Rustyms

* Fixed output of some Pro Forma features, notable charge and adduct ions
* Updated models to be more cohesive, added more neutral losses where they are expected
* Added `Modification::ontology_url` to get the url for modifications from ontologies
* Added `Modification::search` which searches for ontology modifications matching the modification given
* Updated ontologies and IMGT database

* Added python bindings (thanks to RalfG)
* Added `MultiChemical::charge`
* Added `MolecularFormula::from_pro_forma` (was private)
* Added `MolecularFormula::from_psi_mod` (was private)
* Made `MolecularCharge` public
* Updated IMGT database, notably improved for Human IGHA1&2

Very small patch, as `docs.rs` broke on `doc_cfg`. `doc_cfg` has been removed across the board.

First of all, this is a big update in terms of affected API. Many functions have been renamed/moved/updated. All functionality is still available, but any code that uses rustyms needs some manual updates.

Here is a shortlist of the most important things that have changed, if you want a full list check out [cargo-public-api](https://github.com/enselic/cargo-public-api):
* Alignment is more robust and allows for more settings, either global and different matrices are notable additions
* All germline sequences from IMGT are now included as well as methods of iterating over them
* Many more things now support Eq+Ord+Hash thanks to use of the ordered-float crate
* Many tuple results have been refactored to be small structs with named fields
* Amino acids now have different property classes defined
* The functions have been updated to handle ambiguous amino acids throughout so using B/Z now will never crash

* Added custom placement rules for building blocks and isobaric set generation
* Added support for terminal modifications in building blocks and isobaric set generation
* Added support for a custom AA alphabet for building blocks and isobaric set generation
* Added support for selecting a base sequence (possibly based on immonium ions) for building blocks and isobaric set generation
* Added `.assume_simple()` for `LinearPeptide`
* Added `AminoAcid::from_dna()`
* Added conversions into linear and complex peptides
* Added implementation for many more common traits as well as serde serialisation
* Added support for MaxQuant & MaxNovo files
* Added `.binary_search_slice()` on `RawSpectrum` (based on rusteomics)
* Refactored the CSV based file formats to use column names instead on indices for more robust identification file support
* Bug fixes (ppm for negative numbers works now, no crash on noise filter for empty spectrum)