Version 0.6.1

Fixed

• Missing region check should ignore subregion field, if it exists

Changed

• Use SPDX license data from SPDX themselves instead of spdx-license-list because it is deprecated and outdated
• Require Python 3.9+
• Don't run fix.separators() on title or abstract fields
• Don't run whitespace or newline fixes on abstract fields
• Ignore some common non-SPDX licenses
• Ignore __description suffix in filenames meant for SAFBuilder when checking for uncommon file extensions

Updated

• Python dependencies

Version 0.6.0

Changed

• Perform fix for "unnecessary" Unicode characters after we try to fix encoding issues with ftfy
• ftfy heuristics to use is_bad() instead of sequence_weirdness()
• ftfy fix_text() to not change “smart quotes” to "ASCII quotes"

Updated

• Python dependencies
• Metadata field exclude logic

Added

• Ability to drop invalid AGROVOC values with -d when checking AGROVOC values with -a <field.name>
• Ability to add missing UN M.49 regions when both country and region columns are present. Enable with -u (unsafe fixes) for now.

Removed

• Support for reading Excel files (both .xls and .xlsx) as it was completely untested

Version 0.5.0

Added

• Ability to check for, and fix, "mojibake" characters using ftfy
• Ability to check if the item's title exists in the citation
• Ability to check if an item has countries, but no matching regions (only suggests missing regions if there is a region field in the CSV)

Updated

• Python dependencies

Fixed

• Regular expression to match all citation fields (dc.identifier.citation as well as dcterms.bibliographicCitation) in experimental.correct_language()
• Regular expression to match dc.title and dcterms.title, but ignore dc.title.alternative check.duplicate_items()
• Missing field name in fix.newlines() output

Version 0.4.7

Changed

• Fixing invalid multi-value separators like | and ||| is no longer classified as "unsafe" as I have yet to see a case where this was intentional
• Not user visible, but now checks only print a warning to the screen instead of returning a value and re-writing the DataFrame, which should be faster and use less memory

Added

• Configurable directory for AGROVOC requests cache (to allow running the web version from Google App Engine where we can only write to /tmp)
• Ability to check for duplicate items in the data set (uses a combination of the title, type, and date issued to determine uniqueness)

Removed

• Checks for invalid and unnecessary multi-value separators because now I fix them whenever I see them, so there is no need to have checks for them

Updated

• Run poetry update to update project dependencies

Version 0.4.6

Added

• Validation of dcterms.license field against SPDX license identifiers

Changed

• Use DCTERMS fields where possible in data/test.csv

Updated

• Run poetry update to update project dependencies

Fixed

• Output for all fixes should be green, because it is good

Version 0.4.5

Added

• Check dates in dcterms.issued field as well, not just fields that have the word "date" in them

Updated

• Run poetry update to update project dependencies

Version 0.4.4

Added

• Accept dates formatted in ISO 8601 extended with combined date and time, for example: 2020-08-31T11:04:56Z
• Colorized output: red for errors, yellow for warnings and information, green for changes

Updated

• Run poetry update to update project dependencies

Version 0.4.3

Changed

• Reformat with black
• Now requires Python 3.7+ for pandas 1.2.0

Updated

• Run poetry update to update dependencies
• Expand check/fix for multi-value separators to include metadata with invalid separators at the end, for example Kenya||Tanzania||

Version 0.4.2

Changed

• Add field name to the output for more fixes and checks to help identify where the error is
• Minor optimizations to AGROVOC subject lookup
• Use Poetry instead of Pipenv

Updated

• Update python dependencies to latest versions

Version 0.4.1

Changed

• Reduce minimum Python version to 3.6 by working around the is_normalized() that only works in Python >= 3.8