1
0
mirror of https://github.com/ilri/csv-metadata-quality.git synced 2024-11-04 21:43:00 +01:00
csv-metadata-quality/csv_metadata_quality/util.py
Alan Orth 7cc49b500d
Use licenses.json from SPDX instead of spdx-license-list
spdx-license-list has been deprecated[1] and already has outdated
information compared to recent SPDX data releases. Now I use the
JSON license data directly from SPDX[2] (currently version 3.19).

The JSON file is loaded from the package's data directory using
Python 3's stdlib functions from importlib[3], though we now need
Python 3.9 as a minimum for importlib.resources.files[4].

Also note that the data directory is not properly packaged via
setuptools, so this only works for local installs, and not via
versions published to pypi, for example (I'm currently not doing
this anyways). If I want to publish this in the future I will
need to modify setup.py/pyproject.toml to include the data files.

[1] https://gitlab.com/uniqx/spdx-license-list
[2] https://github.com/spdx/license-list-data/blob/main/json/licenses.json
[3] https://copdips.com/2022/09/adding-data-files-to-python-package-with-setup-py.html
[4] https://docs.python.org/3/library/importlib.resources.html#importlib.resources.files
2022-12-13 10:39:17 +03:00

66 lines
1.9 KiB
Python

# SPDX-License-Identifier: GPL-3.0-only
import json
from importlib.resources import files
from ftfy.badness import is_bad
def is_nfc(field):
"""Utility function to check whether a string is using normalized Unicode.
Python's built-in unicodedata library has the is_normalized() function, but
it was only introduced in Python 3.8. By using a simple utility function we
are able to run on Python >= 3.6 again.
See: https://docs.python.org/3/library/unicodedata.html
Return boolean.
"""
from unicodedata import normalize
return field == normalize("NFC", field)
def is_mojibake(field):
"""Determines whether a string contains mojibake.
We commonly deal with CSV files that were *encoded* in UTF-8, but decoded
as something else like CP-1252 (Windows Latin). This manifests in the form
of "mojibake", for example:
- CIAT Publicaçao
- CIAT Publicación
This uses the excellent "fixes text for you" (ftfy) library to determine
whether a string contains characters that have been encoded in one encoding
and decoded in another.
Inspired by this code snippet from Martijn Pieters on StackOverflow:
https://stackoverflow.com/questions/29071995/identify-garbage-unicode-string-using-python
Return boolean.
"""
if not is_bad(field):
# Nothing weird, should be okay
return False
try:
field.encode("sloppy-windows-1252")
except UnicodeEncodeError:
# Not CP-1252 encodable, probably fine
return False
else:
# Encodable as CP-1252, Mojibake alert level high
return True
def load_spdx_licenses():
"""Returns a Python list of SPDX short license identifiers."""
with open(files("csv_metadata_quality").joinpath("data/licenses.json")) as f:
licenses = json.load(f)
# List comprehension to extract the license ID for each license
return [license["licenseId"] for license in licenses["licenses"]]