Move CSL-JSON builder from commented code to own script

scripts/build_zenodo_json.py

@@ -102,58 +102,6 @@ def convert_source_to_reference(source: pd.Series) -> str:
   description_html = description_html.replace(match.group(0), match.group(2))
 
 
-# --- Build references from CSL ----
-# import citeproc
-# import citeproc_styles
-
-
-# def convert_source_to_csl(source: pd.Series) -> dict:
-#   """Convert source to CSL-JSON."""
-#   source = source.astype(object, copy=True)
-#   source[source.isnull()] = None
-#   csl = {
-#     'id': source['id'],
-#     'author': [
-#       {'literal': author}
-#       for author in source['author'].split(' | ')
-#     ],
-#     'issued': {'date-parts': [[int(source['year'])]]},
-#     'type': source['type'],
-#     'title': source['title'],
-#     'URL': source['url'],
-#     'language': source['language'],
-#     'container-title': source['container_title'],
-#     'volume': source['volume'],
-#     'issue': source['issue'],
-#     'page': source['page'],
-#     'version': source['version'],
-#     'editor': [
-#       {'literal': editor}
-#       for editor in source['editor'].split(' | ')
-#     ] if source['editor'] else None,
-#     'collection-title': source['collection_title'],
-#     'collection-number': source['collection_number'],
-#     'publisher': source['publisher']
-#   }
-#   # Remove None values
-#   return {key: value for key, value in csl.items() if value is not None}
-
-
-# csl_json = [
-#   convert_source_to_csl(sources.loc[i])
-#   for i in sources.query('type.ne("personal-communication")').index
-# ]
-# csl = citeproc.source.json.CiteProcJSON(csl_json)
-# style_path = citeproc_styles.get_style_filepath('apa')
-# style = citeproc.CitationStylesStyle(style_path)
-# bibliography = citeproc.CitationStylesBibliography(style, csl, citeproc.formatter.plain)
-# for key in csl:
-#   citation = citeproc.Citation([citeproc.CitationItem(key)])
-#   bibliography.register(citation)
-
-# references = [str(item) for item in bibliography.bibliography()]
-
-
 # ---- Build zenodo.json ----
 
 zenodo = {

scripts/print_source_csljson.py

@@ -0,0 +1,63 @@
+import json
+from typing import Dict
+
+import pandas as pd
+
+
+def convert_source_to_csl(source: Dict[str, str]) -> dict:
+  """
+  Convert source to CSL-JSON.
+
+  Names of people are currently represented with their full names only:
+  `{"literal": "Name"}`.
+  Extracting given and family names could be achieved with the following heuristics,
+  depending on the writing system:
+
+  * Latin: Assume last word is the family name
+    (e.g. "Jakob F. Steiner" -> {"family": "Steiner", "given": "Jakob F."}).
+    Although this assumption is valid in most cases, it is wrong when
+    particles (e.g. "Ward J. J. Van Pelt", "Roderik S. W. van de Wal"),
+    suffixes (e.g. "E. Calvin Alexander Jr."),
+    or double family names ("Guillermo Cobos Campos") are present.
+  * Cyrillic: Assume last word of the original and transliteration is the
+    family name (e.g. "Н. Г. Разумейко [N. G. Razumeiko]" ->
+    {"family": "Иванов [Razumeiko]", "given": "Н. Г. [N. G.]"}).
+  * Chinese: Assume first character and word of the original and
+    transliteration, respectively, is the family name (e.g. "孙维君 [Sun Weijun]" ->
+    "family": "孙 [Sun]", "given": "Weijun [维君]").
+  * Hangul: Same as for Chinese (e.g. "안진호 [Ahn Jinho]" ->
+    {"family": "안 [Ahn]", "given": "진호 [Jinho]"}).
+  """
+  csl = {
+    'id': source['id'],
+    'author': [
+      {'literal': author}
+      for author in source['author'].split(' | ')
+    ],
+    'issued': {'date-parts': [[int(source['year'])]]},
+    'type': source['type'],
+    'title': source['title'],
+    'URL': source['url'],
+    'language': source['language'],
+    'container-title': source['container_title'],
+    'volume': source['volume'],
+    'issue': source['issue'],
+    'page': source['page'],
+    'version': source['version'],
+    'editor': [
+      {'literal': editor}
+      for editor in source['editor'].split(' | ')
+    ] if source['editor'] else None,
+    'collection-title': source['collection_title'],
+    'collection-number': source['collection_number'],
+    'publisher': source['publisher']
+  }
+  # Remove None values
+  return {key: value for key, value in csl.items() if value is not None}
+
+
+# ---- Convert source table to CSL-JSON ----
+
+sources = pd.read_csv('data/source.csv', dtype='string').replace({pd.NA: None})
+csl = [convert_source_to_csl(source) for source in sources.to_dict(orient='records')]
+print(json.dumps(csl, indent=2, ensure_ascii=False))