Merge pull request #73 from KumarLabJax/KLAUS-118-add-bout-table-outputs-to-jabs-features

Add nextflow process to merge bout tables by behavior

Author: bergsalex

nextflow/modules/jabs_classifiers.nf

@@ -101,4 +101,57 @@ process BEHAVIOR_TABLE_TO_FEATURES {
     """
     Rscript ${params.support_code_dir}behavior_summaries.R -f ${in_summary_table} -b ${bin_size} -o "${in_summary_table.baseName}_features_${bin_size}.csv"
     """
+}
+
+/**
+* Aggregate bout tables by behavior across all videos.
+*
+* This process uses jabs-postprocess to merge tables by behavior,
+* creating separate merged files for each behavior detected.
+*
+* @param bout_tables List of paths to bout tables to be merged.
+*
+* @return merged_bout_tables List of paths to merged bout tables, one per behavior.
+* @return merge_log Path to the log file detailing the merge process.
+*/
+process AGGREGATE_BOUT_TABLES {
+    label "jabs_table_convert"
+    label "r_jabs_table_convert"
+
+    publishDir "${params.outdir}/merged_behavior_tables", mode: 'copy'
+
+    input:
+    path bout_tables
+
+    output:
+    path("merged_*_bouts_merged.csv"), emit: merged_bout_tables
+    path("merge_log.txt"), emit: merge_log
+
+    script:
+    """
+    # Create a temporary directory for organizing tables
+    mkdir -p table_staging
+    
+    # Copy all bout tables to staging directory
+    for table in ${bout_tables}; do
+        cp "\${table}" table_staging/
+    done
+    
+    # Use jabs-postprocess to merge tables by behavior
+    # This will automatically detect behaviors and create separate merged files for each
+    echo "Starting behavior table merging..." > merge_log.txt
+    echo "Input tables found:" >> merge_log.txt
+    ls table_staging/*.csv >> merge_log.txt
+    
+    uv run python -m jabs_postprocess.cli.main merge-multiple-tables \\
+        --table-folder table_staging \\
+        --table-pattern "*_bouts.csv" \\
+        --output-prefix merged \\
+        --overwrite \\
+        2>&1 | tee -a merge_log.txt
+    
+    # Log completion
+    echo "Merge completed. Output files:" >> merge_log.txt
+    ls merged_*_bouts_merged.csv >> merge_log.txt 2>/dev/null || echo "No merged files generated" >> merge_log.txt
+    """
 }

nextflow/workflows/feature_generation.nf

@@ -29,7 +29,8 @@ include { GENERATE_FEATURE_CACHE;
           PREDICT_CLASSIFIERS;
           GENERATE_BEHAVIOR_TABLES;
           PREDICT_HEURISTICS;
-          BEHAVIOR_TABLE_TO_FEATURES } from "${projectDir}/nextflow/modules/jabs_classifiers"
+          BEHAVIOR_TABLE_TO_FEATURES;
+          AGGREGATE_BOUT_TABLES } from "${projectDir}/nextflow/modules/jabs_classifiers"
 include { EXTRACT_FECAL_BOLI_BINS } from "${projectDir}/nextflow/modules/fecal_boli"
 
 workflow SINGLE_MOUSE_V2_FEATURES {
@@ -90,6 +91,14 @@ workflow SINGLE_MOUSE_V6_FEATURES {
     classifier_predictions = PREDICT_CLASSIFIERS(cached_features, params.single_mouse_classifiers)
     classifier_tables = GENERATE_BEHAVIOR_TABLES(classifier_predictions, params.single_mouse_classifiers)
 
+    // Aggregate bout tables by behavior for downstream analysis
+    all_bout_tables = heuristic_tables
+        .concat(classifier_tables)
+        .map { bout_table, summary_table -> bout_table }
+        .flatten()
+        .collect()
+    merged_bout_tables = AGGREGATE_BOUT_TABLES(all_bout_tables)
+
     // Combine table data into feature file
     all_summary_tables = heuristic_tables
         .concat(classifier_tables)
@@ -121,4 +130,5 @@ workflow SINGLE_MOUSE_V6_FEATURES {
     emit:
     wide_jabs_features
     fecal_boli_table
+    merged_bout_tables
 }