Add protection finding authors (trigger by ins1386663).

modules/harvest_tools/base.py

@@ -5,6 +5,7 @@
 import re
 
 from gluon import current
+from invenio_tools import REG_AUTHOR
 
 DRY_RUN = "dry run"
 
@@ -49,7 +50,7 @@ def format_author_fr(name):
     #                  Family, First
     # To avoid to deal with unicode character
     # look for non empty string \S
-    match = re.match(r'(.+), (\S+)( |\-)*(\S+)*', name)
+    match = REG_AUTHOR.match(name)
 
     # reformat the name as L. Family
     # or keep it as it is

modules/invenio_tools/__init__.py

@@ -10,6 +10,7 @@ from base import (ARXIV,
                   is_thesis,
                   OAI_URL,
                   REG_ARXIV_NUMBER,
+                  REG_AUTHOR,
                   REG_OAI,
                   REG_YEAR,
                   THESIS_DIR)

modules/invenio_tools/base.py

@@ -13,8 +13,20 @@ MSG_NO_THESIS = "Reject no thesis information"
 OAI_URL = "http://%s/record/%s"
 
 REG_ARXIV_NUMBER = re.compile("\d+\.\d+")
-REG_OAI = re.compile('oai:([a-z\.]+):([\d]+)')
-REG_YEAR = re.compile("(\d{4})")
+
+# name are encoded Family, L
+#                  Family, P L
+#                  Family, M -H
+#                  Family Name, J
+#                  Family-Name, J
+#                  Family, F Name
+#                  Family, First
+# To avoid to deal with unicode character
+# look for non empty string \S
+REG_AUTHOR = re.compile(r"(.+), (\S+)( |\-)*(\S+)*")
+
+REG_OAI = re.compile(r"oai:([a-z\.]+):([\d]+)")
+REG_YEAR = re.compile(r"(\d{4})")
 
 THESIS_DIR = u"dir."
 

modules/invenio_tools/checkandfix.py

@@ -5,7 +5,12 @@
 import re
 import regex
 
-from base import MSG_NO_CONF, MSG_NO_THESIS, OAI_URL, REG_OAI, REG_YEAR
+from base import (MSG_NO_CONF,
+                  MSG_NO_THESIS,
+                  OAI_URL,
+                  REG_AUTHOR,
+                  REG_OAI,
+                  REG_YEAR)
 from exception import CheckException
 from filters import CLEAN_REVIEW
 from gluon import current
@@ -339,13 +344,33 @@ class CheckAndFix(object):
         for key in (u"100", u"700"):
             if key in record:
                 if isinstance(record[key], list):
-                    for i in range(len(record[key])):
+                    for i in xrange(len(record[key])):
                         if "a" in record[key][i]:
-                            record[key][i]["a"] = func(record[key][i]["a"])
+
+                            # PROTECTION
+                            # see RecordPubli.author_as_list
+                            value = record[key][i]["a"]
+                            if isinstance(value, unicode):
+                                record[key][i]["a"] = func(value)
+
+                            elif isinstance(value, list):
+                                for elt in value:
+                                    if REG_AUTHOR.match(elt):
+                                        record[key][i]["a"] = func(elt)
 
                 else:
                     if "a" in record[key]:
-                        record[key]["a"] = func(record[key]["a"])
+                        value = record[key]["a"]
+
+                        # PROTECTION
+                        # see RecordPubli.authors_as_list
+                        if isinstance(value, unicode):
+                            record[key]["a"] = func(value)
+
+                        elif isinstance(value, list):
+                            for elt in value:
+                                if REG_AUTHOR.match(elt):
+                                    record[key][i]["a"] = func(elt)
 
     def format_editor(self, record):
         """Format the editor abbreviation since the encoding

modules/invenio_tools/recordpubli.py

@@ -5,7 +5,12 @@
 import re
 
 
-from base import ARXIV, ARXIV_PDF, REG_ARXIV_NUMBER, REG_YEAR, THESIS_DIR
+from base import (ARXIV,
+                  ARXIV_PDF,
+                  REG_ARXIV_NUMBER,
+                  REG_AUTHOR,
+                  REG_YEAR,
+                  THESIS_DIR)
 from filters import CLEAN_COLLABORATION
 from plugin_dbui import CLEAN_SPACES
 from record import Record
@@ -66,7 +71,8 @@ class RecordPubli(Record):
         """
         authors = []
 
-        # NOTE: the content of the 700 field depend on the record type.
+        # NOTE
+        # the content of the 700 field depend on the record type.
         # For thesis it also contains the name of the director
         if u"700" in self and isinstance(self[u"700"], dict):
             if not ("e" in self[u"700"] and self[u"700"]["e"] == THESIS_DIR):
@@ -78,7 +84,20 @@ class RecordPubli(Record):
                     continue
 
                 if "a" in di:
-                    authors.append(di["a"])
+                    author = di["a"]
+
+                    # PROTECTION
+                    # in most of the case the author is a string
+                    # but it can be a list, e.g inspirehep.net/138663:
+                    # [u'Zuniga, J.', u'(the A.N.T.ARES. Collaboration)']
+                    if isinstance(author, unicode):
+                        authors.append(di["a"])
+
+                    elif isinstance(author, list):
+                        for elt in author:
+                            if REG_AUTHOR.match(elt):
+                                authors.append(elt)
+                                break
 
         return authors
 

tests/harvester/CheckAndFix/test_checkandfix_non_conformities.py

+# -*- coding: utf-8 -*-
+"""CheckAndFix non conformities
+
+"""
+from invenio_tools import CheckAndFix, load_record
+from harvest_tools import format_author_fr
+
+
+def test_protection_format_authors():
+    """[u'Zuniga, J.', u'(the A.N.T.ARES. Collaboration)']"""
+    record = load_record('inspirehep.net', 1386663)
+
+    svc = CheckAndFix()
+    svc.format_authors(record, format_author_fr)
+
+    authors = record[u"700"]
+
+    assert len(authors) == 139
+    assert authors[128]["a"] == u"J. Zuniga"
+
+

tests/harvester/Record/test_record_non_conformities.py

@@ -8,6 +8,14 @@ Protection are add in the record method to correct them
 from invenio_tools import load_record
 
 
+def test_protection_authors_as_list():
+    """[u'Zuniga, J.', u'(the A.N.T.ARES. Collaboration)']"""
+    record = load_record('inspirehep.net', 1386663)
+    authors = record.authors_as_list()
+    assert len(authors) == 139
+    assert authors[128] == u"Zuniga, J."
+
+
 def test_protection_oai():
     """['oai:cds.cern.ch:1513204', 'oai:cds.cern.ch:1512766']"""
     record = load_record('cds.cern.ch', 1513204)