Atomic number of gold

1.3.0 - 2022-01-25

Changes

• Updates Pillow to >=9.0.0 in order to fix a critical CVE (SECURITY) by @leblancfg in #157
• Remove Travis CI Integration by @leblancfg in #152
• Remove Appveyor integration by @leblancfg in #151
• Set up GH Actions by @leblancfg in #150
• Update the experimental CI Python 3.11 alpha to rc 4 by @leblancfg in #158
• The initial-setup step in Makefile now also installs development packages.
• black, wheel and twine packages now get installed with requirements-dev.txt

Deprecations

• Deprecates Python 3.6

Based on careful, scientific investigation

This release now uses the headless version of OpenCV. This should contribute to a smaller download, as well as make autocrop an easier library to package in installers.

What's Changed

• 🎥 Add example for video frames by @leblancfg in #136
• Standardize docstring format by @leblancfg in #137
• Create codeql-analysis.yml by @leblancfg in #138
• Update Pillow dependency by @leblancfg in #139
• Remove LGTM tag from README by @leblancfg in #141
• Use headless OpenCV by @leblancfg in #144
• Autocrop version 1.2.0 by @leblancfg in #145
• Add clearing step to pypi in the makefile by @leblancfg in #146

Full Changelog: https://github.com/leblancfg/autocrop/CHANGELOG.md

An excellent trade

Deprecations

• Deprecate Python 3.5
• Deprecate OpenCV 3

API Additions

• User can now specify what file extension to save cropped files at the CLI

Security

• Update Pillow dependency in order to limit possible security issues

Other changes

• Updates to the developer setup tools and documentation
• Updates to the example notebook

Plays well with kids

This release sees quite a bit of change. We introduce a Python class that's usable elsewhere.

Also, the CLI tool conserves the original files.

alpha

Changes

• Autocrop now prints the filename of images where face detection failed
• Internal refactoring and more tests

alpha

Changes to 0.3.1

• Add -r, --reject flag to specify directory where the images that autocrop couldn't find a face in are directed to.
• Instead of having the target files copied then cropped, they are instead cropped and saved to their respective target folder.

alpha

This release

Fixes:

• The long-standing image-stretching when using non-square -w and -H values

Adds features:

• facePercent CLI option, to select how much around the face autocrop leaves in the image
• padding CLI attributes, which lets the user position the face at any position in the image

Big thanks to @SirNarsh for the contribution

alpha

This release fixes a major bug when handling images with no detectable faces.

alpha

Merge pull request #15 from leblancfg/numpy

tightened numpy requirements, bumped to v0.1.4

alpha

First production-ready verison