#!/usr/bin/env bash
# wiki-maintain.sh — Agent Estate Wiki maintenance entrypoint
# Run daily at 03:00 AWST via cron (wiki-stocktake-maintenance job)
# 
# Does three things:
# 1. Regenerates systems/skills-index.md from live ~/.hermes/skills/ tree
# 2. Audits EVERY page for staleness (>21 days since `updated` frontmatter)
# 3. Rebuilds Quartz static site into public/

set -euo pipefail

WIKI_DIR="/home/hermes/workspace/agent-estate-wiki"
SKILLS_DIR="$HOME/.hermes/skills"
STOCKTAKE_DIR="$WIKI_DIR/raw/stocktake"
QUARTZ_CONTENT_DIR="$WIKI_DIR/public"
TODAY=$(date +%Y%m%d)

cd "$WIKI_DIR"

echo "[wiki-maintain] Starting maintenance at $(date -Iseconds)"
echo ""

# =============================================================================
# [1/3] Regenerate skills-index.md from live skill tree
# =============================================================================
echo "[1/3] Regenerating systems/skills-index.md from $SKILLS_DIR..."

python3 << 'PYTHON_SCRIPT'
import sys
from pathlib import Path
from datetime import date
import subprocess

skills_dir = Path("/home/hermes/.hermes/skills")
output = Path("/home/hermes/workspace/agent-estate-wiki/systems/skills-index.md")

# Use find to get all SKILL.md files, then group by their immediate parent category
# This handles nested structures like context-engineering-collection/skills/foo/SKILL.md
result = subprocess.run(
    ["find", str(skills_dir), "-name", "SKILL.md", "-type", "f"],
    capture_output=True, text=True
)
skill_files = [Path(p) for p in result.stdout.strip().split('\n') if p]

# Group by the top-level category (first directory under skills_dir)
categories = {}
for skill_md in skill_files:
    # Get relative path from skills_dir
    try:
        rel = skill_md.relative_to(skills_dir)
    except ValueError:
        continue  # Skip if not under skills_dir (e.g., broken symlinks)
    
    # Top-level category is the first part of the path
    category = rel.parts[0]
    skill_name = skill_md.parent.name  # The directory containing SKILL.md
    
    # Read the description from the skill file
    try:
        content = skill_md.read_text(errors='ignore')[:2000]
        lines = content.split('\n')
        desc = ""
        in_frontmatter = False
        for line in lines:
            if line.strip() == '---':
                in_frontmatter = not in_frontmatter
                continue
            if not in_frontmatter and line.strip():
                desc = line.strip()
                break
        if len(desc) > 200:
            desc = desc[:197] + "..."
    except Exception as e:
        desc = f"[error reading: {e}]"
    
    if category not in categories:
        categories[category] = []
    # Avoid duplicates (same skill_name in same category)
    if not any(name == skill_name for name, _ in categories[category]):
        categories[category].append((skill_name, desc))

# Sort skills within each category
for cat in categories:
    categories[cat] = sorted(categories[cat], key=lambda x: x[0])

# Build the markdown
today = date.today().isoformat()
lines = [
    "---",
    "title: Installed Skills",
    "type: system",
    "status: active",
    "created: 2026-07-22",
    f"updated: {today}",
    f"verified_on: {today}",
    "confidence: high",
    "tags: [skills, hermes]",
    "sources: []",
    "---",
    "",
    "# Installed Skills",
    "",
    f"**Total installed: {sum(len(v) for v in categories.values())} skills** in `~/.hermes/skills/`.",
    "",
    "This page is auto-regenerated from the live skill inventory. Do not hand-edit.",
    "",
    "## Stock-take notes",
    "- Category = directory structure under `~/.hermes/skills/` (`category/skill/SKILL.md`).",
    "- Skills directly under `~/.hermes/skills/` group under `uncategorized`.",
    "",
]

# Write categories (sorted, excluding 'uncategorized' which goes last)
for cat_name in sorted(categories.keys()):
    skills = categories[cat_name]
    # Format category name (capitalize first letter of each word)
    cat_display = cat_name.replace('-', ' ').title()
    lines.append(f"## {cat_display} ({len(skills)})")
    lines.append("")
    for skill_name, desc in skills:
        # Format skill name
        skill_display = skill_name.replace('-', ' ')
        if desc:
            lines.append(f"- `{skill_name}` — {desc}")
        else:
            lines.append(f"- `{skill_name}` — (no description)")
    lines.append("")

output.write_text('\n'.join(lines))
print(f"  Written {output} with {sum(len(v) for v in categories.values())} skills in {len(categories)} categories")
PYTHON_SCRIPT

echo ""

# =============================================================================
# [2/3] Freshness audit — check every page's `updated` frontmatter
# =============================================================================
REPORT="$STOCKTAKE_DIR/freshness-$TODAY.txt"
echo "[2/3] Auditing wiki freshness (threshold: >21 days)..."
python3 "$STOCKTAKE_DIR/_freshness_audit.py" "$WIKI_DIR" "$REPORT" 21

echo ""

# =============================================================================
# [3/3] Rebuild Quartz static site
# =============================================================================
echo "[3/3] Checking Quartz build..."
# Quartz is already built into public/ — if there's a build script, run it
# For now, the public/ dir is the Quartz output, regenerated when content changes
# If Quartz CLI is available, we'd run: npx quartz build
if command -v npx &> /dev/null && [ -f "$WIKI_DIR/quartz.config.ts" ]; then
    echo "  Running Quartz build..."
    cd "$WIKI_DIR"
    npx quartz build 2>&1 | head -20 || echo "  Quartz build had issues (non-fatal)"
else
    echo "  Quartz build skipped (no quartz.config.ts or npx not available)"
    echo "  public/ directory contains the current static site"
fi

echo ""
echo "[wiki-maintain] Maintenance complete at $(date -Iseconds)"
echo "  Freshness report: $REPORT"
echo ""

# Print summary
head -5 "$REPORT"
tail -3 "$REPORT"