claude-evolve 1.6.0 → 1.6.1

package/bin/claude-evolve-clean-corrupted

@@ -0,0 +1,178 @@
+#!/bin/bash
+
+set -e
+
+# Function to show help
+show_help() {
+  cat <<EOF
+claude-evolve clean-corrupted - Remove corrupted records from evolution CSV
+
+USAGE:
+  claude-evolve clean-corrupted [OPTIONS]
+
+OPTIONS:
+  --dry-run   Show what would be removed without modifying the CSV
+  --help      Show this help message
+
+DESCRIPTION:
+  Removes corrupted records with invalid ID formats, such as:
+  - IDs with line numbers and pipe characters (e.g., "00648| gen43-001")
+  - IDs with leading digits followed by spaces
+  - Any ID containing pipe (|) characters
+
+  Creates a backup before modifying the CSV.
+
+EXAMPLES:
+  claude-evolve clean-corrupted           # Remove corrupted records
+  claude-evolve clean-corrupted --dry-run # Preview what would be removed
+EOF
+}
+
+# Parse arguments
+DRY_RUN=false
+
+while [[ $# -gt 0 ]]; do
+  case $1 in
+  --dry-run)
+    DRY_RUN=true
+    shift
+    ;;
+  --help)
+    show_help
+    exit 0
+    ;;
+  *)
+    echo "[ERROR] Unknown option: $1" >&2
+    exit 1
+    ;;
+  esac
+done
+
+# Load configuration
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# shellcheck source=../lib/config.sh
+source "$SCRIPT_DIR/../lib/config.sh"
+
+# Use CLAUDE_EVOLVE_CONFIG if set, otherwise default
+if [[ -n ${CLAUDE_EVOLVE_CONFIG:-} ]]; then
+  load_config "$CLAUDE_EVOLVE_CONFIG"
+else
+  # Check if config.yaml exists in current directory
+  if [[ -f "config.yaml" ]]; then
+    # Don't export to avoid collision with parallel runs
+    CONFIG_FILE="$(pwd)/config.yaml"
+    load_config "$CONFIG_FILE"
+  else
+    load_config
+  fi
+fi
+
+# Validate configuration
+if ! validate_config; then
+  echo "[ERROR] Configuration validation failed" >&2
+  exit 1
+fi
+
+# Check if CSV exists
+if [[ ! -f "$FULL_CSV_PATH" ]]; then
+  echo "[ERROR] Evolution CSV not found: $FULL_CSV_PATH" >&2
+  echo "Run 'claude-evolve setup' first or navigate to the correct directory" >&2
+  exit 1
+fi
+
+echo "Checking for corrupted records in: $FULL_CSV_PATH"
+
+# Use Python to identify and optionally remove corrupted records
+"$PYTHON_CMD" - <<EOF
+import csv
+import sys
+import re
+import shutil
+from datetime import datetime
+
+csv_path = "$FULL_CSV_PATH"
+dry_run = $DRY_RUN
+
+def is_valid_candidate_id(candidate_id):
+    """Check if a candidate ID is valid."""
+    if not candidate_id or candidate_id == "id":
+        return True  # Header row
+
+    # Reject IDs containing pipe characters (line number artifacts)
+    if '|' in candidate_id:
+        return False
+
+    # Valid ID should match: baseline-NNN or genNN-NNN format
+    # Also accept special IDs like "000", "0", etc.
+    if re.match(r'^(baseline|gen\d{2})-\d{3}$', candidate_id):
+        return True
+    if re.match(r'^(000|0|gen00-000)$', candidate_id):
+        return True
+
+    # Reject anything with leading digits followed by non-standard format
+    if re.match(r'^\d+\s', candidate_id):
+        return False
+
+    return True
+
+# Read CSV
+with open(csv_path, 'r') as f:
+    reader = csv.reader(f)
+    rows = list(reader)
+
+# Identify corrupted records
+corrupted_records = []
+valid_rows = [rows[0]]  # Keep header
+
+for i, row in enumerate(rows[1:], start=1):
+    if not row or len(row) == 0:
+        continue
+
+    candidate_id = row[0] if len(row) > 0 else ""
+
+    if not is_valid_candidate_id(candidate_id):
+        corrupted_records.append((i, candidate_id, row))
+    else:
+        valid_rows.append(row)
+
+# Report findings
+total_records = len(rows) - 1  # Exclude header
+corrupted_count = len(corrupted_records)
+
+print(f"\\nTotal records: {total_records}")
+print(f"Corrupted records: {corrupted_count}")
+print(f"Valid records: {len(valid_rows) - 1}")  # Exclude header
+
+if corrupted_count > 0:
+    print("\\n🔍 Corrupted records found:")
+    for line_num, candidate_id, row in corrupted_records[:20]:
+        status = row[4] if len(row) > 4 else "unknown"
+        print(f"  Line {line_num}: {candidate_id} (status: {status})")
+
+    if len(corrupted_records) > 20:
+        print(f"  ... and {len(corrupted_records) - 20} more")
+
+    if not dry_run:
+        # Create backup
+        backup_path = f"{csv_path}.backup.{datetime.now().strftime('%Y%m%d-%H%M%S')}"
+        shutil.copy2(csv_path, backup_path)
+        print(f"\\n✅ Created backup: {backup_path}")
+
+        # Write cleaned CSV
+        with open(csv_path, 'w', newline='') as f:
+            writer = csv.writer(f, quoting=csv.QUOTE_NONNUMERIC)
+            for row in valid_rows:
+                writer.writerow(row)
+
+        print(f"✅ Removed {corrupted_count} corrupted records from CSV")
+        print(f"✅ Cleaned CSV saved to: {csv_path}")
+    else:
+        print("\\n[DRY RUN] No changes made. Run without --dry-run to remove corrupted records.")
+else:
+    print("\\n✅ No corrupted records found!")
+
+sys.exit(0)
+EOF
+
+echo ""
+echo "Done!"

package/bin/claude-evolve-main

@@ -66,6 +66,7 @@ COMMANDS:
     status             Show evolution progress and current leader
     autostatus         Auto-updating status display (real-time)
     cleanup            Clean up unchanged algorithms and descendants
+    clean-corrupted    Remove corrupted records from evolution CSV
     cleanup-duplicates Alias for cleanup (deprecated)
     help               Show this help message
 
@@ -92,16 +93,17 @@ show_menu() {
   echo
   echo "What would you like to do?"
   echo
-  echo "  1) setup      - Initialize evolution workspace"
-  echo "  2) ideate     - Generate new algorithm ideas"
-  echo "  3) run        - Execute evolution candidates"
-  echo "  4) analyze    - Analyze evolution results"
-  echo "  5) edit       - Manage candidate statuses by generation"
-  echo "  6) status     - Show evolution progress and current leader"
-  echo "  7) autostatus - Auto-updating status display (real-time)"
-  echo "  8) config     - Manage configuration settings"
-  echo "  9) help       - Show help message"
-  echo "  0) exit       - Exit"
+  echo "  1) setup           - Initialize evolution workspace"
+  echo "  2) ideate          - Generate new algorithm ideas"
+  echo "  3) run             - Execute evolution candidates"
+  echo "  4) analyze         - Analyze evolution results"
+  echo "  5) edit            - Manage candidate statuses by generation"
+  echo "  6) status          - Show evolution progress and current leader"
+  echo "  7) autostatus      - Auto-updating status display (real-time)"
+  echo "  8) clean-corrupted - Remove corrupted records from CSV"
+  echo "  9) config          - Manage configuration settings"
+  echo "  h) help            - Show help message"
+  echo "  0) exit            - Exit"
   echo
 
   # Show workspace status
@@ -152,7 +154,7 @@ check_for_updates
 # Main logic
 if [[ $# -eq 0 ]]; then
   show_menu
-  read -r -p "Enter your choice (1-9, 0): " choice
+  read -r -p "Enter your choice (1-9, h, 0): " choice
 
   case $choice in
   1) exec "$SCRIPT_DIR/claude-evolve-setup" ;;
@@ -162,14 +164,15 @@ if [[ $# -eq 0 ]]; then
   5) exec "$SCRIPT_DIR/claude-evolve-edit" ;;
   6) exec "$SCRIPT_DIR/claude-evolve-status" ;;
   7) exec "$SCRIPT_DIR/claude-evolve-autostatus" ;;
-  8) exec "$SCRIPT_DIR/claude-evolve-config" ;;
-  9) show_help ;;
+  8) exec "$SCRIPT_DIR/claude-evolve-clean-corrupted" ;;
+  9) exec "$SCRIPT_DIR/claude-evolve-config" ;;
+  h|H) show_help ;;
   0)
     echo "Goodbye!"
     exit 0
     ;;
   *)
-    echo -e "${RED}Invalid choice. Please select 1-9 or 0.${NC}"
+    echo -e "${RED}Invalid choice. Please select 1-9, h, or 0.${NC}"
     exit 1
     ;;
   esac
@@ -214,6 +217,10 @@ cleanup-duplicates|cleanup)
   shift
   exec "$SCRIPT_DIR/claude-evolve-cleanup" "$@"
   ;;
+clean-corrupted)
+  shift
+  exec "$SCRIPT_DIR/claude-evolve-clean-corrupted" "$@"
+  ;;
 config)
   shift
   exec "$SCRIPT_DIR/claude-evolve-config" "$@"

package/lib/csv_fixer.py

@@ -2,25 +2,87 @@
 """
 CSV format fixer for claude-evolve
 Ensures proper quoting of CSV fields, especially descriptions
+Also filters out corrupted records with invalid ID formats
 """
 
 import csv
 import sys
+import re
+
+def is_valid_candidate_id(candidate_id):
+    """
+    Check if a candidate ID is valid.
+    Valid formats:
+    - baseline-000
+    - gen00-000
+    - gen01-001, gen02-042, etc.
+
+    Invalid formats (to reject):
+    - 00648| gen43-001 (line numbers with pipes)
+    - Any ID containing | character
+    - Any ID with leading numbers followed by |
+    """
+    if not candidate_id or candidate_id == "id":
+        return True  # Header row
+
+    # Reject IDs containing pipe characters (line number artifacts)
+    if '|' in candidate_id:
+        return False
+
+    # Valid ID should match: baseline-NNN or genNN-NNN format
+    # Also accept special IDs like "000", "0", etc.
+    if re.match(r'^(baseline|gen\d{2})-\d{3}$', candidate_id):
+        return True
+    if re.match(r'^(000|0|gen00-000)$', candidate_id):
+        return True
+
+    # Reject anything with leading digits followed by non-standard format
+    if re.match(r'^\d+\s', candidate_id):
+        return False
+
+    return True
 
 def fix_csv_format(input_file, output_file):
     """
     Read a CSV file and ensure all fields are properly quoted.
     The csv module handles quoting automatically based on content.
+    Also filters out rows with invalid candidate IDs.
     """
     with open(input_file, 'r') as infile:
         reader = csv.reader(infile)
         rows = list(reader)
-    
+
+    rejected_count = 0
+    filtered_rows = []
+
+    for i, row in enumerate(rows):
+        # Always keep header
+        if i == 0:
+            filtered_rows.append(row)
+            continue
+
+        # Skip empty rows
+        if not row or len(row) == 0:
+            continue
+
+        candidate_id = row[0] if len(row) > 0 else ""
+
+        # Check if candidate ID is valid
+        if not is_valid_candidate_id(candidate_id):
+            rejected_count += 1
+            print(f"[WARN] Rejecting corrupted record with invalid ID: {candidate_id}", file=sys.stderr)
+            continue
+
+        filtered_rows.append(row)
+
+    if rejected_count > 0:
+        print(f"[INFO] Filtered out {rejected_count} corrupted records", file=sys.stderr)
+
     with open(output_file, 'w', newline='') as outfile:
         writer = csv.writer(outfile, quoting=csv.QUOTE_NONNUMERIC)
-        
+
         # Write all rows - csv.writer handles quoting automatically
-        for row in rows:
+        for row in filtered_rows:
             writer.writerow(row)
 
 if __name__ == "__main__":

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-evolve",
-  "version": "1.6.0",
+  "version": "1.6.1",
   "bin": {
     "claude-evolve": "./bin/claude-evolve",
     "claude-evolve-main": "./bin/claude-evolve-main",