test new metadata parsing

fgcz · Oct 17, 2024 · 0657dc1 · 0657dc1
1 parent 9d4690e
commit 0657dc1
Show file tree

Hide file tree

Showing 9 changed files with 150 additions and 3 deletions.
diff --git a/src/depiction/persistence/imzml/parser/parse_metadata.py b/src/depiction/persistence/imzml/parser/parse_metadata.py
@@ -13,11 +13,11 @@ def ibd_checksums(self) -> dict[str, str]:
         elements = self._etree.findall(f".//{self._ns}fileDescription/{self._ns}fileContent/{self._ns}cvParam")
         checksums = {}
         for element in elements:
-            if element.attrib["accession"] == "MS:1000568":
+            if element.attrib["accession"] in ("MS:1000568", "IMS:1000090"):
                 checksums["md5"] = element.attrib["value"].lower()
-            elif element.attrib["accession"] == "MS:1000569":
+            elif element.attrib["accession"] in ("MS:1000569", "IMS:1000091"):
                 checksums["sha1"] = element.attrib["value"].lower()
-            elif element.attrib["accession"] == "MS:1003151":
+            elif element.attrib["accession"] in ("MS:1003151", "IMS:1000092"):
                 checksums["sha256"] = element.attrib["value"].lower()
         return checksums
 

diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/checksums_md5.xml b/tests/integration/imzml_parser/chunks/parse_metadata/checksums_md5.xml
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <fileDescription>
+    <fileContent>
+      <cvParam accession="IMS:1000030" cvRef="IMS" name="continuous"/>
+      <cvParam accession="IMS:1000090" cvRef="IMS" name="ibd MD5" value="00000000000111111111112222222222"/>
+      <cvParam accession="MS:1000294" cvRef="MS" name="mass spectrum"/>
+    </fileContent>
+  </fileDescription>
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/checksums_multiple.xml b/tests/integration/imzml_parser/chunks/parse_metadata/checksums_multiple.xml
@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <fileDescription>
+    <fileContent>
+      <cvParam accession="IMS:1000030" cvRef="IMS" name="continuous"/>
+      <cvParam accession="IMS:1000090" cvRef="IMS" name="ibd MD5" value="00000000000111111111112222222222"/>
+      <cvParam accession="IMS:1000092" cvRef="IMS" name="ibd SHA-256" value="aaaaaa"/>
+      <cvParam accession="MS:1000294" cvRef="MS" name="mass spectrum"/>
+    </fileContent>
+  </fileDescription>
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/checksums_none.xml b/tests/integration/imzml_parser/chunks/parse_metadata/checksums_none.xml
@@ -0,0 +1,3 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/checksums_sha1.xml b/tests/integration/imzml_parser/chunks/parse_metadata/checksums_sha1.xml
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <fileDescription>
+    <fileContent>
+      <cvParam accession="IMS:1000030" cvRef="IMS" name="continuous"/>
+      <cvParam accession="IMS:1000091" cvRef="IMS" name="ibd SHA-1" value="abcdef"/>
+      <cvParam accession="MS:1000294" cvRef="MS" name="mass spectrum"/>
+    </fileContent>
+  </fileDescription>
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/checksums_sha256.xml b/tests/integration/imzml_parser/chunks/parse_metadata/checksums_sha256.xml
@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <fileDescription>
+    <fileContent>
+      <cvParam accession="IMS:1000030" cvRef="IMS" name="continuous"/>
+      <cvParam accession="IMS:1000092" cvRef="IMS" name="ibd SHA-256" value="aaaaaa"/>
+      <cvParam accession="MS:1000294" cvRef="MS" name="mass spectrum"/>
+    </fileContent>
+  </fileDescription>
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/pixel_size_2d.xml b/tests/integration/imzml_parser/chunks/parse_metadata/pixel_size_2d.xml
@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <cvList count="3">
+    <cv URI="http://purl.obolibrary.org/obo/ms.obo" fullName="Proteomics Standards Initiative Mass Spectrometry Ontology" id="MS" version="4.1.35"/>
+    <cv URI="http://purl.obolibrary.org/obo/uo.obo" fullName="Unit Ontology" id="UO" version="2019-03-29"/>
+    <cv URI="http://www.maldi-msi.org/download/imzml/imagingMS.obo" fullName="Imaging MS Ontology" id="IMS" version="0.9.1"/>
+  </cvList>
+  <scanSettingsList count="1">
+    <scanSettings id="scanSettings0">
+      <cvParam accession="IMS:1000401" cvRef="IMS" name="top down"/>
+      <cvParam accession="IMS:1000410" cvRef="IMS" name="meandering"/>
+      <cvParam accession="IMS:1000490" cvRef="IMS" name="linescan right left"/>
+      <cvParam accession="IMS:1000480" cvRef="IMS" name="horizontal line scan"/>
+      <cvParam accession="IMS:1000042" cvRef="IMS" name="max count of pixels x" value="212"/>
+      <cvParam accession="IMS:1000043" cvRef="IMS" name="max count of pixels y" value="270"/>
+      <cvParam accession="IMS:1000044" cvRef="IMS" name="max dimension x" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="10600"/>
+      <cvParam accession="IMS:1000045" cvRef="IMS" name="max dimension y" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="13500"/>
+      <cvParam accession="IMS:1000053" cvRef="IMS" name="absolute position offset x" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="0"/>
+      <cvParam accession="IMS:1000054" cvRef="IMS" name="absolute position offset y" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="0"/>
+      <cvParam accession="IMS:1000046" cvRef="IMS" name="pixel size x" value="50"/>
+      <cvParam accession="IMS:1000047" cvRef="IMS" name="pixel size y" value="20"/>
+    </scanSettings>
+  </scanSettingsList>
+</mzML>
diff --git a/tests/integration/imzml_parser/chunks/parse_metadata/pixel_size_none.xml b/tests/integration/imzml_parser/chunks/parse_metadata/pixel_size_none.xml
@@ -0,0 +1,22 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<mzML version="1.1.0" xmlns="http://psi.hupo.org/ms/mzml" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://psi.hupo.org/ms/mzml http://psidev.info/files/ms/mzML/xsd/mzML1.1.0.xsd">
+  <cvList count="3">
+    <cv URI="http://purl.obolibrary.org/obo/ms.obo" fullName="Proteomics Standards Initiative Mass Spectrometry Ontology" id="MS" version="4.1.35"/>
+    <cv URI="http://purl.obolibrary.org/obo/uo.obo" fullName="Unit Ontology" id="UO" version="2019-03-29"/>
+    <cv URI="http://www.maldi-msi.org/download/imzml/imagingMS.obo" fullName="Imaging MS Ontology" id="IMS" version="0.9.1"/>
+  </cvList>
+  <scanSettingsList count="1">
+    <scanSettings id="scanSettings0">
+      <cvParam accession="IMS:1000401" cvRef="IMS" name="top down"/>
+      <cvParam accession="IMS:1000410" cvRef="IMS" name="meandering"/>
+      <cvParam accession="IMS:1000490" cvRef="IMS" name="linescan right left"/>
+      <cvParam accession="IMS:1000480" cvRef="IMS" name="horizontal line scan"/>
+      <cvParam accession="IMS:1000042" cvRef="IMS" name="max count of pixels x" value="212"/>
+      <cvParam accession="IMS:1000043" cvRef="IMS" name="max count of pixels y" value="270"/>
+      <cvParam accession="IMS:1000044" cvRef="IMS" name="max dimension x" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="10600"/>
+      <cvParam accession="IMS:1000045" cvRef="IMS" name="max dimension y" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="13500"/>
+      <cvParam accession="IMS:1000053" cvRef="IMS" name="absolute position offset x" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="0"/>
+      <cvParam accession="IMS:1000054" cvRef="IMS" name="absolute position offset y" unitAccession="UO:0000017" unitCvRef="UO" unitName="micrometer" value="0"/>
+    </scanSettings>
+  </scanSettingsList>
+</mzML>
diff --git a/tests/integration/imzml_parser/test_parse_metadata.py b/tests/integration/imzml_parser/test_parse_metadata.py
@@ -0,0 +1,57 @@
+from pathlib import Path
+from xml.etree import ElementTree
+
+import pytest
+
+from depiction.persistence.imzml.parser.parse_metadata import ParseMetadata
+from depiction.persistence.pixel_size import PixelSize
+
+
+@pytest.fixture()
+def xml_path(request) -> Path:
+    return Path(__file__).parent / "chunks" / "parse_metadata" / f"{request.param}.xml"
+
+
+@pytest.fixture()
+def etree(xml_path) -> ElementTree.ElementTree:
+    return ElementTree.ElementTree(ElementTree.fromstring(xml_path.read_text()))
+
+
+@pytest.fixture()
+def parse_metadata(etree) -> ParseMetadata:
+    return ParseMetadata(etree=etree)
+
+
+@pytest.fixture()
+def expected_checksums(xml_path) -> dict[str, str]:
+    if xml_path.stem == "checksums_none":
+        return {}
+    elif xml_path.stem == "checksums_md5":
+        return {"md5": "00000000000111111111112222222222"}
+    elif xml_path.stem == "checksums_sha1":
+        return {"sha1": "abcdef"}
+    elif xml_path.stem == "checksums_sha256":
+        return {"sha256": "aaaaaa"}
+    elif xml_path.stem == "checksums_multiple":
+        return {"md5": "00000000000111111111112222222222", "sha256": "aaaaaa"}
+    else:
+        raise NotImplementedError
+
+
+@pytest.mark.parametrize(
+    "xml_path",
+    ["checksums_none", "checksums_md5", "checksums_sha1", "checksums_sha256", "checksums_multiple"],
+    indirect=True,
+)
+def test_ibd_checksums(parse_metadata: ParseMetadata, expected_checksums) -> None:
+    assert parse_metadata.ibd_checksums == expected_checksums
+
+
+@pytest.mark.parametrize("xml_path", ["pixel_size_2d", "pixel_size_none"], indirect=True)
+def test_pixel_size(parse_metadata: ParseMetadata, xml_path: Path) -> None:
+    if xml_path.stem == "pixel_size_2d":
+        assert parse_metadata.pixel_size == PixelSize(size_x=50, size_y=20, unit="micrometer")
+    elif xml_path.stem == "pixel_size_none":
+        assert parse_metadata.pixel_size is None
+    else:
+        raise NotImplementedError