aboutsummaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/importers
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2020-07-23 15:31:03 -0700
committerBryan Newbold <bnewbold@robocracy.org>2020-07-23 15:31:05 -0700
commit578ce63bd17df11d3ca7cd1083feebe4b92e457c (patch)
tree306d218688171edd41a768daf65926cd20df5102 /python/fatcat_tools/importers
parent1b28ba68e3b25a0cded2e88b88134d8d332f8d96 (diff)
downloadfatcat-578ce63bd17df11d3ca7cd1083feebe4b92e457c.tar.gz
fatcat-578ce63bd17df11d3ca7cd1083feebe4b92e457c.zip
simple lint (flake8) fixes over python codebase
These should not have any behavior changes, though a number of exception catches are now more general, and there may be long-tail exceptions getting thrown in these statements.
Diffstat (limited to 'python/fatcat_tools/importers')
-rw-r--r--python/fatcat_tools/importers/common.py1
-rw-r--r--python/fatcat_tools/importers/crossref.py14
-rw-r--r--python/fatcat_tools/importers/datacite.py14
-rw-r--r--python/fatcat_tools/importers/grobid_metadata.py2
-rw-r--r--python/fatcat_tools/importers/jalc.py2
5 files changed, 16 insertions, 17 deletions
diff --git a/python/fatcat_tools/importers/common.py b/python/fatcat_tools/importers/common.py
index c0578224..c692a38d 100644
--- a/python/fatcat_tools/importers/common.py
+++ b/python/fatcat_tools/importers/common.py
@@ -692,7 +692,6 @@ class Bs4XmlLargeFilePusher(RecordPusher):
def run(self):
elem_iter = ET.iterparse(self.xml_file, ["start", "end"])
- i = 0
root = None
for (event, element) in elem_iter:
if not root and event == "start":
diff --git a/python/fatcat_tools/importers/crossref.py b/python/fatcat_tools/importers/crossref.py
index 854e3d9f..71f08952 100644
--- a/python/fatcat_tools/importers/crossref.py
+++ b/python/fatcat_tools/importers/crossref.py
@@ -278,15 +278,15 @@ class CrossrefImporter(EntityImporter):
# license slug
license_slug = None
license_extra = []
- for l in obj.get('license', []):
- if l['content-version'] not in ('vor', 'unspecified'):
+ for lic in obj.get('license', []):
+ if lic['content-version'] not in ('vor', 'unspecified'):
continue
- slug = lookup_license_slug(l['URL'])
+ slug = lookup_license_slug(lic['URL'])
if slug:
license_slug = slug
- if 'start' in l:
- l['start'] = l['start']['date-time']
- license_extra.append(l)
+ if 'start' in lic:
+ lic['start'] = lic['start']['date-time']
+ license_extra.append(lic)
# references
refs = []
@@ -297,7 +297,7 @@ class CrossrefImporter(EntityImporter):
# NOTE: are there crossref works with year < 100?
if year > 2025 or year < 100:
year = None
- except:
+ except (TypeError, ValueError):
year = None
ref_extra = dict()
key = rm.get('key')
diff --git a/python/fatcat_tools/importers/datacite.py b/python/fatcat_tools/importers/datacite.py
index ebb29feb..50d73798 100644
--- a/python/fatcat_tools/importers/datacite.py
+++ b/python/fatcat_tools/importers/datacite.py
@@ -466,7 +466,7 @@ class DataciteImporter(EntityImporter):
try:
_ = int(first_page) < int(last_page)
pages = '{}-{}'.format(first_page, last_page)
- except ValueError as err:
+ except ValueError as err: # noqa: F841
# TODO(martin): This is more debug than info.
# print('[{}] {}'.format(doi, err), file=sys.stderr)
pass
@@ -478,11 +478,11 @@ class DataciteImporter(EntityImporter):
license_slug = None
license_extra = []
- for l in attributes.get('rightsList', []):
- slug = lookup_license_slug(l.get('rightsUri'))
+ for lic in attributes.get('rightsList', []):
+ slug = lookup_license_slug(lic.get('rightsUri'))
if slug:
license_slug = slug
- license_extra.append(l)
+ license_extra.append(lic)
# Release type. Try to determine the release type from a variety of
# types supplied in datacite. The "attributes.types.resourceType" is
@@ -524,7 +524,7 @@ class DataciteImporter(EntityImporter):
value = attributes.get('language', '') or ''
try:
language = pycountry.languages.lookup(value).alpha_2
- except (LookupError, AttributeError) as err:
+ except (LookupError, AttributeError) as err: # noqa: F841
pass
# TODO(martin): Print this on debug level, only.
# print('[{}] language lookup miss for {}: {}'.format(doi, value, err), file=sys.stderr)
@@ -549,7 +549,7 @@ class DataciteImporter(EntityImporter):
if isinstance(text, list):
try:
text = "\n".join(text)
- except TypeError as err:
+ except TypeError:
continue # Bail out, if it is not a list of strings.
# Limit length.
@@ -760,7 +760,7 @@ class DataciteImporter(EntityImporter):
i = 0
for c in creators:
if not set_index:
- i = None
+ i = None
nameType = c.get('nameType', '') or ''
if nameType in ('', 'Personal'):
creator_id = None
diff --git a/python/fatcat_tools/importers/grobid_metadata.py b/python/fatcat_tools/importers/grobid_metadata.py
index 5ec6cc3c..a811c856 100644
--- a/python/fatcat_tools/importers/grobid_metadata.py
+++ b/python/fatcat_tools/importers/grobid_metadata.py
@@ -104,7 +104,7 @@ class GrobidMetadataImporter(EntityImporter):
if raw.get('date'):
try:
year = int(raw['date'].strip()[:4])
- except:
+ except (IndexError, ValueError):
pass
for key in ('volume', 'url', 'issue', 'publisher'):
if raw.get(key):
diff --git a/python/fatcat_tools/importers/jalc.py b/python/fatcat_tools/importers/jalc.py
index 38aa00eb..9bf2621c 100644
--- a/python/fatcat_tools/importers/jalc.py
+++ b/python/fatcat_tools/importers/jalc.py
@@ -308,7 +308,7 @@ class JalcImporter(EntityImporter):
work_id=None,
title=title,
original_title=clean(original_title),
- release_type="article-journal",
+ release_type=release_type,
release_stage='published',
release_date=release_date,
release_year=release_year,