MultiQC plugins for TIN scores and ALFA

Substitute injected .png's for ALFA and TIN scores with our custom MultiQC plugins (repo).

Notable points:

• Adjusted MultiQC rule to use our new container.
• Removed 2/3 rules on both sides (ALFA, TIN): we do not need any merging, plotting, image concatenation etc. what is left is just the calculation of respective scores.
• Pipeline documentation updated accordingly to (1)
• Quasi-unit/quasi-integration test for ALFA turned off in the CI. I see that this test was initially calling Snakemake to run 3 ALFA-related rules and generate the final ALFA plots. Since now for ALFA we have only one rule and the main output are per-sample TSV table a "separate 1-rule Snakemake call" is not necessary. Proper unit testing of this functionality should be a bash ALFA call but on a separate input, since everything else (evaluation of wildcards, jobs parallelization) is taken care of in the integration testing.

I will attach here a report after GCN4_project analysis, so that you may take a look at plugins in action: multiqc_summary.zip

---

@katsanto @gypas @kanitz

workflow/scripts/zarp_multiqc_config.py

@@ -89,7 +89,7 @@ report_header_info:
   - Analysis Author: "{author_name}"
   - Contact E-mail: "{author_email}"
 
-top_modules:
+module_order:
 
   - fastqc:
       path_filters:
@@ -109,17 +109,13 @@ top_modules:
       path_filters:
       - "*/*/map_genome/*"
 
-  - alfa:
-      name: "ALFA"
-      anchor: "ALFA"
+  - ALFA:
       path_filters:
-      - "*/ALFA_plots.concat_mqc.png"
+      - "*/*/ALFA/*/*ALFA_feature_counts.tsv"
 
-  - TIN_scores:
-      name: "TIN_scores"
-      anchor: "TIN_scores"
+  - tin-score:
       path_filters:
-      - "*/TIN_scores_boxplot_mqc.png"
+      - "*/*/TIN/TIN_score.tsv"
 
   - salmon:
       path_filters:
@@ -143,8 +139,6 @@ fn_clean_exts:
   - '.pseudo'
   - '.salmon'
   - '.sam'
-  - 'mqc'
-  - '.png'
 ..."""
 
     with open(options.config, "w") as config: