Merge pull request #146 from joe-antognini/jma/arxiv

Arxiv support for pubs add.

pubs/apis.py

@@ -1,9 +1,55 @@
 """Interface for Remote Bibliographic APIs"""
 
 import requests
+import bibtexparser
+from bibtexparser.bibdatabase import BibDatabase
+import feedparser
 from bs4 import BeautifulSoup
 
 
+class ReferenceNotFoundException(Exception):
+    pass
+
+
+def get_bibentry_from_api(id_str, id_type, rp):
+    """Return a bibtex string from various ID methods.
+
+    This is a wrapper around functions that will return a bibtex string given
+    one of:
+
+    * DOI
+    * IBSN
+    * arXiv ID
+
+    Args:
+        id_str: A string with the ID.
+        id_type: Name of the ID type.  Must be one of `doi`, `isbn`, or `arxiv`.
+        rp: A `Repository` object.
+
+    Returns:
+        A bibtex string.
+
+    Raises:
+        ValueError: if `id_type` is not one of `doi`, `isbn`, or `arxiv`.
+    """
+
+    id_fns = {
+        'doi': doi2bibtex,
+        'isbn': isbn2bibtex,
+        'arxiv': arxiv2bibtex,
+    }
+
+    if id_type not in id_fns.keys():
+        raise ValueError('id_type must be one of `doi`, `isbn`, or `arxiv`.')
+
+    bibentry_raw = id_fns[id_type](id_str)
+    bibentry = rp.databroker.verify(bibentry_raw)
+    if bibentry is None:
+        raise ReferenceNotFoundException(
+            'invalid {} {} or unable to retrieve bibfile from it.'.format(id_type, id_str))
+    return bibentry
+
+
 def doi2bibtex(doi):
     """Return a bibtex string of metadata from a DOI"""
 
@@ -25,3 +71,32 @@ def isbn2bibtex(isbn):
     citation = soup.find("textarea").text
 
     return citation
+
+
+def arxiv2bibtex(arxiv_id):
+    """Return a bibtex string of metadata from an arXiv ID"""
+
+    url = 'https://export.arxiv.org/api/query?id_list=' + arxiv_id
+    r = requests.get(url)
+    feed = feedparser.parse(r.text)
+    entry = feed.entries[0]
+
+    if 'title' not in entry:
+        raise ReferenceNotFoundException('arXiv ID not found.')
+    elif 'arxiv_doi' in entry:
+        bibtex = doi2bibtex(entry['arxiv_doi'])
+    else:
+        # Create a bibentry from the metadata.
+        db = BibDatabase()
+        author_str = ' and '.join(
+            [author['name'] for author in entry['authors']])
+        db.entries = [{
+            'ENTRYTYPE': 'article',
+            'ID': arxiv_id,
+            'author': author_str,
+            'title': entry['title'],
+            'year': str(entry['published_parsed'].tm_year),
+            'Eprint': arxiv_id,
+        }]
+        bibtex = bibtexparser.dumps(db)
+    return bibtex

pubs/commands/add_cmd.py

@@ -26,8 +26,10 @@ def parser(subparsers, conf):
     parser = subparsers.add_parser('add', help='add a paper to the repository')
     parser.add_argument('bibfile', nargs='?', default=None,
                         help='bibtex file')
-    parser.add_argument('-D', '--doi', help='doi number to retrieve the bibtex entry, if it is not provided', default=None, action=ValidateDOI)
+    id_arg = parser.add_mutually_exclusive_group()
-    parser.add_argument('-I', '--isbn', help='isbn number to retrieve the bibtex entry, if it is not provided', default=None)
+    id_arg.add_argument('-D', '--doi', help='doi number to retrieve the bibtex entry, if it is not provided', default=None, action=ValidateDOI)
+    id_arg.add_argument('-I', '--isbn', help='isbn number to retrieve the bibtex entry, if it is not provided', default=None)
+    id_arg.add_argument('-X', '--arxiv', help='arXiv ID to retrieve the bibtex entry, if it is not provided', default=None)
     parser.add_argument('-d', '--docfile', help='pdf or ps file', default=None)
     parser.add_argument('-t', '--tags', help='tags associated to the paper, separated by commas',
                         default=None
@@ -85,24 +87,22 @@ def command(conf, args):
 
     # get bibtex entry
     if bibfile is None:
-        if args.doi is None and args.isbn is None:
+        if args.doi is None and args.isbn is None and args.arxiv is None:
             bibentry = bibentry_from_editor(conf, ui, rp)
         else:
-            if args.doi is not None:
+            bibentry = None
-                bibentry_raw = apis.doi2bibtex(args.doi)
+            try:
-                bibentry = rp.databroker.verify(bibentry_raw)
+                if args.doi is not None:
-                if bibentry is None:
+                    bibentry = apis.get_bibentry_from_api(args.doi, 'doi', rp)
-                    ui.error('invalid doi {} or unable to retrieve bibfile from it.'.format(args.doi))
+                elif args.isbn is not None:
-                    if args.isbn is None:
+                    bibentry = apis.get_bibentry_from_api(args.isbn, 'isbn', rp)
-                        ui.exit(1)
+                    # TODO distinguish between cases, offer to open the error page in a webbrowser.
-            if args.isbn is not None:
+                    # TODO offer to confirm/change citekey
-                bibentry_raw = apis.isbn2bibtex(args.isbn)
+                elif args.arxiv is not None:
-                bibentry = rp.databroker.verify(bibentry_raw)
+                    bibentry = apis.get_bibentry_from_api(args.arxiv, 'arxiv', rp)
-                if bibentry is None:
+            except apis.ReferenceNotFoundException as e:
-                    ui.error('invalid isbn {} or unable to retrieve bibfile from it.'.format(args.isbn))
+                ui.error(e.message)
-                    ui.exit(1)
+                ui.exit(1)
-                # TODO distinguish between cases, offer to open the error page in a webbrowser.
-                # TODO offer to confirm/change citekey
     else:
         bibentry_raw = content.get_content(bibfile, ui=ui)
         bibentry = rp.databroker.verify(bibentry_raw)

requirements.txt

@@ -4,3 +4,4 @@ python-dateutil
 requests
 configobj
 beautifulsoup4
+feedparser

tests/test_apis.py

@@ -7,7 +7,7 @@ import dotdot
 
 from pubs.p3 import ustr
 from pubs.endecoder import EnDecoder
-from pubs.apis import doi2bibtex, isbn2bibtex
+from pubs.apis import arxiv2bibtex, doi2bibtex, isbn2bibtex
 
 
 class TestDOI2Bibtex(unittest.TestCase):
@@ -60,5 +60,30 @@ class TestISBN2Bibtex(unittest.TestCase):
             self.endecoder.decode_bibdata(bib)
 
 
+class TestArxiv2Bibtex(unittest.TestCase):
+
+    def setUp(self):
+        self.endecoder = EnDecoder()
+
+    def test_parses_to_bibtex_with_doi(self):
+        bib = arxiv2bibtex('astro-ph/9812133')
+        b = self.endecoder.decode_bibdata(bib)
+        self.assertEqual(len(b), 1)
+        entry = b[list(b)[0]]
+        self.assertEqual(entry['author'][0], 'Perlmutter, S.')
+        self.assertEqual(entry['year'], '1999')
+
+    def test_parses_to_bibtex_without_doi(self):
+        bib = arxiv2bibtex('math/0211159')
+        b = self.endecoder.decode_bibdata(bib)
+        self.assertEqual(len(b), 1)
+        entry = b[list(b)[0]]
+        self.assertEqual(entry['author'][0], 'Perelman, Grisha')
+        self.assertEqual(entry['year'], '2002')
+        self.assertEqual(
+                entry['title'],
+                'The entropy formula for the Ricci flow and its geometric applications')
+
+
 # Note: apparently ottobib.com uses caracter modifiers for accents instead
 # of the correct unicode characters. TODO: Should we convert them?