action.yml

name: 'Classify Failure'
description: 'Analyze failure logs and classify root cause. Pure Python, no dependencies.'
inputs:
  log_file:
    description: 'Path to the failure log file'
    required: true
  exit_code:
    description: 'Exit code of the failed command'
    required: false
    default: '1'
outputs:
  failure_class:
    description: 'Classification: flaky_test, network_timeout, race_condition, permission_denied, real_bug, unknown'
    value: ${{ steps.classify.outputs.failure_class }}
  failure_summary:
    description: 'One-line summary of the failure'
    value: ${{ steps.classify.outputs.failure_summary }}
  should_retry:
    description: 'Whether to retry (true/false)'
    value: ${{ steps.classify.outputs.should_retry }}
runs:
  using: 'composite'
  steps:
    - name: Classify failure
      id: classify
      shell: bash
      run: |
        LOG_FILE="${{ inputs.log_file }}"
        EXIT_CODE="${{ inputs.exit_code }}"

        export LOG_FILE EXIT_CODE

        CLASS=$(python3 <<'PY'
import os
import sys

log = ""
try:
    with open(os.environ.get("LOG_FILE", ""), encoding="utf-8", errors="replace") as f:
        log = f.read()
except OSError:
    log = ""

exit_code_raw = os.environ.get("EXIT_CODE", "1")
exit_code = int(exit_code_raw) if exit_code_raw.isdigit() else 1
log_lower = log.lower()

if any(p in log_lower for p in [
    "timeout", "timed out", "connection refused", "name resolution",
    "ssl", "tls", "network", "dns", "errno", "etimedout",
    "econnrefused", "econnreset", "httperror", "502", "503", "504",
]):
    print("network_timeout")
    sys.exit(0)

if any(p in log_lower for p in [
    "permission denied", "forbidden", "401", "403",
    "authentication failed", "token expired", "unauthorized",
    "secret", "not allowed",
]):
    print("permission_denied")
    sys.exit(0)

if any(p in log_lower for p in [
    "conflict", "409", "race", "concurrent",
    "another process", "locked", "busy",
    "already exists", "duplicate",
]):
    print("race_condition")
    sys.exit(0)

if any(p in log_lower for p in [
    "flaky", "intermittent", "sometimes fails",
    "passed on retry", "random", "unstable",
]):
    print("flaky_test")
    sys.exit(0)

if any(p in log_lower for p in [
    "assertionerror", "assert", "test failed",
    "pytest", "unittest", "coverage",
]):
    print("flaky_test")
    sys.exit(0)

if any(p in log_lower for p in [
    "syntaxerror", "typeerror", "nameerror", "attributeerror",
    "importerror", "modulenotfounderror", "keyerror", "indexerror",
    "valueerror", "zerodivisionerror", "segmentation fault",
    "core dumped", "fatal", "panic",
]):
    print("real_bug")
    sys.exit(0)

print("unknown")
PY
        )


        # Generate summary
        SUMMARY=$(python3 <<'PY'
import os

try:
    with open(os.environ.get("LOG_FILE", ""), encoding="utf-8", errors="replace") as f:
        lines = f.readlines()
    relevant = [line.strip() for line in lines if line.strip()][-3:]
    summary = " | ".join(relevant)[:200]
    print(summary if summary else "No log content available")
except OSError:
    print("Could not read log file")
PY
        )


        # Determine should_retry
        if [ "$CLASS" = "real_bug" ]; then
          RETRY="false"
        elif [ "$CLASS" = "permission_denied" ]; then
          RETRY="false"
        else
          RETRY="true"
        fi

        echo "failure_class=$CLASS" >> "$GITHUB_OUTPUT"
        echo "failure_summary=$SUMMARY" >> "$GITHUB_OUTPUT"
        echo "should_retry=$RETRY" >> "$GITHUB_OUTPUT"

        echo "📋 Classification: $CLASS"
        echo "📝 Summary: $SUMMARY"
        echo "🔄 Should retry: $RETRY"