Add a citations fetcher for bibtex files

libbmc/citations/bbl.py

@@ -7,7 +7,7 @@ import re
 import subprocess
 
 from libbmc import tools
-from libbmc.citations import bbl
+from libbmc.citations import plaintext
 
 
 # Regex to match bibitems
@@ -18,7 +18,7 @@ ENDTHEBIBLIOGRAPHY_REGEX = re.compile(r"\\end\{thebibliography}.*")
 
 def bibitem_as_plaintext(bibitem):
     """
-    Return a plaintext representation of the bibitem from the ``.bbl`` file.
+    Return a plaintext representation of a bibitem from the ``.bbl`` file.
 
     .. note::
 
@@ -67,16 +67,16 @@ def get_plaintext_citations(bbl):
     return cleaned_bbl
 
 
-def get_cited_DOIs(bbl_input):
+def get_cited_DOIs(bbl):
     """
-    Get the DOIs of the papers cited in this .bbl file.
+    Get the DOIs of the papers cited in a .bbl file.
 
-    :param bbl_input: Either the path to a .bbl file or the content \
+    :param bbl: Either the path to a .bbl file or the content \
             of a .bbl file.
 
     :returns: A dict of cleaned plaintext citations and their associated DOI.
     """
     # Get the plaintext citations from the bbl file
-    plaintext_citations = get_plaintext_citations(bbl_input)
+    plaintext_citations = get_plaintext_citations(bbl)
     # Use the plaintext citations parser on these citations
-    return bbl.get_cited_DOIs(plaintext_citations)
+    return plaintext.get_cited_DOIs(plaintext_citations)

libbmc/citations/bibtex.py

@@ -0,0 +1,74 @@
+"""
+This files contains all the functions to extract DOIs of citations from
+BibTeX files.
+"""
+import bibtexparser
+import os
+
+from bibtexparser.bparser import BibTexParser
+from bibtexparser.customization import convert_to_unicode
+
+from libbmc import tools
+from libbmc.citations import plaintext
+
+
+def bibentry_as_plaintext(bibentry):
+    """
+    Return a plaintext representation of a bibentry from BibTeX file.
+
+    .. note::
+
+        This plaintext representation can be super ugly, contain URLs and so \
+        on.
+
+    :param bibentry: A bibentry as parsed by ``bibtexparser``.
+    :returns: A cleaned plaintext citation from the bibentry.
+    """
+    # Just flatten the bibentry
+    return tools.clean_whitespaces(" ".join([bibentry[k] for k in bibentry]))
+
+
+def get_plaintext_citations(bibtex):
+    """
+    Parse a BibTeX file to get a clean list of plaintext citations.
+
+    :param bibtex: Either the path to the BibTeX file or the content of a \
+            BibTeX file.
+    :returns:  A list of cleaned plaintext citations.
+    """
+    parser = BibTexParser()
+    parser.customization = convert_to_unicode
+    # Load the BibTeX
+    if os.path.isfile(bibtex):
+        with open(bibtex) as fh:
+            bib_database = bibtexparser.load(fh, parser=parser)
+    else:
+        bib_database = bibtexparser.loads(bibtex, parser=parser)
+    # Convert bibentries to plaintext
+    bibentries = [bibentry_as_plaintext(bibentry)
+                  for bibentry in bib_database.entries]
+    # Return them
+    return bibentries
+
+
+def get_cited_DOIs(bibtex):
+    """
+    Get the DOIs of the papers cited in a BibTeX file.
+
+    .. note::
+
+        For now, this function is actually flattening the BibTeX file \
+                (loosing any structure provided by the BibTeX) and calling \
+                the matching method for plaintext citations, relying on \
+                CrossRef API. This is the best method I have found so far, \
+                although it can be quite frustrating. Let me know if you have \
+                anything better!
+
+    :param bibtex: Either the path to a BibTeX file or the content of a \
+            BibTeX file.
+    :returns: A dict of cleaned plaintext citations and their associated DOI.
+    """
+    # Get the plaintext citations from the bbl file
+    plaintext_citations = get_plaintext_citations(bibtex)
+    # Use the plaintext citations parser on these citations
+    return plaintext.get_cited_DOIs(plaintext_citations)

requirements.txt

@@ -1,3 +1,4 @@
 arxiv2bib>=1.0.7
+bibtexparser>=0.6.2
 isbnlib>=3.5.7
 requests>=2.9.1