fix(cff): doi structure parsing (#121)

* fix: DOI from dict, not flat value

* fix:make cff example correct

* docs: adapt docstring example to match real CFF structure

* fix:docstring still fucked

* fix:typo docstring

* fix:chatGPT's suggestion for docstring formatting

* fix:OK, no multiline, and double escape newlines

* feat: support multiple DOI's

* refactor(cff): reduce nesting

* chore(cff): use type hint for list from python standard collection

Co-authored-by: Cyril Matthey-Doret <cyril.matthey-doret@epfl.ch>

---------

Co-authored-by: Cyril Matthey-Doret <cyril.matthey-doret@epfl.ch>
Co-authored-by: cmdoret <cyril.mattheydoret@gmail.com>

gimie/parsers/cff.py

@@ -34,21 +34,22 @@ def __init__(self, subject: str):
         super().__init__(subject)
 
     def parse(self, data: bytes) -> Graph:
-        """Extracts a DOI link and list of authors from a CFF file and returns a
-        graph with a single triple <subject> <schema:citation> <doi>
+        """Extracts DOIs and list of authors from a CFF file and returns a
+        graph with triples <subject> <schema:citation> <doi>
         and a number of author objects with <schema:name> and <md4i:orcid> values.
-        If no DOI is found, it will not be included in the graph.
-        If no authors are found, it will not be included in the graph.
-        If neither authors nor DOI are found, an empty graph is returned.
+        If no DOIs are found, they will not be included in the graph.
+        If no authors are found, they will not be included in the graph.
+        If neither authors nor DOIs are found, an empty graph is returned.
         """
         extracted_cff_triples = Graph()
-        doi = get_cff_doi(data)
+        dois = get_cff_doi(data)
         authors = get_cff_authors(data)
 
-        if doi:
-            extracted_cff_triples.add(
-                (self.subject, SDO.citation, URIRef(doi))
-            )
+        if dois:
+            for doi in dois:
+                extracted_cff_triples.add(
+                    (self.subject, SDO.citation, URIRef(doi))
+                )
         if not authors:
             return extracted_cff_triples
         for author in authors:
@@ -119,8 +120,8 @@ def doi_to_url(doi: str) -> str:
     return f"https://doi.org/{doi_match}"
 
 
-def get_cff_doi(data: bytes) -> Optional[str]:
-    """Given a CFF file, returns the DOI, if any.
+def get_cff_doi(data: bytes) -> Optional[list[str]]:
+    """Given a CFF file, returns a list of DOIs, if any.
 
     Parameters
     ----------
@@ -129,15 +130,16 @@ def get_cff_doi(data: bytes) -> Optional[str]:
 
     Returns
     -------
-    str, optional
-        doi formatted as a valid url
+    list of str, optional
+        DOIs formatted as valid URLs
 
     Examples
     --------
-    >>> get_cff_doi(bytes("doi:   10.5281/zenodo.1234", encoding="utf8"))
-    'https://doi.org/10.5281/zenodo.1234'
+    >>> get_cff_doi(bytes("identifiers:\\n    - type: doi\\n      value: 10.5281/zenodo.1234\\n    - type: doi\\n      value: 10.5281/zenodo.5678", encoding="utf8"))
+    ['https://doi.org/10.5281/zenodo.1234', 'https://doi.org/10.5281/zenodo.5678']
+    >>> get_cff_doi(bytes("identifiers:\\n    - type: doi\\n      value: 10.5281/zenodo.9012", encoding="utf8"))
+    ['https://doi.org/10.5281/zenodo.9012']
     >>> get_cff_doi(bytes("abc: def", encoding="utf8"))
-
     """
 
     try:
@@ -146,18 +148,25 @@ def get_cff_doi(data: bytes) -> Optional[str]:
         logger.warning("cannot read CITATION.cff, skipped.")
         return None
 
+    doi_urls = []
+
     try:
-        doi_url = doi_to_url(cff["doi"])
-    # No doi in cff file
+        identifiers = cff["identifiers"]
     except (KeyError, TypeError):
-        logger.warning("CITATION.cff does not contain a 'doi' key.")
-        doi_url = None
-    # doi is malformed
-    except ValueError as err:
-        logger.warning(err)
-        doi_url = None
-
-    return doi_url
+        logger.warning(
+            "CITATION.cff does not contain a valid 'identifiers' key."
+        )
+        return None
+
+    for identifier in identifiers:
+        if identifier.get("type") == "doi":
+            try:
+                doi_url = doi_to_url(identifier["value"])
+                doi_urls.append(doi_url)
+            except ValueError as err:
+                logger.warning(err)
+
+    return doi_urls or None
 
 
 def get_cff_authors(data: bytes) -> Optional[List[dict[str, str]]]:

tests/test_cff.py

@@ -73,12 +73,23 @@ def test_broken_cff(cff_file):
 def test_parse_doi():
     cff_file = b"""
     cff-version: 1.2.0
-    title: gimie
-    doi: 10.5281/zenodo.1234567
+    message: If you use this software, please cite it using these metadata.
+    title: 'napari: a multi-dimensional image viewer for Python'
+    identifiers:
+    - type: doi
+      value: 10.5281/zenodo.3555620
+    - type: doi
+      value: 10.21105/joss.01274
     """
-    obj = next(
+    parsed_dois = list(
         CffParser(subject=URIRef("https://example.org/"))
         .parse(data=cff_file)
         .objects()
     )
-    assert URIRef("https://doi.org/10.5281/zenodo.1234567") == obj
+    expected_dois = [
+        URIRef("https://doi.org/10.5281/zenodo.3555620"),
+        URIRef("https://doi.org/10.21105/joss.01274"),
+    ]
+    # parsed_dois already contains all parsed DOI objects
+    for doi in expected_dois:
+        assert doi in parsed_dois