Feature/cells argument

.gitignore

@@ -3,4 +3,6 @@ build/
 CITE_seq_Count.egg-info/
 __pycache__
 *.pyc
-.vscode
+.vscode
+.ruff_cache
+.pytest_cache

CHANGELOG.md

@@ -4,10 +4,41 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](http://keepachangelog.com/)
 and this project adheres to [Semantic Versioning](http://semver.org/).
 
-## [1.4.4] - 10.12.2020
+## [1.5.0] - XXXX
+### Added
+  - `CITE-seq-Count` is now Compatible with trimmed data. There is a new `too_short` category in the `run_report.yaml`
+    that will let you know how much you lost due to reads being too short. #123
+  - UMI correction is now also parallelized and will use the threads given.
+  - Added a check at the end of the mapping. If more than 99% of the reads are unmapped, CITE-seq-Count will exit. #62
+  - (BETA) New functionnality that will fetch the chemistry definition from a remote repo to simplify usage and reduce human errors.
+  - Added cython dependency based on issue #117
+
 ### Changed
-  - Hotfix linked to the issues like #125 showing that python3.8 and above crash on modified loop.
-  - Fixed bug that checks filenames length instead of number of files.
+  - The `features.tsv` now has different columns for the tag name and the tag sequence. This keeps the relevant information
+    in the output files as well as simplifies reading the mtx format when processing the data.
+  - The mapping step has been changed. It will first write chunks of reads to files and then read in the chunks in each child process.
+    This should solve the io bottleneck from before.
+  - There are new options now for parallel computing. `--chunk_size` Determines how many reads will be read per chunk. This should fix issues like #99.
+  - `--sliding-window` now only checks for exact matches.
+  - The main results dict now uses an `int` as keys reducing memory footprint.
+  - Fixed the issue #92 with using `--bc_collapsing_dist 0`.
+  - Fixed issue #122 and now properly checks number of files.
+  - Fixed the error in the documentation pointed by issue #132.
+  - The report is now a proper yaml file. Issue #133
+  - Removed distance checking on the whitelist because too slow for long whitelists.
+  - Tags csv file now requires a header with at least "sequence" and "feature_name".
+  - Updated tags file parsing to make it more reliable.
+  - Added new tests to help out contributions.
+  - If no clustered cells found, the dense output matrix will not be written.
+  - Barcode whitelists are now called reference lists.
+  - The reference list file now requires a header `reference`. There is now an optional column called `translation`. This is specific to chemistries such as 10xV3 that use different barcodes for mRNA and Antibody tag capture sequences. See more details in the documentation. #139 and #141
+  - Bumped UMI_tools to 1.1.1
+  - Changed `-cells` paramter to `-n_cells` for more explicit argument.
+  - Cell barcode correction when reference list is provided is now discarding cells with only unmapped reads reducing run time.
+  - Aberrant cells wording replaced by clustered cells to be more specific.
+
+  ### Removed
+  - Unmmapped reads are not umi corrected anymore reducing run time and memory usage.
 
 
 ## [1.4.3] - 05.10.2019