wip: Further patches to file processing with precomputed boltzgen results

Author: FloWuenne

assets/NO_MSA

@@ -198,6 +198,23 @@ process BOLTZ2_REFOLD {
     echo "Affinity predictions: \${AFFINITY_COUNT}"
     echo "Output directory: ${meta.id}_boltz2_output"
     echo "============================================"
+
+    # Fail if no structures were produced (critical for downstream processes)
+    if [ "\${CIF_COUNT}" -eq 0 ]; then
+        echo ""
+        echo "ERROR: No CIF structures were produced by Boltz-2!"
+        echo "This will cause downstream processes (IPSAE, PRODIGY) to have no input."
+        echo ""
+        echo "Debug info - checking boltz2_results directory structure:"
+        find boltz2_results -type f 2>/dev/null | head -50 || echo "No boltz2_results directory found"
+        exit 1
+    fi
+
+    # Warn if no PAE files (needed for IPSAE)
+    if [ "\${NPZ_COUNT}" -eq 0 ]; then
+        echo ""
+        echo "WARNING: No PAE NPZ files were produced. IPSAE will not be able to run."
+    fi
     
     # Create summary file
     cat > ${meta.id}_boltz2_output/prediction_summary.txt <<SUMMARY

assets/NO_TEMPLATE

@@ -21,32 +21,40 @@ process CONSOLIDATE_METRICS {
     script:
     def pae_cutoff = params.ipsae_pae_cutoff ?: 10
     def dist_cutoff = params.ipsae_dist_cutoff ?: 10
-    def seq_dir_flag = sequence_files.name != 'NO_SEQUENCE_FILES' ? '--sequence_dir "sequences"' : ''
 
     """
     # Make script executable
     chmod +x ${consolidate_script}
 
+    # Create directories if they don't exist (handles empty input lists)
+    mkdir -p ipsae prodigy foldseek sequences
+
     # Debug: List staged files in subdirectories
     echo "=== Staged ipSAE files ==="
-    ls -la ipsae/ 2>/dev/null || echo "No ipsae directory"
+    ls -la ipsae/ 2>/dev/null || echo "No ipsae files"
     echo ""
     echo "=== Staged Prodigy files ==="
-    ls -la prodigy/ 2>/dev/null || echo "No prodigy directory"
+    ls -la prodigy/ 2>/dev/null || echo "No prodigy files"
     echo ""
     echo "=== Staged Foldseek files ==="
-    ls -la foldseek/ 2>/dev/null || echo "No foldseek directory"
+    ls -la foldseek/ 2>/dev/null || echo "No foldseek files"
     echo ""
     echo "=== Staged Sequence files ==="
-    ls -la sequences/ 2>/dev/null || echo "No sequences directory"
+    ls -la sequences/ 2>/dev/null || echo "No sequence files"
     echo ""
 
+    # Build command with optional sequence directory (only if files exist)
+    SEQ_FLAG=""
+    if [ -n "\$(ls -A sequences/ 2>/dev/null)" ]; then
+        SEQ_FLAG="--sequence_dir sequences"
+    fi
+
     # Run consolidation script with staged subdirectories
     python ${consolidate_script} \\
         --ipsae_dir "ipsae" \\
         --prodigy_dir "prodigy" \\
         --foldseek_dir "foldseek" \\
-        ${seq_dir_flag} \\
+        \${SEQ_FLAG} \\
         --output_html design_metrics_report.html \\
         --output_csv design_metrics_summary.csv \\
         --title "Protein Design Metrics Report" \\

modules/local/boltz2_refold.nf

@@ -141,9 +141,10 @@ workflow PROTEIN_DESIGN {
             // ================================================================
             // Prepare Target MSA from Samplesheet
             // ================================================================
+            // Use actual placeholder files in assets/ for k8s compatibility (avoids staging non-existent files)
             ch_target_msa = ch_input
                 .map { meta, design_yaml, structure_files, target_msa, target_sequence, target_template, boltzgen_output_dir ->
-                    def msa_file = target_msa ?: file('NO_MSA')
+                    def msa_file = target_msa ?: file("${projectDir}/assets/NO_MSA", checkIfExists: true)
                     [meta.id, msa_file]
                 }
 
@@ -152,7 +153,7 @@ workflow PROTEIN_DESIGN {
             // ================================================================
             ch_target_template = ch_input
                 .map { meta, design_yaml, structure_files, target_msa, target_sequence, target_template, boltzgen_output_dir ->
-                    def template_file = target_template ?: file('NO_TEMPLATE')
+                    def template_file = target_template ?: file("${projectDir}/assets/NO_TEMPLATE", checkIfExists: true)
                     [meta.id, template_file]
                 }
 
@@ -398,30 +399,31 @@ workflow PROTEIN_DESIGN {
 
         // Collect output files from each analysis process
         // These will be staged into the consolidation task's work directory
+        // Use empty lists [] instead of non-existent placeholder files for k8s compatibility
 
         // ipSAE scores (the .txt files, not byres)
         ch_ipsae_files = (params.run_ipsae && params.run_proteinmpnn && params.run_boltz2_refold)
             ? IPSAE_CALCULATE.out.scores
                 .map { meta, file -> file }
                 .collect()
-                .ifEmpty { file('NO_IPSAE_FILES') }
-            : Channel.value(file('NO_IPSAE_FILES'))
+                .ifEmpty { [] }
+            : Channel.value([])
 
         // Prodigy results (.txt files)
         ch_prodigy_files = (params.run_prodigy && params.run_proteinmpnn && params.run_boltz2_refold)
             ? PRODIGY_PREDICT.out.results
                 .map { meta, file -> file }
                 .collect()
-                .ifEmpty { file('NO_PRODIGY_FILES') }
-            : Channel.value(file('NO_PRODIGY_FILES'))
+                .ifEmpty { [] }
+            : Channel.value([])
 
         // Foldseek summaries (.tsv files)
         ch_foldseek_files = (params.run_foldseek && params.run_proteinmpnn && params.run_boltz2_refold)
             ? FOLDSEEK_SEARCH.out.summary
                 .map { meta, file -> file }
                 .collect()
-                .ifEmpty { file('NO_FOLDSEEK_FILES') }
-            : Channel.value(file('NO_FOLDSEEK_FILES'))
+                .ifEmpty { [] }
+            : Channel.value([])
 
         // ====================================================================
         // Collect binder sequences from ProteinMPNN for the report
@@ -435,9 +437,9 @@ workflow PROTEIN_DESIGN {
                     fasta_list.collect { fasta_file -> fasta_file }
                 }
                 .collect()
-                .ifEmpty { file('NO_SEQUENCE_FILES') }
+                .ifEmpty { [] }
         } else {
-            ch_sequence_files = Channel.value(file('NO_SEQUENCE_FILES'))
+            ch_sequence_files = Channel.value([])
         }
 
         // Run consolidation with staged files