -> `e-05 ` MATLAB SBML post-processing item likewise dropped. Test renames: - tests/test_ec_model.py: EcData.empty_for_reactions -> EcData.empty. New tests: - test_load_ec_model.test_sbml_extension_raises_with_helpful_message - test_save_ec_model.test_sbml_extension_raises_with_helpful_message Updated tests: - test_load_ec_model.test_unknown_extension_raises: new "YAML only" message. - test_load_ec_model.test_legacy_sequence_top_level_is_merged: now builds a complete bare-`-` doc from the canonical fixture and asserts the load succeeds (proving the merge in raven-python's reader is reached through the geckopy wrapper). - test_save_ec_model.test_unknown_extension_raises: same message update. Dependency: - raven-python pin moved from @main to @develop. raven-python's main branch is currently just the initial commit; all active development lives on develop, including the typed model.ec / EcData feature this commit depends on. README and pyproject pins updated. Docs updated: - raven_integration.md: rewritten to describe the new clean split (raven-python owns YAML+schema, geckopy owns three application concerns). - yaml_format.md: ec-* keys now listed as raven-python-owned. - README: drop SBML mention, fix pin. Full suite: 1254 passed, 1 pre-existing solver-state flake unrelated to this change. Depends on raven-python PR #12 (feat: typed model.ec via EcData) landing on develop before geckopy install resolves the pin.
edkerk
force-pushed
the
feat/yaml-delegate-to-raven-python
branch
from
6b679a3 to
4e422a9
Compare
edkerk
added a commit
that referenced
this pull request
May 30, 2026
edkerk
added a commit
that referenced
this pull request
May 30, 2026
Roll-up commit for the second alpha. Bumps the version, adds the CHANGELOG entry, ships the MIT LICENSE text the pyproject already declared, drops the redundant requirements.txt, and tucks internal planning notes under docs/internal/ so the user-facing docs/ tree only carries reference documentation. Highlights covered by [0.1.0a2] (see CHANGELOG.md for full notes): - gecko-light ecModel end-to-end (PR #20). - YAML I/O + typed EcData delegated to raven-python; SBML ecModel I/O removed (PR #19). - ec_fseof re-aligned with raven_python.analysis.fseof (PR #21). - get_conc_control_coeffs solver-state guard fix. - raven-python pin moved from @main to @develop. Repo housekeeping: - LICENSE: MIT text added at repo root (the classifier and the pyproject license field already declared MIT, but no LICENSE file shipped previously — blocks GitHub repo metadata and any future PyPI publish). - requirements.txt: deleted. It was a pip-freeze snapshot duplicating pyproject.toml's declared dependencies; pyproject is the source of truth. - docs/internal/: brenda_refresh_plan, code_review, openkineticspredictor_plan, porting_plan, raven_inventory moved out of the user-facing docs/ tree. Two source-file docstrings updated to point at the new location. - pyproject.toml: version 0.1.0a1 -> 0.1.0a2. Release smoke: - Full pytest suite green: 1259 passed, 1 skipped (network), 1 deselected (smoke), 1 xfailed (documented). - Light tutorial runs end-to-end against ecTestGEM. - Both tutorial scripts compile cleanly. The full tutorial's end-to-end run is exercised piecewise by the unit suite; a live BRENDA + DLKcat + sensitivity-tuning replay was skipped for the release smoke (too long, no signal beyond the unit tests).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
save_ec_modelandload_ec_modelnow delegate the cobra-side round-trip and the opaque preservation of GECKO top-level keys (ec-rxns,ec-enzymes,gecko_light,metaData) toraven_python.io.yaml. geckopy still owns the typedEcDatainterpretation, the legacy MATLAB ec-model normalisation, and the protein-direction flip for older models with reverse-signusage_prot_*reactions.What moved to raven-python
!!omapblocks for reactions/metabolites/genes/compartments).model.notes["_yaml_sections"]— so geckopy's GECKO blocks land at the document root, not nested insidenotes.Depends on raven-python PR #11 (merged).
What geckopy still owns
save_ec_model: buildsec-rxns/ec-enzymes/gecko_lightfromEcData(omitting emptysource/notes/eccodes,kcat==0→ 'no kcat assigned', omitting NaNmw/concsand emptysequence), pre-coerces numpy/ruamel scalars via_to_native, stashes everything onmodel.noteswithin a try/finally so the caller's model is restored after the write, then delegates toraven_python.io.yaml.write_yaml_model.load_ec_model: keeps its own_read_yaml(handles legacy bare--sequence-of-single-key-maps), applies_normalize_legacy_layout(metaData lifting + smiles → annotation), hands the cleaned dict to raven-python'smodel_from_yaml_data, then readsec-*blocks back frommodel.notes["_yaml_sections"]into a typedEcData.Tests
Full test suite passes locally. Round-trip semantics covered by the existing
test_save_load_ec_model.pyandtest_yaml_roundtrip.pysuites.Migration note
Users now need
raven-pythonavailable; install via:(Already listed in
pyproject.tomlfrom the earlier delegation work.)