pysec2pri 0.2.0__tar.gz → 0.2.2__tar.gz

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pysec2pri
-Version: 0.2.0
+Version: 0.2.2
 Summary: Secondary to primary identifier mapping
 Keywords: snekpack,cookiecutter
 Author: Javier Millán Acosta
@@ -40,7 +40,7 @@ Requires-Dist: numpy
 Requires-Dist: rdkit
 Requires-Dist: polars-runtime-32
 Requires-Dist: sssom-schema==1.1.0a5
-Requires-Dist: mapkgsutils>=0.0.4
+Requires-Dist: mapkgsutils>=0.0.5
 Maintainer: Javier Millán Acosta
 Maintainer-email: Javier Millán Acosta <javier.millan.acosta@gmail.com>
 Requires-Python: >=3.10
@@ -96,20 +96,33 @@ uv pip install git+https://github.com/jmillanacosta/pysec2pri.git
 
 ### Generating mapping sets
 
-To obtain the secondary to primary identifier SSSOM mapping set for ChEBI:
+Mapping sets can be generated from bash:
 
 ```bash
-pysec2pri chebi ids
+# ID mapping sets
+pysec2pri chebi ids (--help)
+pysec2pri ensembl ids (--help)
+pysec2pri hgnc ids (--help)
+pysec2pri vgnc ids (--help)
+pysec2pri ncbi ids (--help)
+pysec2pri hmdb-protein ids (--help)
+pysec2pri hmdb-gene ids (--help)
+pysec2pri uniprot ids (--help)
+# Label mapping sets
+pysec2pri chebi labels (--help)
+pysec2pri hgnc labels (--help)
+pysec2pri ensembl labels (--help)
+pysec2pri vgnc labels (--help)
 ```
 
-This will automatically download the latest ChEBI release and generate an SSSOM
-mapping file in your current directory.
+Or python:
 
-To process local files and specify the output:
-
-```bash
-pysec2pri chebi ids ChEBI_complete_3star.sdf --output my_mappings.sssom.tsv
 ```
+from pysec2pri import generate_ensembl_labels, generate_ensembl
+```
+
+Replacing Ensembl by the supported database. These functions return either a
+`IdMappingSet` or `LabelMappingSet`, SSSOM `MappingSet`s.
 
 For more options and help on any command:
 
@@ -122,14 +135,23 @@ The default output is in [SSSOM](https://mapping-commons.github.io/sssom/)
 
 ### Updating IDs and labels
 
-A generated mapping set can be used to update IDs and labels in Python:
+A generated mapping set can be used to update IDs and labels in Python: ChEBI
+synonyms:
 
 ```python
 from pysec2pri import generate_chebi_synonyms, resolve_labels
-cs = generate_chebi_synonyms()
+chebi_ms = generate_chebi_synonyms()
 resolve_labels(["Glucose", "ATP", "Guanine"], cs)
 ```
 
+Ensembl gene identifiers in a dataframe:
+
+```python
+from pysec2pri import update_ids, generate_ensembl
+ens_ms = generate_ensembl(version="115", species="9606")
+df_with_new_column = update_ids(mapping_set=ens_ms, ids = df, at="Ensembl_id")  # `at` is the name of the column
+```
+
 Or from the command line, given a TSV file `gene_ex.tsv`:
 
 ```

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/README.md

@@ -43,20 +43,33 @@ uv pip install git+https://github.com/jmillanacosta/pysec2pri.git
 
 ### Generating mapping sets
 
-To obtain the secondary to primary identifier SSSOM mapping set for ChEBI:
+Mapping sets can be generated from bash:
 
 ```bash
-pysec2pri chebi ids
+# ID mapping sets
+pysec2pri chebi ids (--help)
+pysec2pri ensembl ids (--help)
+pysec2pri hgnc ids (--help)
+pysec2pri vgnc ids (--help)
+pysec2pri ncbi ids (--help)
+pysec2pri hmdb-protein ids (--help)
+pysec2pri hmdb-gene ids (--help)
+pysec2pri uniprot ids (--help)
+# Label mapping sets
+pysec2pri chebi labels (--help)
+pysec2pri hgnc labels (--help)
+pysec2pri ensembl labels (--help)
+pysec2pri vgnc labels (--help)
 ```
 
-This will automatically download the latest ChEBI release and generate an SSSOM
-mapping file in your current directory.
+Or python:
 
-To process local files and specify the output:
-
-```bash
-pysec2pri chebi ids ChEBI_complete_3star.sdf --output my_mappings.sssom.tsv
 ```
+from pysec2pri import generate_ensembl_labels, generate_ensembl
+```
+
+Replacing Ensembl by the supported database. These functions return either a
+`IdMappingSet` or `LabelMappingSet`, SSSOM `MappingSet`s.
 
 For more options and help on any command:
 
@@ -69,14 +82,23 @@ The default output is in [SSSOM](https://mapping-commons.github.io/sssom/)
 
 ### Updating IDs and labels
 
-A generated mapping set can be used to update IDs and labels in Python:
+A generated mapping set can be used to update IDs and labels in Python: ChEBI
+synonyms:
 
 ```python
 from pysec2pri import generate_chebi_synonyms, resolve_labels
-cs = generate_chebi_synonyms()
+chebi_ms = generate_chebi_synonyms()
 resolve_labels(["Glucose", "ATP", "Guanine"], cs)
 ```
 
+Ensembl gene identifiers in a dataframe:
+
+```python
+from pysec2pri import update_ids, generate_ensembl
+ens_ms = generate_ensembl(version="115", species="9606")
+df_with_new_column = update_ids(mapping_set=ens_ms, ids = df, at="Ensembl_id")  # `at` is the name of the column
+```
+
 Or from the command line, given a TSV file `gene_ex.tsv`:
 
 ```

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "uv_build"
 
 [project]
 name = "pysec2pri"
-version = "0.2.0"
+version = "0.2.2"
 description = "Secondary to primary identifier mapping"
 readme = "README.md"
 authors = [
@@ -71,7 +71,7 @@ dependencies = [
     "rdkit", # Chemistry toolkit for SDF parsing
     "polars-runtime-32", # Polars runtime for 32-bit systems
     "sssom-schema==1.1.0a5", # SSSOM schema for validation
-    "mapkgsutils>=0.0.4",
+    "mapkgsutils>=0.0.5",
 ]
 
 # see https://peps.python.org/pep-0735/ and https://docs.astral.sh/uv/concepts/dependencies/#dependency-groups
@@ -255,7 +255,7 @@ known-first-party = [
 docstring-code-format = true
 
 [tool.bumpversion]
-current_version = "0.2.0"
+current_version = "0.2.2"
 parse = "(?P<major>\\d+)\\.(?P<minor>\\d+)\\.(?P<patch>\\d+)(?:-(?P<release>[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*))?(?:\\+(?P<build>[0-9A-Za-z-]+(?:\\.[0-9A-Za-z-]+)*))?"
 serialize = [
     "{major}.{minor}.{patch}-{release}+{build}",

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/api.py

@@ -878,7 +878,7 @@ def _generate_ensembl_all_species(
         version = check_ensembl_release().version
         if version is None:
             raise ValueError("Could not determine the latest Ensembl release.")
-    release_date = resolve_release_date("ensembl", version)
+    release_date = resolve_release_date("ensembl", version, species=9606)
 
     species_list = discover_ensembl_species(version)
     iterator: Any = species_list
@@ -1001,7 +1001,7 @@ def generate_ensembl_labels(
         species: Canonical NCBI taxon ID, or ``"all"`` (default) to process
             every species Ensembl publishes and combine them into one
             mapping set -- see :func:`_generate_ensembl_all_species`. A
-            symbol shared by two species becomes genuinely ambiguous in
+            symbol shared by two species becomes ambiguous in
             that combined output (there's no longer a per-species scope to
             disambiguate it). This is network-heavy (~276 species); pass
             an explicit taxon ID for a fast, single-species run.

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/cli.py

@@ -464,7 +464,7 @@ def ensembl_label_history_cmd(
     Ensembl's core schema has no previous-gene-symbol table, so this walks
     every historical release (or a bounded --from-version/--to-version
     range) and diffs each release's primary-label snapshot to recover
-    genuine previous -> current symbol transitions. Network-heavy and
+    true previous -> current symbol transitions. Network-heavy and
     resumable; run on demand, not part of normal mapping generation.
     """
     from pysec2pri.api import generate_ensembl_label_history
@@ -643,6 +643,7 @@ def _make_consolidate_cmd(cfg_id: str, extra_opts: list[Callable[..., Any]]) ->
         cache_dir: Path | None,
         force: bool,
         no_progress: bool,
+        output: Path | None,
         **extra_kwargs: Any,
     ) -> None:
         from pysec2pri.consolidate import _sssom_output_path, consolidate_mapping_dates
@@ -653,18 +654,19 @@ def _make_consolidate_cmd(cfg_id: str, extra_opts: list[Callable[..., Any]]) ->
             f"(mode={mode}{', ' + extras if extras else ''}, {mapping_sets})..."
         )
         try:
-            path = consolidate_mapping_dates(
+            path, _ = consolidate_mapping_dates(
                 cfg_id,
                 mode=mode,
                 cache_dir=cache_dir,
                 mapping_sets=mapping_sets,
                 show_progress=not no_progress,
                 force=force,
+                output=output,
                 **extra_kwargs,
             )
         except ValueError as exc:
             raise click.ClickException(str(exc)) from exc
-        click.echo(f"Wrote consolidated mapping set -> {_sssom_output_path(path)}")
+        click.echo(f"Wrote consolidated mapping set -> {output or _sssom_output_path(path)}")
 
     decorators: list[Callable[..., Any]] = [
         click.option(
@@ -693,6 +695,13 @@ def _make_consolidate_cmd(cfg_id: str, extra_opts: list[Callable[..., Any]]) ->
             default=False,
             help="Re-scan every release, ignoring resume state.",
         ),
+        click.option(
+            "-o",
+            "--output",
+            type=PathType,
+            default=None,
+            help="Also write the consolidated SSSOM mapping set to this path.",
+        ),
         _opt_no_progress,
     ]
     for dec in reversed(decorators):

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/config/ensembl.yaml

@@ -7,13 +7,6 @@ prefix: "ENSEMBL"
 curie_base_url: "http://identifiers.org/ensembl/"
 entity_types: ["genes"]
 
-# Canonical key = NCBI taxon ID. `token` is Ensembl's own species directory
-# component (e.g. https://ftp.ensembl.org/pub/release-115/mysql/homo_sapiens_core_115_38/).
-# `default: "all"` processes every species Ensembl publishes for the
-# release and combines them into one mapping set (see
-# pysec2pri.api._generate_ensembl_all_species) -- network-heavy (~276
-# species, one download each); pass an explicit taxon ID for a fast,
-# single-species run.
 species:
   default: "all"
   available:

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/config/ncbi.yaml

@@ -9,14 +9,6 @@ prefix: "NCBIGene"
 curie_base_url: "http://identifiers.org/ncbigene/"
 entity_types: ["genes"]
 
-# NCBI's own token IS the canonical taxon ID (no translation needed, unlike
-# Ensembl) -- gene_history/gene_info are filtered directly on #tax_id, which
-# accepts ANY taxon ID, not just the ones listed here. `available` exists
-# only to drive CLI help/choices for common species; NCBIParser never calls
-# species_token(), so it never restricts which taxon IDs are accepted.
-# `default: "all"` processes every organism in the file instead of
-# filtering to one (see NCBIParser.ALL_SPECIES) -- cheap here since
-# gene_info/gene_history are single global files already being downloaded.
 species:
   default: "all"
   available:

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/config/vgnc.yaml

@@ -7,22 +7,6 @@ prefix: "VGNC"
 curie_base_url: "http://identifiers.org/vgnc/"
 entity_types: ["genes"]
 
-# VGNC (Vertebrate Gene Nomenclature Committee) names genes for vertebrate
-# species that lack their own nomenclature committee (human/mouse/rat are
-# deliberately out of scope -- HGNC/MGI/RGD cover those). Unlike Ensembl
-# (one download URL per species), VGNC publishes a single global TSV per
-# file with a `taxon_id` column, filtered the same way NCBI's
-# gene_info/gene_history are filtered by `#tax_id` -- see VGNCParser.
-# `default: "all"` processes every species together (see
-# VGNCParser.ALL_SPECIES) -- cheap for ids (the global withdrawn set is
-# always parsed in full regardless), heavier for labels (a symbol shared
-# by two species becomes genuinely ambiguous with no per-species scope to
-# disambiguate it). VGNCParser never calls species_token(), so `available`
-# exists only to drive CLI help/choices and never restricts which taxon IDs
-# are accepted. Taxon IDs below were verified against the live data (each
-# species' own vgnc/tsv/<species>/*_vgnc_gene_set_All.tsv, cross-referenced
-# by VGNC ID against vgnc/tsv/all/all_vgnc_gene_set_All.tsv's taxon_id
-# column).
 species:
   default: "all"
   available:

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/consolidate.py

@@ -161,6 +161,59 @@ _read_meta = _consolidate.read_meta
 _write_meta = _consolidate.write_meta
 
 
+def _save_optional_output(mapping_set: BaseMappingSet, output: Path | None) -> None:
+    """Write *mapping_set* to *output* as SSSOM when an output path is given."""
+    if output is not None:
+        mapping_set.save("sssom", output)
+
+
+def _consolidate_by_date_full(
+    cache_path: Path,
+    meta_path: Path,
+    *,
+    run_one_version: Callable[[], Any],
+) -> bool:
+    """Date mode that caches the *full* mapping set, not just dated rows.
+
+    :func:`mapkgsutils.consolidate.consolidate_by_date` keeps only rows that
+    carry a real per-row ``mapping_date`` and drops the rest, so a single
+    parse of e.g. HGNC labels would cache only its ~12k dated previous-symbol
+    mappings and lose the ~48k alias mappings. This builds a complete cache
+    from one (current) parse: every mapping is kept, stamped with its own
+    per-row date when present and left undated otherwise (exactly how the
+    release walk treats an unresolvable date).
+
+    Returns ``True`` (cache written) when the parse yields at least one real
+    per-row date, and ``False`` (cache untouched) when it yields none -- so
+    date-less sources (ChEBI, UniProt) still fall back to a release walk
+    rather than caching a fully undated set.
+    """
+    mapping_set = run_one_version()
+    mappings = mapping_set.mappings or []
+    if not any(getattr(m, "mapping_date", None) for m in mappings):
+        return False
+
+    version_label = str(getattr(mapping_set, "mapping_set_version", None) or "current")
+    records: dict[str, dict[str, str]] = {}
+    for m in mappings:
+        # record_id is release-scoped; its trailing 16 hex chars are the
+        # version-independent pair hash (see mapkgsutils.parsers.base.pair_hash).
+        pair_key = str(getattr(m, "record_id", None) or "")[-16:]
+        if not pair_key:
+            continue
+        date_str = str(m.mapping_date) if getattr(m, "mapping_date", None) else ""
+        records[pair_key] = {
+            "first_seen_version": version_label,
+            "first_seen_date": date_str,
+            "last_seen_version": version_label,
+            "last_seen_date": date_str,
+            "fields_json": _consolidate._mapping_fields_json(m),
+        }
+    _write_cache(cache_path, records)
+    _write_meta(meta_path, version_label)
+    return True
+
+
 def load_mapping_dates(
     datasource: str,
     cache_dir: Path | None = None,
@@ -359,8 +412,12 @@ def _build_consolidated_mapping_set(
 
 def _write_consolidated_sssom(
     datasource: str, mapping_sets: str, cache_path: Path, meta_path: Path
-) -> Path:
-    """Build and save the companion SSSOM mapping set next to the cache file."""
+) -> tuple[Path, Any]:
+    """Build and save the companion SSSOM mapping set next to the cache file.
+
+    Returns ``(output_path, mapping_set)`` -- see
+    :func:`mapkgsutils.consolidate.write_consolidated_sssom`.
+    """
     from pysec2pri.parsers.base import IdMappingSet, LabelMappingSet
 
     config = ALL_DATASOURCES[datasource]
@@ -383,8 +440,9 @@ def consolidate_mapping_dates(
     mapping_sets: str = "ids",
     show_progress: bool = True,
     force: bool = False,
+    output: Path | None = None,
     **kwargs: Any,
-) -> Path:
+) -> tuple[Path, BaseMappingSet]:
     """Build/update the first-seen-date index for *datasource*.
 
     Two strategies, selected via *mode*:
@@ -393,17 +451,18 @@ def consolidate_mapping_dates(
       (oldest first, resuming from the last completed version unless
       *force* is set). For every mapping seen, records the version/date it
       first appeared and keeps bumping the version/date it was last seen.
-      This is a slow, network-heavy operation (~250 releases for ChEBI) and
+      This is a slow, network-heavy operation (e.g. ~250 releases for ChEBI) and
       is meant to be run manually/as a one-off, not as part of normal
       mapping generation. Requires a versioned archive (ChEBI, HGNC,
       UniProt); raises :class:`ValueError` for datasources without one
       (e.g. NCBI).
-    - ``"date"``: a single (current) parse, capturing whatever real
-      per-row ``mapping_date`` the datasource's parser already produces
-      (e.g. HGNC's ``date_symbol_changed``, NCBI's ``Discontinue_Date``).
-      Fast -- no historical walk. If *datasource* never produces a real
-      per-row date (ChEBI, UniProt), falls back to ``"release"`` mode and
-      warns the user.
+    - ``"date"``: a single (current) parse, caching the *full* mapping set
+      with whatever real per-row ``mapping_date`` the datasource's parser
+      already produces (e.g. HGNC's ``date_symbol_changed``, NCBI's
+      ``Discontinue_Date``); rows without such a date are still cached, just
+      left undated. Fast -- no historical walk. If *datasource* never
+      produces a real per-row date at all (ChEBI, UniProt), falls back to
+      ``"release"`` mode and warns the user.
 
     Either way, alongside the internal cache file this also writes a
     companion real SSSOM mapping set (see :func:`_sssom_output_path`) where
@@ -420,13 +479,18 @@ def consolidate_mapping_dates(
             (``"release"`` mode only).
         force: Re-scan every release from scratch, ignoring any existing
             cache/resume state (``"release"`` mode only).
+        output: Optional path to also write the consolidated SSSOM mapping
+            set to. The internal cache-adjacent copy (see
+            :func:`_sssom_output_path`) is always written; when *output* is
+            given, the same full mapping set is additionally saved there.
         **kwargs: Datasource-specific knobs (``subset`` for ChEBI, ``species``
             for NCBI/Ensembl); ignored for datasources with no such config
             block.
 
     Returns:
-        Path to the written cache TSV (see :func:`_sssom_output_path` for
-        the companion SSSOM mapping set written alongside it).
+        ``(cache_path, mapping_set)``: the path to the written cache TSV
+        (see :func:`_sssom_output_path` for the companion SSSOM mapping set
+        written alongside it) and the in-memory consolidated mapping set.
 
     Raises:
         ValueError: For an unknown *mode*, an unsupported *datasource*, an
@@ -461,14 +525,17 @@ def consolidate_mapping_dates(
     meta_path = _meta_path(cache_dir, datasource, mapping_sets, **kwargs)
 
     if mode == "date":
-        got_dates = _consolidate.consolidate_by_date(
+        got_dates = _consolidate_by_date_full(
             cache_path,
             meta_path,
             run_one_version=lambda: _run_one_version(datasource, None, mapping_sets, **kwargs),
         )
         if got_dates:
-            _write_consolidated_sssom(datasource, mapping_sets, cache_path, meta_path)
-            return cache_path
+            _, mapping_set = _write_consolidated_sssom(
+                datasource, mapping_sets, cache_path, meta_path
+            )
+            _save_optional_output(mapping_set, output)
+            return cache_path, mapping_set
         msg = (
             f"{datasource!r} has no parseable per-row mapping dates; "
             "falling back to mode='release'."
@@ -492,12 +559,13 @@ def consolidate_mapping_dates(
         show_progress=show_progress,
         force=force,
     )
-    _write_consolidated_sssom(datasource, mapping_sets, cache_path, meta_path)
-    return cache_path
+    _, mapping_set = _write_consolidated_sssom(datasource, mapping_sets, cache_path, meta_path)
+    _save_optional_output(mapping_set, output)
+    return cache_path, mapping_set
 
 
 # Cross-release label history (e.g. Ensembl, whose core schema has no
-# previous-gene-symbol table -- genuine previous->current symbol transitions
+# previous-gene-symbol table: previous->current symbol transitions
 # are recovered by diffing each release's current-label snapshot).
 
 

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/parsers/hmdb.py

@@ -17,7 +17,6 @@ import gzip
 import io
 import re
 import zipfile
-from collections.abc import Callable
 from pathlib import Path
 from typing import IO, TYPE_CHECKING
 
@@ -85,7 +84,6 @@ class HMDBParser(BaseParser):
         element_tag: str,
         prefix: str,
         desc: str,
-        secondary_normaliser: Callable[[str], str] | None = None,
     ) -> BaseMappingSet:
         """Parse an HMDB XML file and return a mapping set.
 
@@ -95,8 +93,6 @@ class HMDBParser(BaseParser):
                 ``"protein"``).
             prefix: CURIE prefix (``"HMDB"`` for both record types).
             desc: Progress-bar description string.
-            secondary_normaliser: Optional callable ``(raw_str) -> str``
-                applied to each raw secondary accession before prefixing.
 
         Returns:
             :class:`BaseMappingSet` with computed cardinalities and
@@ -130,10 +126,10 @@ class HMDBParser(BaseParser):
                         accession_elem = elem.find("accession")
                     if accession_elem is not None and accession_elem.text:
                         primary_ids_found.add(f"{prefix}:{accession_elem.text.strip()}")
-                    rows_data.extend(self._process_record(elem, prefix, secondary_normaliser))
+                    rows_data.extend(self._process_record(elem, prefix))
                     elem.clear()
         except DefusedET.ParseError:
-            rows_data = self._parse_simple_xml(file_path, element_tag, prefix, secondary_normaliser)
+            rows_data = self._parse_simple_xml(file_path, element_tag, prefix)
 
         mappings = self._build_mappings(
             rows_data, fixed, desc="Building HMDB mappings", total=len(rows_data)
@@ -147,9 +143,14 @@ class HMDBParser(BaseParser):
         self,
         elem: Element,
         prefix: str,
-        secondary_normaliser: Callable[[str], str] | None,
     ) -> list[dict[str, str]]:
-        """Extract row dicts from a single record element."""
+        """Extract row dicts from a single record element.
+
+        Bare-numeric ``secondary_accessions`` entries are skipped: in the
+        proteins file these are HMDB's internal database row ids (not legacy
+        accessions), and a bare integer is never a valid HMDB accession for
+        either record type. See issue #44.
+        """
         # Primary accession
         accession_elem = elem.find("hmdb:accession", HMDB_NS)
         if accession_elem is None:
@@ -177,8 +178,8 @@ class HMDBParser(BaseParser):
             if not sec_elem.text:
                 continue
             raw_sec = sec_elem.text.strip()
-            if secondary_normaliser is not None:
-                raw_sec = secondary_normaliser(raw_sec)
+            if _BARE_NUM_RE.match(raw_sec):
+                continue
             secondary_id = f"{prefix}:{raw_sec}"
             rows.append(
                 {
@@ -200,7 +201,6 @@ class HMDBParser(BaseParser):
         file_path: Path,
         element_tag: str,
         prefix: str,
-        secondary_normaliser: Callable[[str], str] | None,
     ) -> list[dict[str, str]]:
         """Fallback parser for test files without a namespace declaration."""
         rows_data: list[dict[str, str]] = []
@@ -208,7 +208,7 @@ class HMDBParser(BaseParser):
             tree = DefusedET.parse(file_path)
             root = tree.getroot()
             for record in root.findall(f".//{element_tag}"):
-                rows_data.extend(self._process_record(record, prefix, secondary_normaliser))
+                rows_data.extend(self._process_record(record, prefix))
         except Exception as e:
             logger.warning("Failed to parse HMDB XML: %s", e)
         return rows_data
@@ -329,9 +329,15 @@ class HMDBProteinParser(HMDBParser):
 
     Reads configuration from ``hmdb_proteins.yaml``.
     Primary accessions have the form ``HMDBP…`` (e.g. ``HMDBP00001``).
-    Secondary accessions may be bare numbers (legacy format, e.g. ``5229``)
-    or full ``HMDBP`` accessions; both are normalised by
-    :meth:`_normalise_protein_secondary`.
+
+    A record's ``<secondary_accessions>`` block holds two distinct kinds of
+    value. Only the first is a real identifier:
+
+    * **``HMDBP``-prefixed accessions** (e.g. ``HMDBP05261``): genuine retired
+      accessions that map to the record's primary accession.
+    * **Bare numeric values** (e.g. ``5229``): HMDB's internal database row id,
+      *not* a legacy accession. These are dropped by the shared
+      :meth:`~HMDBParser._process_record` (see issue #44).
     """
 
     datasource_name = "hmdb_proteins"
@@ -341,26 +347,6 @@ class HMDBProteinParser(HMDBParser):
         """Proteins download URL from ``hmdb_proteins.yaml``."""
         return self.get_download_url("proteins") or ""
 
-    @staticmethod
-    def _normalise_protein_secondary(raw: str) -> str:
-        """Normalise a raw secondary protein accession to a ``HMDBP``-prefixed form.
-
-        Bare numeric values (legacy IDs) are zero-padded to five digits and
-        prefixed with ``HMDBP`` to match the primary accession pattern.
-        Already-prefixed values are returned unchanged.
-
-        Args:
-            raw: Raw secondary accession string (e.g. ``"5229"`` or
-                ``"HMDBP05229"``).
-
-        Returns:
-            Normalised accession string (e.g. ``"HMDBP05229"``).
-        """
-        raw = raw.strip()
-        if _BARE_NUM_RE.match(raw):
-            return f"HMDBP{int(raw):05d}"
-        return raw
-
     def parse(self, input_path: Path | str | None) -> BaseMappingSet:
         """Parse ``hmdb_proteins.xml`` (or ``.zip`` / ``.gz``).
 
@@ -377,7 +363,6 @@ class HMDBProteinParser(HMDBParser):
             element_tag="protein",
             prefix="HMDBP",
             desc="Parsing HMDB proteins XML",
-            secondary_normaliser=self._normalise_protein_secondary,
         )
 
 

{pysec2pri-0.2.0 → pysec2pri-0.2.2}/src/pysec2pri/version.py

@@ -12,7 +12,7 @@ __all__ = [
     "get_version",
 ]
 
-VERSION = "0.2.0"
+VERSION = "0.2.2"
 
 
 def get_git_hash() -> str: