/debug-failed

You are debugging a failed VGP workflow for a specific species.

Your Task

The species name or assembly ID should be provided as an argument. Example: /debug-failed bTaeGut2

1. Check metadata first (token-efficient):

   cat metadata/metadata_run.json | jq '.bTaeGut2'
   

   • Identify which workflow failed
   • Get the invocation ID
   • Check the failure state

2. Read log file efficiently (NOT the entire file):

   # Read only the end where errors appear
   tail -100 {assembly_id}/planemo_log/{assembly_id}_Workflow_{N}.log
   
   # Or filter for errors
   grep -A 20 -i "error\|fail" {assembly_id}/planemo_log/{assembly_id}_Workflow_{N}.log | head -100
   

3. Identify error pattern:

   • Check against known VGP failure patterns:
     • WF0: No mitochondrial reads (expected, not a real error)
     • WF8: Missing Hi-C files or WF4 not complete
     • WF9: NCBI datasets tool not installed
     • General: Parameter issues, file upload failures

4. Provide diagnosis:

   • What failed and why
   • Whether it's expected (e.g., WF0 no mito reads)
   • Specific line numbers or error messages
   • Root cause analysis

5. Suggest fixes:

   • Specific parameter changes needed
   • Commands to edit job YAML if needed
   • Command to retry:

     vgp-run-all --resume --retry-failed -p profile.yaml -m ./metadata/
     

   • Whether to use --no-cache for fresh execution

Output Format

Provide:

• 🔍 Error Summary: What failed
• 📋 Diagnosis: Why it failed
• 💡 Solution: How to fix it
• ⚙️ Commands: Exact commands to run