Refactored set year

Pipfile

@@ -7,7 +7,7 @@ name = "pypi"
 
 [packages]
 
-Django = "*"
+django = "*"
 dj-database-url = "*"
 django-dirtyfields = "*"
 django-filter = "*"
@@ -16,7 +16,7 @@ django-tequila = {git = "https://github.com/epfl-idevelop/django-tequila.git", e
 "django-bootstrap4" = "*"
 gevent = "*"
 gunicorn = "*"
-Markdown = "*"
+markdown = "*"
 "psycopg2-binary" = "*"
 pymarc = "*"
 pytz = "*"
@@ -31,9 +31,9 @@ codecov = "*"
 django-debug-toolbar = "*"
 django-extensions = "*"
 django-nose = "*"
-"Fabric3" = "*"
-factory_boy = "*"
-Faker = "*"
+"fabric3" = "*"
+factory-boy = "*"
+faker = "*"
 "flake8" = "*"
 graypy = "*"
 ipdb = "*"
@@ -44,6 +44,7 @@ nose-progressive = "*"
 selenium = "*"
 requests = "*"
 rope = "*"
+pytest = "*"
 
 
 [requires]

docker-compose-dev.yml

@@ -52,4 +52,4 @@ selenium:
   image: selenium/standalone-chrome-debug:3.11
   ports:
     - "4444:4444"   # Selenium
-    - "5900:5900"   # VNC server, pass is "secret"
+    - "${VNC_PORT}:5900"   # VNC server, pass is "secret"

env/django.env

@@ -19,6 +19,7 @@ SERVER_HOST=${ALLOWED_HOSTS}
 # -> 80 in production
 # -> 8000 or 8080 in dev
 DEV_PORT=8000
+VNC_PORT=5900
 
 # `SITE_PATH` is the root path of the application (it not hosted on a bare domain)
 SITE_PATH=/publications-exports

infoscience_exports/exports/marc21xml.py

@@ -3,7 +3,7 @@
 """
 Parse a marc-21-xml file
 """
-
+from logging import getLogger
 from django.utils.translation import gettext as _
 from django.conf import settings
 from os.path import dirname, splitext
@@ -12,27 +12,22 @@
 from pymarc import marcxml
 import unicodedata
 
+logger = getLogger(__name__)
 
-def get_attributes(subfields):
-    res_value = {}
-    for element1 in subfields:
-        for key1, value1 in element1.items():
-            res_value[key1] = value1
-    return res_value
 
+class Author:
 
-# Authors is a list of a dictionary of author: full name, initial name, url in infoscience
-def set_authors(authors):
-    result = []
-    for author in authors:
-        author_record = {}
-        author_record['fullname'] = author
-        author_record['search_url'] = "{}/search?p={}".format(
+    def __init__(self, author):
+        self.fullname = author
+        self.search_url = "{}/search?p={}".format(
             settings.SITE_DOMAIN, author.replace(",", "+").replace(" ", "+"))
+        self.initname = self.compute_name()
 
-        names = author.split(',')
+    def compute_name(self):
+        names = self.fullname.split(',')
         family = names[0].strip() if len(names) > 0 else ''
         fnames = names[1].split(' ') if len(names) > 1 else ''
+
         initname = ""
         for fname in fnames:
             if not fname:
@@ -53,23 +48,22 @@ def set_authors(authors):
         if family:
             initname += family
 
-        author_record['initname'] = initname
-        result.append(author_record)
+        return initname
 
-    return result
+
+# Authors is a list of Author instance: full name, initial name, url in infoscience
+def set_authors(authors):
+    return [Author(author) for author in authors]
 
 
 # get only the year in a date-string
 def set_year(date):
-    if len(date) == 4:
-        return date
-    dates = date.split("-")
-    year = date
-    for val in dates:
-        if len(val) == 4:
-            year = val
-            break
-    return year
+    dates = [val for val in date.split("-") if len(val) == 4]
+    if len(dates) == 1:
+        return dates[0]
+    else:
+        logger.warning("Year not found in %s, decomposed as %s", date, dates)
+        return ''
 
 
 # get fulltext: link to pdf or link to repository if several links
@@ -110,6 +104,14 @@ def set_fulltext(fulltexts):
     return result
 
 
+def get_attributes(subfields):
+    res_value = {}
+    for element1 in subfields:
+        for key1, value1 in element1.items():
+            res_value[key1] = value1
+    return res_value
+
+
 # get dictionary (icon, fulltexts) of ELA
 def get_ELA_fields(field):
     ela_fulltexts = []

infoscience_exports/exports/options_notices.py

@@ -13,7 +13,7 @@
 
 
 DOC_TYPE_ORDERED = {
-    #('ARTICLE', _("Journal Articles")),
+    # ('ARTICLE', _("Journal Articles")),
     'ARTICLE': _("Articles & Reviews"),
     'CONF': _("Conference Papers"),
     'REVIEW': _("Reviews"),
@@ -75,17 +75,17 @@ def get_sorted_by_doc_types(notices):
     # add doc_types not listed in DOC_TYPE_ORDERED
     for index, head in enumerate(groups_head):
         if head not in doc_type_keys:
-            groups_list_ordered.extend(groups_list[index])  
+            groups_list_ordered.extend(groups_list[index])
     return groups_list_ordered
 
 
 def get_sorted_by_year(notices, url):
     queries = parse_qs(urlsplit(url).query)
-    is_ascending = queries.get('so', ['d'])[0] == "a" 
-    if is_ascending:		
-        notices = sorted(notices, key=lambda k: k['Publication_Year'])		
-    else:		
-        notices = sorted(notices, key=lambda k: k['Publication_Year'], reverse=True)		
+    is_ascending = queries.get('so', ['d'])[0] == "a"
+    if is_ascending:
+        notices = sorted(notices, key=lambda k: k['Publication_Year'])
+    else:
+        notices = sorted(notices, key=lambda k: k['Publication_Year'], reverse=True)
     return notices
 
 

infoscience_exports/exports/test/test_authors.py

@@ -0,0 +1,26 @@
+from django.conf import settings
+
+from exports.marc21xml import Author
+
+
+def expected_url(name):
+    return "{}/search?p={}".format(settings.SITE_DOMAIN, name)
+
+
+def test_standard_case():
+    author = Author("Emmanuel Bréton")
+    assert author.fullname == "Emmanuel Bréton"
+    assert author.search_url == expected_url("Emmanuel+Bréton")
+    assert author.initname == "Emmanuel Bréton"
+
+
+def test_search_url():
+    assert Author("Jean-Marc Marco").search_url == expected_url("Jean-Marc+Marco")
+    assert Author("Jean, Marc Marco").search_url == expected_url("Jean++Marc+Marco")
+
+
+def test_initial_names():
+    assert Author("Marco Jean Marc").initname == "Marco Jean Marc"
+    assert Author("Marco, Jean").initname == "J. Marco"
+    assert Author("Marco, Jean Marc").initname == "J. M. Marco"
+    assert Author("Marco, Jean-Marc").initname == "J.-M. Marco"

infoscience_exports/exports/test/test_dates.py

@@ -0,0 +1,14 @@
+from exports.marc21xml import set_year
+
+
+def test_year_only():
+    assert set_year('2018') == '2018'
+
+
+def test_full_date():
+    assert set_year('2018-01-01') == '2018'
+    assert set_year('01-01-2018') == '2018'
+
+
+def test_error():
+    assert set_year('2018.01.01') == ''

infoscience_exports/exports/versions.py

@@ -3,7 +3,7 @@
 
 # the release comes from git and should not be modified
 # => read-only
-_release = '0.3.7-3-g3a4e66e'
+_release = '0.3.7-16-g5e0cae8'
 
 # you can set the next version number manually
 # if you do not, the system will make sure that version > release
@@ -13,5 +13,4 @@
 # the build number will generate conflicts on each PR merge
 # just keep yours every time
 # => read-only
-
-_build = '086e59f5d9fdf7c643968137b2f1fe3936657d3d'
+_build = '5e0cae8b46c21d3448ab566f21295f59d76bbf8f'

requirements-dev.txt

@@ -1,6 +1,7 @@
 -r requirements.txt
 appnope==0.1.0; sys_platform == 'darwin'
 asn1crypto==0.24.0
+attrs==17.4.0
 autopep8==1.3.4
 bcrypt==3.1.4
 blessings==1.6.1
@@ -9,19 +10,21 @@ cffi==1.11.5
 chardet==3.0.4
 click==6.7
 codecov==2.0.15
+colorama==0.3.9; sys_platform == 'win32'
 coverage==4.5.1
 coveralls==1.3.0
 cryptography==2.1.4
 decorator==4.2.1
 django-debug-toolbar==1.9.1
-django-extensions==2.0.5
+django-extensions==2.0.6
 django-nose==1.4.5
 django==2.0.3
 docopt==0.6.2
 fabric3==1.14.post1
 factory-boy==2.10.0
 faker==0.8.12
 flake8==3.5.0
+funcsigs==1.0.2; python_version < '3.0'
 graypy==0.2.14
 idna==2.6
 ipdb==0.11
@@ -42,14 +45,17 @@ parso==0.1.1
 pbr==3.1.1
 pexpect==4.4.0; sys_platform != 'win32'
 pickleshare==0.7.4
+pluggy==0.6.0
 prompt-toolkit==1.0.15
 ptyprocess==0.5.2
+py==1.5.2
 pyasn1==0.4.2
 pycodestyle==2.3.1
 pycparser==2.18
 pyflakes==1.6.0
 pygments==2.2.0
 pynacl==1.2.1
+pytest==3.4.2
 python-dateutil==2.7.0
 pytz==2018.3
 pyyaml==3.12