v1.0.1

Changed

Switch to headless opencv dependency to improve stability

Full Changelog: v1.0.0...v1.0.1

v1.0.0

What's Changed

Major new version that removes dependency on older machine learning approaches, including spaCy and NLTK, which tend to have version conflicts with Docassemble and perform poorly and at a higher cost than modern LLM technologies.

Changed

• Add information about presence of PDF tags for accessibility by @nonprofittechy in #104
• Force reportlab to use latest security patch by @BryceStevenWilley in #105
• Sorted simplified words to prep for new words by @plocket in #106
• Add simple words by @plocket in #107
• don't parse [] terms in YAML, allow overwrite of PDF by @nonprofittechy in #111
• If the PDF has no text, OCR it by @nonprofittechy in #112
• Add has_fields() function by @nonprofittechy in #115
• Typing issue by @nonprofittechy in #116
• Correct mypy issues by @BryceStevenWilley in #121
• Remove use of pickled (joblib) files for now by @nonprofittechy in #119
• Safer checking for field annotations in a PDF by @BryceStevenWilley in #122
• Fix issue with duplicate fields by @BryceStevenWilley in #124
• Only send first 5000 characters to Spot by @ClaireSimmonds in #126
• Migrate to the stable OpenAI Python client (1.0) by @nonprofittechy in #129
• Add some DOCX modification functions by @nonprofittechy in #130
• Fix mypy typing issues by @BryceStevenWilley in #136
• 25 detect sensitive fields by @codestronger in #134
• Unpin scikit-learn because it conflicts with docassemble by @nonprofittechy in #138
• quick and dirty patch by @nonprofittechy in #141
• Switch to ubuntu-latest action runner by @BryceStevenWilley in #143
• Migrate to more standard .env file usage by @nonprofittechy in #146
• We're not really using spaCy but it was still a dependency by @nonprofittechy in #148
• Replace passivepy with a call to an LLM by @nonprofittechy in #147
• Use raw strings for regexes to prevent warning in python 3.12 / black by @nonprofittechy in #151
• Typing and unit test fixes by @nonprofittechy in #152
• Allow working without .env; pull creds from docassemble config by @nonprofittechy in #150
• Finish the migration to LLMs; removing NLTK, etc. by @nonprofittechy in #153
• Migrate from spa cy and nltk by @nonprofittechy in #154

New Contributors

• @plocket made their first contribution in #106
• @ClaireSimmonds made their first contribution in #126
• @codestronger made their first contribution in #134

Full Changelog: v0.2.0...v1.0.0

v0.3.0a3

What's Changed

This release now depends on access to the tools.suffolklitlab.org endpoint for several features. In the future, we will likely migrate these to a public LLM endpoint, like OpenAI, which can be used more cost effectively for most deployments

• Remove all requirements to install Spacy by @colarusso in #141

Full Changelog: v0.3.0a2...v0.3.0a3

v0.3.0a2

What's Changed

• Fix issue with duplicate fields by @BryceStevenWilley in #124
• Only send first 5000 characters to Spot by @ClaireSimmonds in #126
• Migrate to the stable OpenAI Python client (1.0) by @nonprofittechy in #129
• Add some DOCX modification functions by @nonprofittechy in #130
• Fix mypy typing issues by @BryceStevenWilley in #136
• 25 detect sensitive fields by @codestronger in #134
• Unpin scikit-learn because it conflicts with docassemble by @nonprofittechy in #138

New Contributors

• @ClaireSimmonds made their first contribution in #126
• @codestronger made their first contribution in #134

Full Changelog: v0.3.0a1...v0.3.0a2

v0.3.0a1

What's Changed

New

• Add information about presence of PDF tags for accessibility by @nonprofittechy in #104
• Force reportlab to use latest security patch by @BryceStevenWilley in #105
• Add simple words by @plocket in #107
• If the PDF has no text, OCR it by @nonprofittechy in #112
• Add has_fields() function by @nonprofittechy in #115

Fixed

• don't parse [] terms in YAML, allow overwrite of PDF by @nonprofittechy in #111
• Typing issue by @nonprofittechy in #116
• To handle new release of Pandas, remove use of pickled (joblib) files for now by @nonprofittechy in #119
• Safer checking for field annotations in a PDF by @BryceStevenWilley in #122

New Contributors

• @plocket made their first contribution in #106

Full Changelog: v0.2.0...v0.3.0a1

v0.2.0

Added

• Add suggestions to remove complex and gendered terms from text of forms by @nonprofittechy in #101
• Improve complexity score coverage by @nonprofittechy in #102

Changed

• Work with both tools.suffolklitlab.org and locally by @BryceStevenWilley in #96

Fixed

• If no sentences, just return by @BryceStevenWilley in #95

Full Changelog: v0.1.1...v0.2.0

v0.1.1

What's Changed

https://github.com/SuffolkLITLab/FormFyxer/blob/main/CHANGELOG.md#version-v011

v0.0.10.1

Fix types

v0.0.10

What's Changed

• small fixes and cleanups in lit_explorer by @BryceStevenWilley in #45
• Better brackets by @BryceStevenWilley in #57
• Improve field finding in PDFs: by @BryceStevenWilley in #55
• Use PikePDF and pdfminer.six by @nonprofittechy in #64
• Reformat with Black, start adding types by @nonprofittechy in #69
• Fix types by @BryceStevenWilley in #74
• Add sentence count, passive voice, citations and first alpha of "time to answer" score by @nonprofittechy in #70

Full Changelog: v0.0.9...v0.0.10

v0.0.9

A new click version broke typer which broke spacy, which then broke the Weaver. New package requirements enforces the right version of typer, which works with the latest click.