Amplimap

- Support running without annotation or with different sets of Annovar indices
- Dockerfile improvements
- Handle variant calls outside target regions
- Add pipeline diagram
- Add test for variant merging

- fix minor code warnings
- fix dependencies (require UMI-tools <1.0)
- add Dockerfile and update installation instructions

- improve error message when no reads with matching primers were found
- fix error when `use_raw_reads` is true and aligner is BWA

- updated documentation, add list of supported protocols
- renamed `tagged_bams_in` to `mapped_bams_in`
- added check to make sure only one input directory exists
- added check to confirm that `use_raw_reads` is set when using `unmapped_bams_in` or `mapped_bams_in`
- add link to cluster log documentation when showing error message
- internal changes:
+ made input rules conditional on input directory existing, to remove some ambiguity when input files are missing
+ make sure `stats_alignment` files are created even with `mapped_bams_in`
+ added test for multiple input directory check

- improved code documentation
- improved support for running the Snakefile stand-alone rather than through amplimap
- fixed naive mapping of untrimmed reads
- updated tutorials
- removed some deprecated code
- added `--latency-wait` parameter
- added new test cases

- fix version check for newer versions of picard tools
- add `requirements.txt` file
- internal changes:
+ add tests for config file reading
+ add test for UMI dedup
+ add/improve code documentation
+ refactor config file reading
+ stop recording versions for individual files
+ catch missing cythonized modules