generate citation from csl record

docs/usage.rst

@@ -13,9 +13,22 @@ To create a CITATION.cff file of a release, you must supply the doi that is asso
 
 The generated file must be checked for correctness and you are encouraged to enrich it further.
 
+The command above works well for Zenodo doi's, which contain especially rich metadata.
+It is also possible to produce a CITATION.cff file from any available information. This feature is experimental:
+
+.. code-block:: bash
+
+   doi2cff init --experimental <doi>
+
+   # For example
+   doi2cff init 10.1051/0004-6361/202037850 --experimental
+
+
 Whenever a new release is made of the software the CITATION.cff must be updated with new doi/version/release date.
-This process can be automated by running
+This process can be automated by running (only for Zenodo dois)
 
 .. code-block:: bash
 
     doi2cff update <doi>
+
+

doi2cff/cli.py

@@ -4,6 +4,7 @@
 import sys
 import re
 from datetime import datetime
+from typing import Any, Tuple
 
 import click
 from nameparser import HumanName
@@ -20,6 +21,7 @@ def main():
     Current supported DOI types:
 
         * Zenodo upload of a GitHub release (https://guides.github.com/activities/citable-code/)
+        * Any DOIs with suitable metadata (experimental).
 
     """
     pass
@@ -29,21 +31,85 @@ def is_software_zenodo(zenodo_record):
     return zenodo_record['metadata']['resource_type']['type'] == 'software'
 
 
+def zenodo_record_to_cff_yaml(zenodo_record: dict, template) -> Tuple[ruamel.yaml.YAML, Any]:
+    yaml = ruamel.yaml.YAML()
+    data = yaml.load(template)
+    data['title'] = zenodo_record['metadata']['title']
+    data['doi'] = zenodo_record['doi']
+    tagurl = tagurl_of_zenodo(zenodo_record)
+    if 'version' in zenodo_record['metadata']:
+        data['version'] = re.sub('^(v)', '', zenodo_record['metadata']['version'])
+    else:
+        data['version'] = tagurl2version(tagurl)
+    data['license'] = zenodo_record['metadata']['license']['id']
+    data['date-released'] = datetime.strptime(zenodo_record['metadata']['publication_date'], "%Y-%m-%d").date()
+    data['repository-code'] = tagurl2repo(tagurl)
+    data['authors'] = authors_of_zenodo(zenodo_record)
+    references = references_of_zenodo(zenodo_record)
+    fixme = 'FIXME generic is too generic, ' \
+            'see https://citation-file-format.github.io/1.2.0/specifications/#/reference-types for more specific types'
+    if references:
+        data['references'] = yaml.seq(references)
+        for idx, r in enumerate(references):
+            if r['type'] == 'generic':
+                data['references'].yaml_add_eol_comment(fixme, idx)
+
+    return yaml, data
+
+
+def csljson_to_cff_yaml(cffjson: dict, template) -> Tuple[ruamel.yaml.YAML, Any]:
+    yaml = ruamel.yaml.YAML()
+    data = yaml.load(template)
+
+    data['title'] = cffjson['title']
+    if '\n' in data['title']:
+        data.yaml_add_eol_comment("FIXME: title contains new line: this is strange", "title")
+
+    data['doi'] = cffjson['DOI']
+
+    if 'license' in cffjson:
+        data['license'] = cffjson['license']
+
+    data['date-released'] = datetime(*cffjson['published']['date-parts'][0], 1).date()
+    data['authors'] = authors_of_csl(cffjson)
+
+    references = cffjson.get('reference', None)
+    fixme = 'FIXME generic is too generic, ' \
+            'see https://citation-file-format.github.io/1.2.0/specifications/#/reference-types for more specific types'
+    if references:
+        data['references'] = yaml.seq(references)
+        for idx, r in enumerate(references):
+            if r.get('type', 'generic') == 'generic':
+                data['references'].yaml_add_eol_comment(fixme, idx)
+
+    # In CFF 1.2.0 these fields are optional
+    # https://github.com/citation-file-format/citation-file-format/releases/tag/1.2.0
+    del data['version']
+    del data['repository-code']    
+
+    return yaml, data
+
+
 @main.command()
 @click.argument('doi')
 @click.option('--cff_fn',
               type=click.File('x'),
               default='CITATION.cff',
               help='Name of citation formatted output file',
               show_default=True)
-def init(doi, cff_fn):
+@click.option('--experimental/--no-experimental',
+              is_flag=True,
+              default=False,
+              help='experimental parsing of non-zenodo links',
+              show_default=True)
+def init(doi, cff_fn, experimental):
     """Generate CITATION.cff file based on a Zenodo DOI of a Github release.
 
     * DOI, The Digital Object Identifier (DOI) name of a Zenodo upload of a GitHub release
     """
     template = '''# YAML 1.2
 # Metadata for citation of this software according to the CFF format (https://citation-file-format.github.io/)
-cff-version: 1.0.3
+cff-version: 1.2.0
 message: If you use this software, please cite it using these metadata.
 # FIXME title as repository name might not be the best name, please make human readable
 title: x
@@ -56,35 +122,24 @@ def init(doi, cff_fn):
 license: x
     '''
 
-    if not doi_is_from_zenodo(doi):
-        raise click.UsageError('Unable to process DOI name, only accept DOI name which is a Zenodo upload')
+    if doi_is_from_zenodo(doi):
+        zenodo_record = fetch_zenodo_by_doiurl(doi)
 
-    zenodo_record = fetch_zenodo_by_doiurl(doi)
+        if not is_software_zenodo(zenodo_record):
+            raise click.UsageError('Unable to process DOI name, only accept DOI name '
+                                   'which is a Zenodo upload of type software')
 
-    if not is_software_zenodo(zenodo_record):
-        raise click.UsageError('Unable to process DOI name, only accept DOI name which is a Zenodo upload of type software')
-
-    yaml = ruamel.yaml.YAML()
-    data = yaml.load(template)
-    data['title'] = zenodo_record['metadata']['title']
-    data['doi'] = zenodo_record['doi']
-    tagurl = tagurl_of_zenodo(zenodo_record)
-    if 'version' in zenodo_record['metadata']:
-        data['version'] = re.sub('^(v)', '', zenodo_record['metadata']['version'])
+        yaml, data = zenodo_record_to_cff_yaml(zenodo_record, template)
     else:
-        data['version'] = tagurl2version(tagurl)
-    data['license'] = zenodo_record['metadata']['license']['id']
-    data['date-released'] = datetime.strptime(zenodo_record['metadata']['publication_date'], "%Y-%m-%d").date()
-    data['repository-code'] = tagurl2repo(tagurl)
-    data['authors'] = authors_of_zenodo(zenodo_record)
-    references = references_of_zenodo(zenodo_record)
-    fixme = 'FIXME generic is too generic, ' \
-            'see https://citation-file-format.github.io/1.0.3/specifications/#/reference-types for more specific types'
-    if references:
-        data['references'] = yaml.seq(references)
-        for idx, r in enumerate(references):
-            if r['type'] == 'generic':
-                data['references'].yaml_add_eol_comment(fixme, idx)
+        if experimental:
+            click.echo("Trying experimental parsing of arbitrary DOI")
+            csljson = fetch_csljson(doi)
+            yaml, data = csljson_to_cff_yaml(csljson, template)
+        else:
+            raise click.UsageError('Unable to process DOI name, normally we only accept DOI name '
+                                   'which is a Zenodo upload'
+                                   'You can try experimental parsing of other DOIs '
+                                   '(see --experimental option).')
 
     yaml.dump(data, cff_fn)
 
@@ -103,6 +158,12 @@ def update(doi, cff_fn):
 
     * DOI, The Digital Object Identifier (DOI) name of a Zenodo upload of a GitHub release
     """
+
+    if not doi_is_from_zenodo(doi):
+        raise click.UsageError('CITATION.cff update is only possible with Zenodo DOI. '
+                               'For non-Zenodo DOIs, please consider recreating the citation, enabling experimental features: '
+                               '`doi2cff init <doi> --experimental`')
+
     update_version(doi, cff_fn)