Vermouth

New Features
* add zenodo authors file by fgrunewald in https://github.com/marrink-lab/vermouth-martinize/pull/431
* resids from the input structure can now be preserved using the -resid flag; this fixes 421 153 & 366
* elastic networks can be fine-tuned via the eunit flag which now allows selection of resid ranges
* default mapping force-field universal is removed; mapping now happens by default from CHARMM; users can select AMBER as well. The only difference to universal is that titratable AAs are picked up correctly.

Bug Fixes
* Issue 424 & 439 regarding a bug, which prevented removal of atoms is now fixed.
* Correct the version check in Molecule.remove_interaction by 484

Library Updates
* Issues 425 426 426 regarding issues with legacy Martini2 data-files were fixed
* Add charmm data for NCAP and CCAP mappings by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/477
* Proline termini mappings by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/476
* Update aminoacids.ff by fgrunewald in https://github.com/marrink-lab/vermouth-martinize/pull/473
* Protonation states are now implemented by resname for AMBER and CHARMM
* HSD modification was fixed for AMBER
* implement small molecule DB by fgrunewald in https://github.com/marrink-lab/vermouth-martinize/pull/456
* Add citation of M3 small molecules (address issue 468)
API Updates
* Fix write_pdb to always define open by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/434
* Fix/432 by making sure modifications are connected components by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/436
* Update gh actions by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/475
*
Dependency Updates
* Remove redistributed KDTree, and make scipy a hard requirement by pckroon in https://github.com/marrink-lab/vermouth-martinize/pull/438

This release implements some performance enhancements, small bug fixes and updated API.

Performance Improvements:
- accelerated making of residue graphs related to issue 379
- accelerated parsing of itp files related to issue 377
- accelerated mapping of molecules related to issue 408

Bug Fixes:
- itp reader now reads virtual_sites1 and settle directives
- itp reader can handle else statements

API Updates:
- the deferred file writer doesn’t overwrite the built-in open anymore instead now we expose deffered_open
- the PDB writer gets a new flag ‘defer_writing’ that either uses the deferred file writing or just regular file writing

Dependency Updates:
- Martinize2 now supports the latest version of Sphinx
- Martinize2 now runs with the latest version of pylint

This bugfix release fixes two issues:
- Warnings issued by MakeBonds now have a `type`.
- The N-ter modification in the 'universal' force field was modified to also fit on N-terminal prolines. This also affects the mapping weights of N termini.

Alanine was missing a bond between BB and SC1 in the force field martini3001 if the FLEXIBLE flag was used. This release fixes that.

With the publication of the Martini 3 paper we can include the relevant data files, and switch the default output force field to Martini 3!

You can find the paper here: https://www.nature.com/articles/s41592-021-01098-3

This release has a few important usability features:
- Support for citations from the data files, which make martinize2 print the relevant citations.
- Make the -nter and -cter CLI flags only affect protein residues. This does mean that non protein molecules don't get their termini patched automatically.

In addition, there are a number of bug fixes:
- Fixed a bug in the ISMAGS symmetry detection.
- Fix 311.
- Change CI to Github actions.
- Fix 317, which sometimes triggered DSSP errors when one of the terminal oxygens is missing.
- Fix issues where resids from the PDB file were ignored when itp files are used as input forcefield.
- Fix 354 which would trigger if there were multiple unknown modifications in a residue.