okkevaneck / prospr Goto Github PK

Visualization can be made more pretty and include extra information. Don't forget to update documentation figures as well!

• Note: Y-axis not rendered properly when placing all amino acids in a row.
• Fix visualization compatibility with new version of prospr_core

Implement manylinux and auditwheel.

Somehow the Pybind11 compilation fails on the Windows runners.

Just putting this here ;)

Example

Change the way datasets are provided. Rename folder 'datasets' to 'data', use package_data in setup.py, and find good file structure.

Make a decision on how to speed up the CI. Non-master pushes do not need all wheel builds, maybe only do a ubuntu-32bit build, or a 64-bit build on all platforms.

Fix 3D plots of proteins.

Change auto-zipping to support multiple archive types. Create a separate folder for the zips as well, which is referenced in the README. Also change the pre-commit file to be in a .githooks folder instead of the github folder.

Happens when running the Pytests on a macOS runner. Ubuntu runner with the same code works flawlessly.

Caching the Python environment will speed up the CI as the dependencies are not reinstalled for each step.

Check on:

• Imports
• Calling attributes instead of getters
• New functions from Protein

The build version of the code needs to be tested. In order to do so, a tool like pytest needs to be used to perform tests. The tests need to be added to the CI workflow.

Add documentation on the usage of the Protein module and how algorithms can be used.

Check differences with Protein module.
Add the notion of AminoAcids.

See if the Furo theme suits the documentation.

For GitHub actions:

• Set python-version in setup_python for build_source_distribution, build_wheels, and deploy
• Set right version for pypa/gh-action-pypi-publish@master in deploy

For fixing imports:

• Add AminoAcid, load_vanEck250 and load_vanEck1000 to __all__ list.
• Add load_vanEck250 and load_vanEck1000 to import from prospr.data

Only some parts of the core are tested with PyTest. Extend the test suite to also test all other functionalities.

• Algorithms don't need to return the pointer, folding algorithms can have rettype void.
• For deterministic algorithms, build load_tmp functions for loading intermediate results from a hash_fold()

From the README.md:

The PDF version of the complete documentation can be found here.

Cibuildwheels gives the error that a [project] section inside pyproject.toml file is invalid for Python 11. See what is the issue and fix it to support Python11.

Change core into:

• Change to doubly linked list for storing conformations
• Remove required move for removal of amino acids
• Add possibility to remove / place multiple amino acids in a given direction (add to docs)
• Add possibility to rotate the chain at a given point (add to docs)
• Add Amino Acid class for storing the values in the values of space
• Add copyright to all headers with https link to the project's LICENSE file

The core code is available as from prospr import prospr_core instead of directly as from prosper import ....
The visualize code needs to be tested on importability, as well as the helper code.

As a side task, the visualize code also needs to be tested on functionality.