Alphapulldown

What's Changed
* Fixed the issue with create_individual_features.py
* Allow the user to use custom template databases even while using mmseqs2 by specifying inputs for template_mmcif_dir and pdb70_database_path when running create_individual_features.py

**Full Changelog**: https://github.com/KosinskiLab/AlphaPulldown/compare/0.2-beta...0.22.1

What's Changed
* Now can generate MSAs using MMseqs2 either remotely or directly parse the results from local MMseqs2 jobs. In some cases, using MMseqs2 could be faster than HHBlits in the previous version of AlphaPulldown but please be aware that these two methods may give different sets of MSAs thus different models at the end. The user has to decide which one to use.

* Added two auxiliary scripts: prepare_seq_names.py and rename_colab_search_a3m.py which can help the user simplify the names of the sequences from Uniprot downloads and rename the a3m files from local MMseqs2 jobs.

**Full Changelog**: https://github.com/KosinskiLab/AlphaPulldown/compare/0.2-beta...0.22.1

What's Changed
* Remove the redundant and unnecessary loop at the end of create_individual_features.py. This can speed up the process compared to the previous versions of AlphaPulldown
* Reduce default pae plot figsize to avoid ipython kernel crashes

**Full Changelog**: https://github.com/KosinskiLab/AlphaPulldown/compare/0.2-beta...0.21.6

1. The Jupyter notebook will be created by get_good_pae.py script from _AlphaPulldown_ package itself. Once the notebook is created, the user can open it from Jupyterlab and view it. _jupyterlab_ is now automatically installed when installing _AlphaPulldown_.
2. Fixed errors that jupyter lab may crash on displaying PAE plots
3. Fixed the error with jax that might return a PDB file full of NaN.

_AlphaPulldown_ has entered its version 0.2 era. If you have installed an older version, make sure reinstall or upgrade it within your conda env.