Release v3.1.0

Release v3.0.6

Small release with

 # Adds:
        - Script that can transform primer bed to bedpe based on
          amplicon names

 # Fixes:
        - pangolin update script: might need to be run with
          `--pre-release` option
        - Rmarkdown report crashing if species filtering removes all
          reads. Only hotfix: fragment size distribution plot crashes,
          because it has no mapped reads and information.

Patch Release v3.0.5

 # Adds:

    - mamba support
    - snakemake 6 support
    - automated docker builds

 # Fixes:

    - subprocess calls using sh instead of bash in update_pangolin script
    - all lineage information is now collected in one file

 # Changes:

    - wrapper log file is now next to all the other logs

v3.0.4

Adds:
  - Option: read_filter_len (default is 50)
    Allows to adjust read filter. Some SarsCov Kits use 75BP Reads and
    don't work well with the stringent filter.
  - Option: read_filter_qual (default is 20)
    Used to be 15 and was adjusted to 20. Please use this option with
    15, to recreate old data.
  - warning if new version is available

Changes:
  - requires Snakemake < 6.1, because they require mamba after 6.1

v3.0.3

 # Fixes

    - Version not found when installing with pip

 # Changes:

	- Kraken Parameter name mistake in documentation.
      Closes #85
    - Zenodo Link typo was fixed
    - Docker file updates for version v.3.0.1.
      Sadly automatic builds do not yet work, due to size
      constraints.
	- SAP (Strand Bias Filter) is now optional and disabled by default.
	  This adresses the issue tha NimaGen amplicon sequencing kits failed
	  to detect some variants with the strict filters.
      Closes #84 #96

New Release

 # Changes

  * finished update_pangolin utility
    It is now able to update or install a pangolin environment
    to either the latest release or latest master branch commit.
  * being more verbose in quicktest error messages
  * adjusted quicktest to define realpath, if it is not
    available on a system

 # Fixes

  * Nonetype error when triggering the illumina fallback filename
    parser.
  * Truncated sam file issue in liftoff.
    This was due to it not managing its external resources
    in a parallel friendly manner.
  * feature_db referenced before assignment error in liftoff.
    same reason as above.

New Release v3.0.0

 # Changes

    * resource management for rules for the better.
    * due to major snakemake code refactor.
      (Code cleanup mainly)
    * use NC_045512.2 Wuhan Sars-Cov-2 reference and its corresponding
      annotation gff,
      if no reference is provided.
    * Primer clipping. It  is now performed by bamclipper and
      needs a bed file with primer locations.
    * due to major refactor and move to install the pipeline with setuptools.
    * config option "annotation" to var_annotation
    * --no-annotation wrapper option is now --no-var-annotation

 # Adds

    * Pangolin lineage assignment support.
    * "--pangolin" option to wrapper.
    * annotation Transfer with Liftoff to reconstructed genomes
    * create_bedpe script
      In case you only have amplicons and their insert positions (as bed files),
      these can be translated into a single .bedpe file ready to use
      with the --primer option.
    * quality control mechanism for detecting low coverage in genomic
      regions that are important for detecting variants of concern.

      This also adds a "--variants_of_interest"
      option that accepts a vcf.
    * lineage information to report
    * annotation_gff option to wrapper
      that triggers annotation of reconstructed genomes
    * config option "cns_annotation"

 # Removes

    * "amplification" config option (Now based on the presence of --primer option)
    * config option "max_primer_mismatches" as it is also no longer used
    * explicit "--wgs" wrapper option and related logic because it is now
      implicitly detected. (see above)

 # Fixes

     * missing line endings in reconstructed genome fastas!
     * Some fallback regexes had issues in the filename parser
     * png plot device no longer crashes on large plots for lots of
       samples

- reduced memory footprint of report
- reduced run time of report
- output additional tables
- envs reorder channels
- clean up ptrimmer env
- check for config file presence in Snakefile

Changes:

    * global: require snakemake version greater then 5.26
	* install_script: Order of channels in the setup script (environment solve speed is vastly improved)
    * install_script: removed not needed dependency

Adds:

	* quicktest test: re-introduce quicktest that checks if pipeline can be rerun if only a config is provided
	* conda: .conda folder that contains recipes corresponding to releases

Fixes:

   * fix for logging error if no output directory is provided
   * fix for snakemake preperation run failing if .snakemake migration takes place

Changes:
	- install can deal with missing .git folder closes #57
    - no longer checks for parameter changes on first run
      (should fix arguments to long error on large runs)
	- quicktests have more verbose output

Adds:
	- new regex for rki file schema
    - Debug output for snakemake rerun files

Fixes:
	- third illumina regex

- fixed amplification primer clipping to avoid discarding reads overly strict

Pipeline received major changes: 
- variant caller was swapped from GATK to Freebayes 
- amplification primer clipping is run by ptrimmer (was cutadapt before) 
- amount of output sequences was reduced to low coverage masked sequences either containing IUPAC code nucletoides for mixed variants or these positions N masked 
- code is put into modules for ease of maintenance 
- report includes some enhancements for improved ovreview for large data sets

Major fix of wrapper and some conda environment fixes

Changes:
   + Fixed issue with strict yaml that resulted in boolean config values changing to "yes" and "no".
     This broke the annotation setting when using the wrapper.
   + Added "mscorefonts" to r environment for report writing, which causes the minimal gitlab ci docker image
     to not output broken fonts in plots
   + Generally fixed tool versions in the report writing r environment. Improving reproducibility

Changes:

    - cutadapt setup has changed drastically.
      To-be-cut, sequences for both 3' and 5' ends of amplicons can now
      be given in separate files.
      The config options changed and this will cause incompatibility with piplines that were run
      prior to 1.2.0.

      New Option:
         --primerfasta_3prime is the new otion for the 3' primers/sequences to be cut

Added:
    - multisample variant calling

Fixes:
    - Config parameters were not correctly parsed in the Snakefile.
      This is also the reason why annotation was never performed
    - The --blame option is vastly improved

New Release v1.1.4

Changed Behaviour:
  - No longer looks for oldest config in output folder.
    The user needs to request this explicitly via the --use_last_config option
  - Snakemake is now called with the -d option.
    This causes the .snakemake directory to be stored in the output folder and not
    the current working directory.

    Side Effect: Conda environments, if not told via the --conda_prefix <Folder> to be
                 stored somewhere else, are also stored in the outputfolder/.snakemake/conda
                 directory. We strongly recommend the use of a common --conda_prefix
                 if you plan on using this tool alot.

Added:
   - Sample Validation: This task is no longer relegated to the snakemake code
         It also checks if files are proper two file paired end reads
   - Capability to assign unique alternative  ids to incrementally added samples
   - python 2.7 compatibility
   - New Parser Options:
       - Option: --ignore_old_samples
             Forces program to ignore samples listed in config file and only uses
             the ones provided via the command line
       - Option: --wgs
             Deactivates amplicon mode even if primer file was provided
       - Option:
Fixes:
   - SRA file parse error because missed  adding "1" and "2" as read ids
   - Filename parse allways fell through to the most generic matching pattern

- adding filter step that masks all raw variant calls, if they don't survive filtering steps

Bugfixes:
    - When in tag release __version__.py does no longer complain about
      missing head count and commit

Changed:
   - in-house renaming scheme for samples
Added
  - Cores Option now mandatory in snakemake 5.11.0 +