Protgraph

Finished the Implementation of N-Term and C-Term variable and fixed Modifications. Now ProtGraph can encode all mods, but only 1 (fixed/variable) modification per aminoacid (including N-Term and C-Term).

Release after v0.3.2, which had a BUG in statistics generation.

A small and quick release fixing some bugs.

* Added new Digestion enzyme Glu-C
* Parsing EMBL-Files are now done in the Graph-Generation Process (yields to a slight performance increase)
* Collapsing Edges are now collapsing information in a compact manner and
* BUG-Fix in digestion and Feature-Generation, where we sometimes "forgot" to include feature

A release containing various small fixes (for aspw and exporter) as well as a new functionality: `protgraph_compact_fasta`

Finally! A release of ProtGraph, fixing some problems, having slight improvements and reducing the number of essential dependencies by introducing optional dependencies (mostly for exports).


**Changes**
* Refactoring of many functions, reducing duplicated code. (ProtGraph also seems more responsive, due to removing the delay in TQDM)
* Added more customization for the generated statistics file, you can now select which columns you are interested in!
* This refactoring also led to add multiple execution scripts into protgraph, there are now more cli-commands which can be used for automation
* New easy scripts to generate decoys as well as to replace (extract) header information from fasta files! There is now also a script available which allows to easily sum up columns in generated statistic files!
* Also added support to parse ProPeptide and Peptide FeatureTables (FT) from the SP-EMBL-Format
* Added support for simple PTMs in graphs by adding an additional Attribute `delta_mass`. Currently only the mass shift is supported on aminoacids and NOT on the terminis... (this most probably will be added in the next release!)
* New Exporter: Graphs can be now exported into a **C**ompressed **S**parse **R**ow-Representation (Protein-CSR --> PCSR, Binary-PCSR -> BPCSR)
* New Trie Exporter abusing the filesystem. A slower peptide exporter with an interesting concept!
* New (fast) exporter from graphs into peptides! Using SQLite (with an additionally added script) allows a quick generation of a Fasta-file even for complete UniProt!
* Added new option in peptide traversal, you can now filter by weight instead of only peptide length! (or filter by both)

The changlist got a bit lengthy, so was this release, which took almost 5 months!

Changes and Improvements can be seen on PR 44

Finally, after 3 months of work and pizza, a new release: ProtGraph 0.2.1

This version does not have any breaking changes to the previous one and mostly adds new functionalities.

**Improvements/Changes**:
- The summarizing (merging) of single amino acids in nodes into one Node is significantly faster (Factor ~20x). E.G.: Executing ProtGraph on Titin will now only need roughly ~2 minutes (instead of ~30 minutes)!
- Psycopg2 has been replaced by Psycopg3, therefore an overall performance increase using Postgres should be seen. We also set up connections to use the binary protocol for better efficiency!
- New stopping criterion for started processes in ProtGraph. We noticed, that some processes do not get destroyed after they finished. This is most probably due to some multiprocessing shenanigans in Python itself. We changed the implementation slightly in hope that zombie processes occur less often!

**New Features**:
- New CLI Option, Add no description to the statistics output csv. This reduces the statistics file size considerably (for complete UP, a reduction of ~20 GB could be noticed)
- ProtGraph now collapses parallel edges into one edge, containing information of all parallel edges. This may only happen for some proteins, however, ProtGraph now has an "edge case" for that.
- A new exporter has been added: Cassandra (Cluster)
- We now are able to calculate the number of possible peptides by variants/mutagen and conflicts, giving more insights of the large search space!

The new release of ProtGraph should shortly be availble in PIP (PyPI) and probably tomorrow on BioConda due to new dependencies!