v4.0.0
difPy v4 comes with major updates, code improvements and new features 🎉
- Enhancement: difPy now leverages Python's multiprocessing capabilities which results in significant performance increases. In tests, difPy v4 was on average 10x faster than it's previous versions on the same datasets. Suggested by @TheLastGimbus and @thecodingchicken.
- Enhancement: difPy's processes have been split into
difPy.build
anddifPy.search
so that multiple searches can be performed on the same image repository and not having to re-build it. - Enhancement: difPy is now more lightweight thanks to reduced dependency on external packages and increased usage of Numpy capabilities.
- New feature: difPy now supports
in_folder
search, allowing to search separately among directories, instead of in the union of all directories. Suggested by @audiomuze in #53. - New feature: when running from the CLI, if no folder is specified, difPy will run by default in the working directory.
- Various other minor code improvements
The usage of difPy v4 varies significantly to previous versions. It is therefore recommended to see the updated difPy Usage Documentation.