sbmlutils 0.12.0__tar.gz → 0.12.2__tar.gz

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/.bumpversion.toml

@@ -1,5 +1,5 @@
 [tool.bumpversion]
-current_version = "0.12.0"
+current_version = "0.12.2"
 commit = true
 parse = "(?P<major>\\d+)\\.(?P<minor>\\d+)\\.(?P<patch>\\d+)"
 serialize = ["{major}.{minor}.{patch}"]

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/.claude/issues/followup-after-core-roundtrip.md

@@ -152,7 +152,7 @@ Round-tripping the MathML itself would remove all three and the constant shadowi
 
 ## 10. pymetadata: `resource_normalized` loses information, found by the structural harness of the package work
 
-Measured with pymetadata 0.6.2 (package branch, Task 1 fix round). sbmlutils guards against all of these from 0.12.0 on by writing the resource as given, so the fix belongs in pymetadata.
+Measured with pymetadata 0.6.2 (package branch, Task 1 fix round). sbmlutils guards against all of these from 0.12.0 on by writing the resource as given, so the fix belongs in pymetadata. Reported as [matthiaskoenig/pymetadata#81](https://github.com/matthiaskoenig/pymetadata/issues/81), reproduced with pymetadata 0.6.3.
 
 - **An unknown collection is reduced to the bare term:** `http://identifiers.org/sabiork/1406` becomes `1406`, `urn:miriam:foo:bar` becomes `bar`, `http://identifiers.org/unit/UO:0000040` becomes `UO:0000040`, `http://identifiers.org/ncbigi/gi:16128336` becomes `gi:16128336` (281 uses in the corpus, 138 distinct in `distrib/e_coli_core.xml` alone).
 - **A compact URL of a collection it does not know is reduced to the bare term:** `https://identifiers.org/CMO:0000012` becomes `CMO:0000012` (`icg_body.xml`).
@@ -164,15 +164,15 @@ The invariant which separates a canonicalization from a loss: `resource_normaliz
 
 ## 11. libsbml: the key-value pairs of a species reference are written and never read back
 
-Measured with python-libsbml 5.21.2 alone, no sbmlutils involved (package branch, fbc fixes). The fbc version 3 `listOfKeyValuePairs` of a `<speciesReference>` is written into the document, and reading that document again gives a species reference without any key-value pair; on a `<modifierSpeciesReference>` the same round trip works. sbmlutils writes the pairs of reactants, products and modifiers from 0.12.0 on, so a reactant or product pair is write-only for every libsbml reader, and the structural comparison cannot see it. A test pins the behaviour so that a fixed libsbml is noticed. To report upstream at [sbmlteam/libsbml](https://github.com/sbmlteam/libsbml).
+Measured with python-libsbml 5.21.2 alone, no sbmlutils involved (package branch, fbc fixes). The fbc version 3 `listOfKeyValuePairs` of a `<speciesReference>` is written into the document, and reading that document again gives a species reference without any key-value pair; on a `<modifierSpeciesReference>` the same round trip works. sbmlutils writes the pairs of reactants, products and modifiers from 0.12.0 on, so a reactant or product pair is write-only for every libsbml reader, and the structural comparison cannot see it. A test pins the behaviour so that a fixed libsbml is noticed. Reported as [sbmlteam/libsbml#493](https://github.com/sbmlteam/libsbml/issues/493).
 
 ## 12. libsbml: the core id and name of a comp reference are not written
 
-Measured with python-libsbml 5.21.2 alone (package branch, comp data model). The SBML Level 3 Version 2 core `id` and `name` set on a `ReplacedElement`, a `ReplacedBy` or a nested `sBaseRef` do not survive a write and re-read, nested or not; `metaid`, `sboTerm`, notes, annotation and the four reference attributes do. Measured again at Level 3 Version 1 and Version 2: below L3V2 both setters answer `LIBSBML_UNEXPECTED_ATTRIBUTE` and at L3V2 both answer success, and the written document carries neither attribute at either version, so no SBML version keeps them. A `Port` and a `Deletion` are not affected, comp gives each of them an `id` and a `name` of its own, written as `comp:id` and `comp:name` at both versions. sbmlutils no longer sets them, it reports a value given for either once per document, and the structural comparison reads both sides through libsbml and cannot see the loss. Documented on the `SbaseRef` docstring and on [Model composition](https://matthiaskoenig.github.io/sbmlutils/comp/#what-round-trips).
+Measured with python-libsbml 5.21.2 alone (package branch, comp data model). The SBML Level 3 Version 2 core `id` and `name` set on a `ReplacedElement`, a `ReplacedBy` or a nested `sBaseRef` do not survive a write and re-read, nested or not; `metaid`, `sboTerm`, notes, annotation and the four reference attributes do. Measured again at Level 3 Version 1 and Version 2: below L3V2 both setters answer `LIBSBML_UNEXPECTED_ATTRIBUTE` and at L3V2 both answer success, and the written document carries neither attribute at either version, so no SBML version keeps them. A `Port` and a `Deletion` are not affected, comp gives each of them an `id` and a `name` of its own, written as `comp:id` and `comp:name` at both versions. sbmlutils no longer sets them, it reports a value given for either once per document, and the structural comparison reads both sides through libsbml and cannot see the loss. Documented on the `SbaseRef` docstring and on [Model composition](https://matthiaskoenig.github.io/sbmlutils/comp/#what-round-trips). Reported as [sbmlteam/libsbml#494](https://github.com/sbmlteam/libsbml/issues/494): at Level 3 Version 2 the setters answer success and the object holds the value, which is then not written.
 
 ## 13. libsbml: fbc:strict cannot be written on a comp model definition
 
-Measured with python-libsbml 5.21.2 alone (package branch, `ModelDefinition` as a `Model`). Setting `strict` on the fbc plugin of a `<comp:modelDefinition>` writes the `fbc:strict` attribute twice, and the file then fails to parse with "Duplicate XML attribute". sbmlutils therefore does not write `fbc:strict` on a model definition, and a document whose model definition carries fbc content validates with libsbml error 2020209 (the required attribute is missing): a validation error which a reader survives, instead of a file which no reader can open. `comp/flatten.py` copy-constructs a `libsbml.ModelDefinition` and is affected by the same defect. To report upstream at [sbmlteam/libsbml](https://github.com/sbmlteam/libsbml).
+Measured with python-libsbml 5.21.2 alone (package branch, `ModelDefinition` as a `Model`). Setting `strict` on the fbc plugin of a `<comp:modelDefinition>` writes the `fbc:strict` attribute twice, and the file then fails to parse with "Duplicate XML attribute". sbmlutils therefore does not write `fbc:strict` on a model definition, and a document whose model definition carries fbc content validates with libsbml error 2020209 (the required attribute is missing): a validation error which a reader survives, instead of a file which no reader can open. `comp/flatten.py` copy-constructs a `libsbml.ModelDefinition` and is affected by the same defect. Reported as [sbmlteam/libsbml#495](https://github.com/sbmlteam/libsbml/issues/495).
 
 ## 14. What remains after the package round trip
 
@@ -208,4 +208,4 @@ Sections 1, 3 and 4 are done in 0.12.0. What was deliberately left out of that b
 
 ### Upstream reports
 
-Sections 10 to 13 are the reports this branch owes: `resource_normalized` of pymetadata, and three defects of python-libsbml 5.21.2, the key-value pairs of a `<speciesReference>`, the core `id` and `name` of a comp reference, and `fbc:strict` on a `<comp:modelDefinition>`. Each is measured with the library alone, without sbmlutils, and each is pinned by a test here, so a fixed version is noticed.
+Sections 10 to 13 are reported upstream, [pymetadata#81](https://github.com/matthiaskoenig/pymetadata/issues/81) and libsbml [#493](https://github.com/sbmlteam/libsbml/issues/493), [#494](https://github.com/sbmlteam/libsbml/issues/494) and [#495](https://github.com/sbmlteam/libsbml/issues/495): `resource_normalized` of pymetadata, and three defects of python-libsbml 5.21.2, the key-value pairs of a `<speciesReference>`, the core `id` and `name` of a comp reference, and `fbc:strict` on a `<comp:modelDefinition>`. Each is measured with the library alone, without sbmlutils, and each is pinned by a test here, so a fixed version is noticed.

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/.github/workflows/ci-cd.yml

@@ -27,7 +27,8 @@ jobs:
       matrix:
         os: [ubuntu-latest, windows-latest, macos-latest]
         python-version: ["3.11", "3.12", "3.13", "3.14"]
-        # linux covers the full range, windows and macos only the newest version
+        # Linux covers the full range with simulation support; Windows and
+        # macOS cover its newest version. Core-only 3.15 jobs are added below.
         exclude:
           - os: windows-latest
             python-version: "3.11"
@@ -41,6 +42,15 @@ jobs:
             python-version: "3.12"
           - os: macos-latest
             python-version: "3.13"
+        # Keep the full suite on 3.14 while libroadrunner lacks 3.15 wheels.
+        # uv currently provides the 3.15 macOS interpreter for Intel only.
+        include:
+          - os: ubuntu-latest
+            python-version: "3.15"
+          - os: windows-latest
+            python-version: "3.15"
+          - os: macos-15-intel
+            python-version: "3.15"
 
     steps:
     - uses: actions/checkout@v7

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/CITATION.cff

@@ -8,8 +8,8 @@ authors:
     given-names: "Matthias"
     orcid: "https://orcid.org/0000-0003-1725-179X"
     affiliation: "Humboldt-University Berlin, Faculty of Life Science, Institute for Biology, Biology, Berlin; University Lübeck; University Hospital Schleswig-Holstein, Campus Lübeck, First Department of Medicine, Germany"
-version: "0.12.0"
-date-released: "2026-09-07"
+version: "0.12.2"
+date-released: "2026-09-20"
 doi: "10.5281/zenodo.597149"
 license: MIT
 repository-code: "https://github.com/matthiaskoenig/sbmlutils"

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: sbmlutils
-Version: 0.12.0
+Version: 0.12.2
 Summary: sbmlutils are python utilities for working with SBML.
 Project-URL: Homepage, https://matthiaskoenig.github.io/sbmlutils
 Project-URL: Documentation, https://matthiaskoenig.github.io/sbmlutils
@@ -20,6 +20,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: 3.14
+Classifier: Programming Language :: Python :: 3.15
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Topic :: Scientific/Engineering
 Classifier: Topic :: Scientific/Engineering :: Bio-Informatics
@@ -34,7 +35,7 @@ Requires-Dist: openpyxl>=3.1.5
 Requires-Dist: pandas>=2.2.0
 Requires-Dist: pint>=0.25.3
 Requires-Dist: pydantic>=2.13.5
-Requires-Dist: pymetadata>=0.6.2
+Requires-Dist: pymetadata>=0.6.4
 Requires-Dist: python-libsbml>=5.21.2
 Requires-Dist: requests>=2.34.2
 Requires-Dist: rich>=15.0.0
@@ -92,7 +93,7 @@ If you have any questions or issues please [open an issue](https://github.com/ma
 
 If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi.org/10.5281/zenodo.597149):
 
-> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.10.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22646678
+> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.12.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22853867
 
 ```bibtex
 @software{konig_sbmlutils,
@@ -100,10 +101,10 @@ If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi
   title     = {sbmlutils: Python utilities for SBML},
   year      = {2026},
   month     = sep,
-  version   = {0.10.0},
+  version   = {0.12.0},
   publisher = {Zenodo},
-  doi       = {10.5281/zenodo.22646678},
-  url       = {https://doi.org/10.5281/zenodo.22646678},
+  doi       = {10.5281/zenodo.22853867},
+  url       = {https://doi.org/10.5281/zenodo.22853867},
 }
 ```
 

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/README.md

@@ -29,7 +29,7 @@ If you have any questions or issues please [open an issue](https://github.com/ma
 
 If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi.org/10.5281/zenodo.597149):
 
-> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.10.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22646678
+> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.12.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22853867
 
 ```bibtex
 @software{konig_sbmlutils,
@@ -37,10 +37,10 @@ If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi
   title     = {sbmlutils: Python utilities for SBML},
   year      = {2026},
   month     = sep,
-  version   = {0.10.0},
+  version   = {0.12.0},
   publisher = {Zenodo},
-  doi       = {10.5281/zenodo.22646678},
-  url       = {https://doi.org/10.5281/zenodo.22646678},
+  doi       = {10.5281/zenodo.22853867},
+  url       = {https://doi.org/10.5281/zenodo.22853867},
 }
 ```
 

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/docs/annotations.md

@@ -39,12 +39,14 @@ A resource is written as `collection/term` (`chebi/CHEBI:17234`), as an identifi
 
 ### Resources which are written as given
 
-The normalization keeps the collection and the term of a resource and only changes how they are spelled, so `urn:miriam:chebi:CHEBI%3A17234` and `http://identifiers.org/chebi/CHEBI:17234` are both written as `https://identifiers.org/CHEBI:17234`. For a collection which is not in the registry it cannot do that: `http://identifiers.org/sabiork/1406` would come out as the bare `1406`, which resolves to nothing and no longer says where the term comes from.
+The normalization keeps the collection and the term of a resource and only changes how they are spelled, so `urn:miriam:chebi:CHEBI%3A17234` and `http://identifiers.org/chebi/CHEBI:17234` are both written as `https://identifiers.org/CHEBI:17234`. A collection which is not in the registry cannot be written as a compact identifier and keeps the classic form: `http://identifiers.org/sabiork/1406` is written as `https://identifiers.org/sabiork/1406`.
 
-Such a resource is written exactly as it was given instead. Because this is a property of the collection rather than of a single resource, and a large model can hold tens of thousands of resources of one collection, it is reported once per collection when the document is written, with the count and one example:
+What cannot be normalized is a resource which is malformed to begin with. `urn:miriam:chebi` and `chebi/` name a collection and no term, so there is nothing to write a canonical resource from, and `urn:miriam::1406` names a term and no collection.
+
+Such a resource is written exactly as it was given instead. A large model can hold tens of thousands of resources of one collection, so this is reported once per collection when the document is written, with the count and one example:
 
 ```
-2 annotation resource(s) of the collection 'sabiork' are written as given, e.g. 'http://identifiers.org/sabiork/1406': pymetadata shortens the resource to a bare term, which names no collection.
+2 annotation resource(s) of the collection 'chebi' are written as given, e.g. 'urn:miriam:chebi': pymetadata parses no term from the resource.
 ```
 
 ## Qualifiers

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/docs/development.md

@@ -19,7 +19,7 @@ A pull request can only be merged once the four required checks are green:
 
 | check   | workflow      | content                                                              |
 | ------- | ------------- | -------------------------------------------------------------------- |
-| `tests` | `ci-cd.yml`   | the test matrix, linux, macos and windows with python 3.11 to 3.14    |
+| `tests` | `ci-cd.yml`   | the test matrix, linux with python 3.11 to 3.15, macos and windows with 3.14 and 3.15 |
 | `ruff`  | `ruff.yml`    | `ruff check` and `ruff format --check`                                |
 | `ty`    | `ty.yml`      | `tox r -e ty`                                                         |
 | `docs`  | `docs.yml`    | the zensical build including the api reference and the agent files    |
@@ -92,7 +92,7 @@ From now on every commit is checked with ruff (lint and format) and ty, i.e., th
 
 The tests are written with pytest, tox runs them against every supported python version.
 
-The tox environments are named after the interpreter (`py3.11` to `py3.14`, see `envlist` in `tox.ini`), a single one is run with
+The tox environments are named after the interpreter (`py3.11` to `py3.15`, see `envlist` in `tox.ini`), a single one is run with
 
 ```bash
 tox r -e py3.14
@@ -104,7 +104,7 @@ and the complete matrix, including the `ty` environment, in parallel with
 tox run-parallel
 ```
 
-This needs the interpreters to be available, which uv installs with `uv python install 3.11 3.12 3.13 3.14`. Continuous integration runs the same environments as `uvx --with tox-uv tox -e py3.14`.
+This needs the interpreters to be available, which uv installs with `uv python install 3.11 3.12 3.13 3.14 3.15`. Continuous integration runs the same environments as `uvx --with tox-uv tox -e py3.15`.
 
 What follows `--` is passed on to pytest, so a single module or test can be run in the environment of a tox run rather than against the development environment, which is what continuous integration runs:
 
@@ -115,6 +115,8 @@ tox r -e py3.14 -- tests/test_factory.py::test_model_units -x
 
 Without it the whole suite runs, with the `sbml_testsuite` sweep deselected, which is what the environments do in continuous integration.
 
+Python 3.15 tests the core package without the `examples` extra because libroadrunner does not yet publish CPython 3.15 wheels. Tests requiring simulation skip in that environment; Python 3.11 to 3.14 still install the extra and run them. Python 3.15 uses an Intel macOS runner because uv currently provides its macOS interpreter only for Intel.
+
 The `cobra` environment is the one which installs cobrapy, the optional `cobra` extra, and runs the tests which need it, i.e., the flux balance comparison of `tests/test_package_semantics.py` and `tests/fbc/test_cobra.py`; it is pinned to python 3.14 and has a job of its own in `ci-cd.yml`, which informs and is not part of the required `tests` check.
 
 ```bash

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/docs/index.md

@@ -111,7 +111,7 @@ Continue with [Installation](installation.md) and the [model creation guide](cre
 
 If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi.org/10.5281/zenodo.597149):
 
-> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.10.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22646678
+> König, M. (2026). *sbmlutils: Python utilities for SBML* (Version 0.12.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.22853867
 
 ```bibtex
 @software{konig_sbmlutils,
@@ -119,10 +119,10 @@ If you use `sbmlutils` please cite the archived software on [Zenodo](https://doi
   title     = {sbmlutils: Python utilities for SBML},
   year      = {2026},
   month     = sep,
-  version   = {0.10.0},
+  version   = {0.12.0},
   publisher = {Zenodo},
-  doi       = {10.5281/zenodo.22646678},
-  url       = {https://doi.org/10.5281/zenodo.22646678},
+  doi       = {10.5281/zenodo.22853867},
+  url       = {https://doi.org/10.5281/zenodo.22853867},
 }
 ```
 

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/docs/io.md

@@ -119,7 +119,7 @@ The SBML test suite has no distrib case at all, so distrib is verified only on t
 - **Math is normalized.** Math round trips through the infix notation of libsbml, so it comes back equivalent rather than identical: `<cn> 2 </cn>` becomes `<cn type="integer"> 2 </cn>`, and `a * (b * c)` one `<times/>` of three arguments. In the infix notation an id named like a MathML constant or csymbol, `pi`, `INF`, `NaN`, `time` or `avogadro`, cannot be told apart from the constant, so one of them comes back as the other.
 - **The packages which are not modelled:** `groups`, `layout`, `render`, `qual`, `multi`, `spatial` and `arrays`. `sbml_to_model` declares the packages it reads, so neither the content of such a package nor its declaration survives a round trip.
 - **An external model definition is a reference and stays one.** Its `source`, `modelRef` and `md5` come back unchanged and the file is never opened, so a hierarchical model has to be written next to the files it references, see [Model composition](comp.md#what-round-trips).
-- **Annotation URIs are canonicalized.** The annotations are written through pymetadata, which writes a resource in its canonical form: `urn:miriam:pubmed:10659856` becomes `https://identifiers.org/pubmed:10659856`. This is intentional, and idempotent: round tripping the result again writes the same document. A resource whose collection the canonical form would lose is written exactly as it was read instead, and reported once for its collection, see [Annotations](annotations.md#resources-which-are-written-as-given).
+- **Annotation URIs are canonicalized.** The annotations are written through pymetadata, which writes a resource in its canonical form: `urn:miriam:pubmed:10659856` becomes `https://identifiers.org/pubmed:10659856`. This is intentional, and idempotent: round tripping the result again writes the same document. A collection which is not in the identifiers.org registry keeps the classic form, `http://identifiers.org/sabiork/1406` becomes `https://identifiers.org/sabiork/1406`. A malformed resource, which has no canonical form, is written exactly as it was read instead, and reported once for its collection, see [Annotations](annotations.md#resources-which-are-written-as-given).
 - **Rules are grouped by kind.** A model keeps its assignment rules, rate rules and algebraic rules in a list each, so the round trip writes them in that order rather than in the order of the file. The order of the rules has no meaning in SBML.
 - **Notes are wrapped into a body.** Notes which are a sequence of elements such as `<p>` are written inside a `<body>`, an equivalent form. Notes which are a `<body>` or a complete XHTML document rooted at `<html>` are kept as they are.
 - **Inherited species units are made explicit.** A species without `substanceUnits` is written with the `substanceUnits` of the model, which it inherits anyway.

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/docs/validation.md

@@ -86,6 +86,10 @@ Validation judges the document which was written. Writing it reports what libsbm
 - **Content the document cannot carry** - the key value pairs and the user defined constraints of fbc version 3 in an fbc version 2 document, or in one which declares no fbc - is one error per kind of content, with how many pieces on how many elements, an example and the version to declare. Declaring fbc version 3 keeps all of them at once, so this too is reported once rather than per element.
 - **An annotation resource which is written as given** is one warning per collection, see [Annotations](annotations.md#resources-which-are-written-as-given).
 
+Writing a model definition also gives advice on the definition itself, which is not repeated for a model which was parsed from a file:
+
+- **A parameter or compartment which is `constant=False` and never changed** is one warning per kind of element, with every id. Changing a value takes an assignment rule, a rate rule, an algebraic rule, an event assignment or, for a parameter, a user defined constraint of fbc which names it as a variable; an initial assignment does not count. An element of the comp interface - one with a port, or one which replaces an element of a submodel or is replaced by one - is left out, because what changes it can live in another model. The usual cause is a rule which was forgotten or a target which was misspelled.
+
 ## Checking a libsbml call
 
 `check` is the helper the package itself uses around libsbml calls, which return a status code instead of raising:

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/examples/demo/demo.py

@@ -39,16 +39,16 @@ model = Model(
 )
 model.compartments = [
     Compartment(
-        sid="e", value=1e-06, unit=U.m3, constant=False, name="external compartment"
+        sid="e", value=1e-06, unit=U.m3, constant=True, name="external compartment"
     ),
     Compartment(
-        sid="c", value=1e-06, unit=U.m3, constant=False, name="cell compartment"
+        sid="c", value=1e-06, unit=U.m3, constant=True, name="cell compartment"
     ),
     Compartment(
         sid="m",
         value=1,
         unit=U.m2,
-        constant=False,
+        constant=True,
         spatialDimensions=2,
         name="plasma membrane",
     ),

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/examples/icg/model_body.py

@@ -534,7 +534,7 @@ _m.parameters = [
         "Fblood",
         0.02,
         U.dimensionless,
-        constant=False,
+        constant=True,
         name="blood fraction of organ volume",
         sboTerm=SBO.QUANTITATIVE_SYSTEMS_DESCRIPTION_PARAMETER,
         notes="""

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/pyproject.toml

@@ -20,6 +20,7 @@ classifiers = [
 	"Programming Language :: Python :: 3.12",
 	"Programming Language :: Python :: 3.13",
 	"Programming Language :: Python :: 3.14",
+	"Programming Language :: Python :: 3.15",
 	"Programming Language :: Python :: Implementation :: CPython",
 	"Topic :: Scientific/Engineering",
 	"Topic :: Scientific/Engineering :: Bio-Informatics",
@@ -33,7 +34,7 @@ keywords = [
 ]
 dependencies = [
 	# SBML
-	"pymetadata>=0.6.2",
+	"pymetadata>=0.6.4",
 	"python-libsbml>=5.21.2",
 	"antimony>=3.1.3",
 	# the win_amd64 wheel of antimony 3.2.0 raises a C++ exception in loadAntimonyString

sbmlutils-0.12.2/release-notes/0.12.1.md

@@ -0,0 +1,16 @@
+# Release notes for sbmlutils 0.12.1
+![sbmlutils](https://github.com/matthiaskoenig/sbmlutils/raw/develop/docs/images/sbmlutils-logo-60.png)
+
+We are pleased to release the next version of sbmlutils including the following changes.
+
+## Behaviour changes
+- an annotation resource of a collection the identifiers.org registry does not know is normalized like every other resource: `http://identifiers.org/sabiork/1406` is written as `https://identifiers.org/sabiork/1406`, where 0.12.0 wrote it exactly as given and reported it once per collection. pymetadata 0.6.4 keeps the collection of such a resource in the classic form `https://identifiers.org/<collection>/<term>`, up to 0.6.3 it returned the bare `1406`. The same holds for a term without the prefix its collection embeds, `http://identifiers.org/chebi/000000035`. What is still written as given, and reported, is a resource which is malformed to begin with: one without a term, `urn:miriam:chebi` or `chebi/`, and one without a collection, `urn:miriam::1406` (#479)
+- `create_model` warns about the parameters and compartments which are `constant=False` and never changed, once per kind of element with every id. Changing a value takes an assignment rule, a rate rule, an algebraic rule, an event assignment or, for a parameter, a user defined constraint of fbc which names it as a variable; an initial assignment does not count. An element of the comp interface - one with a port, or one which replaces an element of a submodel or is replaced by one - is left out, because what changes it can live in another model. The warning is an authoring hint, so a model which was parsed from a file is written without it (#463)
+
+## Fixes
+- the compartments of the demo example and `Fblood` of the icg example are `constant=True`; they were declared `constant=False` and nothing changes them
+
+## Development
+- pymetadata >= 0.6.4 is required. It also lower cases the collection of every resource and accepts `_` and `-` in the prefix of an identifiers.org URL, so `http://identifiers.org/ec-code/1.1.1.1` is parsed as a resource of `ec-code`; the `identifiers-org` entry of the round trip whitelist in `tests/structural.py` spells a prefix the same way and accepts the classic form
+
+Your sbmlutils team

sbmlutils-0.12.2/release-notes/0.12.2.md

@@ -0,0 +1,15 @@
+# Release notes for sbmlutils 0.12.2
+
+## Python support
+
+- Add Python 3.15 support for the core package and required CI tests on Linux, Windows, and Intel macOS (#482).
+- Keep simulation coverage on Python 3.11 to 3.14. The optional `examples` extra requires libroadrunner, which does not yet provide Python 3.15 wheels; simulation tests skip in the Python 3.15 environment.
+
+## Development
+
+- Add the `py3.15` tox environment and document the expanded CI matrix.
+- Skip simulation example tests when libroadrunner is unavailable, while continuing to test model creation and other core functionality.
+
+## Fixes
+
+- Reject empty iterators in `Model.merge_models` with `ValueError`, matching the documented behavior for empty lists.

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/src/sbmlutils/__init__.py

@@ -7,7 +7,7 @@ from pathlib import Path
 logging.getLogger(__name__).addHandler(logging.NullHandler())
 
 __author__ = "Matthias König"
-__version__ = "0.12.0"
+__version__ = "0.12.2"
 
 
 program_name = "sbmlutils"

{sbmlutils-0.12.0 → sbmlutils-0.12.2}/src/sbmlutils/factory.py

@@ -6805,6 +6805,146 @@ class ModelDict(TypedDict, total=False):
     layouts: list | None
 
 
+def _math_symbols(math: libsbml.ASTNode | None) -> set[str]:
+    """Get the names the math refers to.
+
+    Args:
+        math: the math, `None` for an element without math
+
+    Returns:
+        the name of every `AST_NAME` node of the math
+    """
+    if math is None:
+        return set()
+    symbols: set[str] = set()
+    nodes: list[libsbml.ASTNode] = [math]
+    while nodes:
+        node = nodes.pop()
+        if node.getType() == libsbml.AST_NAME:
+            symbols.add(node.getName())
+        nodes.extend(node.getChild(k) for k in range(node.getNumChildren()))
+    return symbols
+
+
+def _has_comp_replacement(sbase: libsbml.SBase) -> bool:
+    """Say whether an element replaces an element of a submodel or is replaced.
+
+    Args:
+        sbase: the element
+
+    Returns:
+        `True` if the element carries a `<comp:replacedElement>` or a
+        `<comp:replacedBy>`, `False` if not or if the document has no comp
+    """
+    plugin: libsbml.CompSBasePlugin | None = sbase.getPlugin("comp")
+    if plugin is None:
+        return False
+    return bool(plugin.getNumReplacedElements() > 0 or plugin.isSetReplacedBy())
+
+
+def _warn_never_changed(model: libsbml.Model) -> None:
+    """Report the parameters and compartments which vary and never change.
+
+    `constant=False` says that the value of a parameter or the size of a
+    compartment changes during the simulation. That takes something which
+    changes it: an assignment rule or a rate rule with the element as its
+    variable, an event assignment to it, or an algebraic rule, which names no
+    variable and determines one of the symbols of its math, so that every one
+    of them counts. A parameter which is the variable of a component of a
+    `<fbc:userDefinedConstraint>` counts as well: it is a variable of the
+    optimization problem, whose value the solver determines, which is why fbc
+    requires it to be `constant=False`. An initial assignment does not count,
+    it sets the value once and does so for a constant as well. An element with none of them is
+    a constant which says it is not, which is as a rule the trace of a rule
+    which was forgotten or of a target which was misspelled.
+
+    In a hierarchical model what changes an element can live in another
+    model, so the elements of the comp interface are left out: one with a
+    `<comp:port>` is there to be replaced by the model which instantiates
+    this one, and one which replaces an element of a submodel, or is replaced
+    by one, is the target of the rules of that submodel.
+
+    This reads the created libsbml model rather than the model definition,
+    because that is where the targets are complete: a `Parameter` with a
+    formula creates its assignment rule, and a rule creates the parameter it
+    assigns to if it is missing. It reports once per kind of element, with
+    every id, and changes nothing: whether the rule or the `constant` is
+    wrong is not something which can be decided here.
+
+    It is an authoring hint, advice for a model definition being written. A
+    model which was parsed from a file says what its source said, so nothing
+    is reported inside `Sbase.no_authoring_hints`.
+
+    Args:
+        model: the libsbml model after all of its content was created
+    """
+    if not Sbase._authoring_hints.get():
+        return
+
+    changed: set[str] = set()
+    for k in range(model.getNumRules()):
+        rule: libsbml.Rule = model.getRule(k)
+        if rule.isAlgebraic():
+            changed |= _math_symbols(rule.getMath())
+        else:
+            changed.add(rule.getVariable())
+    for k in range(model.getNumEvents()):
+        event: libsbml.Event = model.getEvent(k)
+        for j in range(event.getNumEventAssignments()):
+            event_assignment: libsbml.EventAssignment = event.getEventAssignment(j)
+            changed.add(event_assignment.getVariable())
+
+    fbc_model: libsbml.FbcModelPlugin | None = model.getPlugin("fbc")
+    if fbc_model is not None:
+        for k in range(fbc_model.getNumUserDefinedConstraints()):
+            constraint: libsbml.UserDefinedConstraint = (
+                fbc_model.getUserDefinedConstraint(k)
+            )
+            for j in range(constraint.getNumUserDefinedConstraintComponents()):
+                component: libsbml.UserDefinedConstraintComponent = (
+                    constraint.getUserDefinedConstraintComponent(j)
+                )
+                changed.add(component.getVariable())
+
+    comp_model: libsbml.CompModelPlugin | None = model.getPlugin("comp")
+    if comp_model is not None:
+        for k in range(comp_model.getNumPorts()):
+            port: libsbml.Port = comp_model.getPort(k)
+            if port.isSetIdRef():
+                changed.add(port.getIdRef())
+
+    compartments: list[libsbml.Compartment] = [
+        model.getCompartment(k) for k in range(model.getNumCompartments())
+    ]
+    parameters: list[libsbml.Parameter] = [
+        model.getParameter(k) for k in range(model.getNumParameters())
+    ]
+    elements: dict[str, list[libsbml.Compartment] | list[libsbml.Parameter]] = {
+        "Compartment": compartments,
+        "Parameter": parameters,
+    }
+    for kind, sbases in elements.items():
+        never_changed: list[str] = [
+            sbase.getId()
+            for sbase in sbases
+            if not sbase.getConstant()
+            and sbase.getId() not in changed
+            and not _has_comp_replacement(sbase)
+        ]
+        if never_changed:
+            logger.warning(
+                "%s '%s' element(s) %s of the model '%s' are 'constant=False', "
+                "but none is ever changed by an assignment rule, a rate rule, "
+                "an algebraic rule, an event assignment or a user defined "
+                "constraint of fbc. Set 'constant=True' or add what changes "
+                "them.",
+                len(never_changed),
+                kind,
+                never_changed,
+                model.getId(),
+            )
+
+
 class Model(Sbase, FrozenClass):
     """Model.
 
@@ -7253,6 +7393,9 @@ class Model(Sbase, FrozenClass):
                 if objects:
                     create_objects(model, obj_iter=objects, key=attr)
 
+        # after everything which can change a value was created
+        _warn_never_changed(model)
+
     def get_sbml(self) -> str:
         """Create SBML model."""
         return Document(model=self).get_sbml()
@@ -7468,6 +7611,7 @@ class Model(Sbase, FrozenClass):
         """
         if isinstance(models, Model):
             return models
+        models = list(models)
         if not models:
             raise ValueError("No models are provided.")
         model = Model("template")