Add basic statistical pipeline metafx chisq and improve feature construction

Author: ivartb

.github/workflows/test.yml

@@ -37,6 +37,11 @@ jobs:
       run: |
         export PATH=bin:$PATH
         metafx colored -t 6 -m 6G -k 31 -i test_data/sample_list_3.txt --n-comps 100 --perc 0.7 -w wd_colored
+    - name: metafx chisq (with depth)
+      run: |
+        export PATH=bin:$PATH
+        metafx chisq -t 6 -m 6G -k 31 -i test_data/sample_list.txt -w wd_chisq_4 --skip-graph -n 1000
+        metafx chisq -t 6 -m 6G -k 31 -i test_data/sample_list_3.txt -w wd_chisq_3 --skip-graph -n 10000 --depth 15
     - name: metafx stats
       run: |
         export PATH=bin:$PATH

bin/metafx

@@ -16,9 +16,9 @@ help_message () {
     echo "    metaspades        Unsupervised feature extraction via metaSpades (https://cab.spbu.ru/software/meta-spades/)"
     echo ""
     echo "    unique            Supervised feature extraction using group-specific k-mers"
+    echo "    chisq             Supervised feature extraction using top significant k-mers by chi-squared test"
     echo "    stats             Supervised feature extraction using statistically significant k-mers"
     echo "    colored           Supervised feature extraction using group-colored de Bruijn graph"
-    echo "    lda               Supervised feature extraction via Latent Dirichlet allocation model"
     echo ""
     echo "    pca               PCA visualisation of samples based on extracted features"
     echo "    fit               Machine Learning methods to train classification model based on extracted features"
@@ -57,6 +57,10 @@ elif [ "$1" = unique ]; then
     echo metafx unique ${@:2} | tee -a $LOGFILE
     { time ${PIPES}/unique.sh ${@:2} 2>&1; echo $? >> $LOGFILE; } | tee -a $LOGFILE
     exit `tail -1 $LOGFILE`
+elif [ "$1" = chisq ]; then
+    echo metafx chisq ${@:2} | tee -a $LOGFILE
+    { time ${PIPES}/chisq.sh ${@:2} 2>&1; echo $? >> $LOGFILE; } | tee -a $LOGFILE
+    exit `tail -1 $LOGFILE`
 elif [ "$1" = stats ]; then
     echo metafx stats ${@:2} | tee -a $LOGFILE
     { time ${PIPES}/stats.sh ${@:2} 2>&1; echo $? >> $LOGFILE; } | tee -a $LOGFILE
@@ -65,11 +69,6 @@ elif [ "$1" = colored ]; then
     echo metafx colored ${@:2} | tee -a $LOGFILE
     { time ${PIPES}/colored.sh ${@:2} 2>&1; echo $? >> $LOGFILE; } | tee -a $LOGFILE
     exit `tail -1 $LOGFILE`
-elif [ "$1" = lda ]; then
-    echo "Not implemented error! Module is under development."
-    exit 1
-    #echo metafx lda ${@:2}
-    #time ${PIPES}/lda.sh ${@:2}
 elif [ "$1" = pca ]; then
     echo metafx pca ${@:2} | tee -a $LOGFILE
     { time ${PIPES}/pca.sh ${@:2} 2>&1; echo $? >> $LOGFILE; } | tee -a $LOGFILE