#!/bin/sh
# docs-sync.cron.sh: Documentation sync orchestrator
#
# Automated cron job (runs daily @ 05:00 UTC):
# 1. Pull latest markdown from git
# 2. Compile markdown → HTML (with .docignore filtering)
# 3. Validate output
# 4. Deploy via atomic symlink swap (zero-downtime)
# 5. Cleanup old versions (30-day retention)
#
# Logs:
#   /var/log/clawdie-docs-sync.log — detailed output
#   .sync-metadata.json — deployment metadata
#
# Lock file prevents concurrent syncs:
#   <repo>/tmp/clawdie-docs-sync.lock

set -e

# ────────────────────────────────────────────────────────────────────────────
# Configuration
# ────────────────────────────────────────────────────────────────────────────

SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
REPO_DIR="$(cd "${SCRIPT_DIR}/.." && pwd)"
DOCS_DIR="${REPO_DIR}/docs/public"
DOCS_POLICY="${REPO_DIR}/docs/internal/DOCUMENTATION-POLICY.md"
SYNC_METADATA="${DOCS_DIR}/.sync-metadata.json"
COMPILE_SCRIPT="${SCRIPT_DIR}/docs-compile.sh"
LOG_FILE="${LOG_FILE:-/var/log/clawdie-docs-sync.log}"
LOCK_FILE="${LOCK_FILE:-${REPO_DIR}/tmp/clawdie-docs-sync.lock}"
RETENTION_DAYS="${RETENTION_DAYS:-30}"
LANGUAGES="${LANGUAGES:-sl,en,de,hr,sr,ru}"  # Slovenian primary; English source; German, Croatian, Serbian, Russian
DATE_FORMAT_SCRIPT="${REPO_DIR}/scripts/date-format.sh"

# Colors (for manual runs; cron will strip these)
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
BLUE='\033[0;34m'
NC='\033[0m'

# Deployment targets (from .sync-metadata.json)
TARGETS="docs-clawdie-si osa-smilepowered.org"

# ────────────────────────────────────────────────────────────────────────────
# Logging & Locking
# ────────────────────────────────────────────────────────────────────────────

log_msg() {
    local msg="$1"
    echo "$(bash "$DATE_FORMAT_SCRIPT" display-ts) [docs-sync] $msg" | tee -a "$LOG_FILE"
}

log_err() {
    local msg="$1"
    echo "$(bash "$DATE_FORMAT_SCRIPT" display-ts) ${RED}[ERROR]${NC} $msg" | tee -a "$LOG_FILE" >&2
}

log_ok() {
    local msg="$1"
    echo "$(bash "$DATE_FORMAT_SCRIPT" display-ts) ${GREEN}[OK]${NC} $msg" | tee -a "$LOG_FILE"
}

acquire_lock() {
    # Prevent concurrent syncs
    if [ -f "$LOCK_FILE" ]; then
        local lock_age=$(($(date +%s) - $(stat -f%m "$LOCK_FILE" 2>/dev/null || echo 0)))
        if [ "$lock_age" -lt 3600 ]; then
            log_err "Sync already in progress (lock age: ${lock_age}s)"
            exit 1
        else
            log_msg "Stale lock removed (age: ${lock_age}s)"
            rm -f "$LOCK_FILE"
        fi
    fi
    mkdir -p "$(dirname "$LOCK_FILE")"
    touch "$LOCK_FILE"
    trap 'rm -f "$LOCK_FILE"' EXIT
}

# ────────────────────────────────────────────────────────────────────────────
# Initialization
# ────────────────────────────────────────────────────────────────────────────

log_msg "════════════════════════════════════════════════════════════════════════"
log_msg "Documentation Sync Started"
log_msg "Repository: $REPO_DIR"
log_msg "Docs source: $DOCS_DIR"

acquire_lock

# Verify dependencies
for cmd in git pandoc jq rsync; do
    command -v "$cmd" >/dev/null 2>&1 || { log_err "$cmd not found"; exit 1; }
done

cd "$REPO_DIR" || exit 1

# ────────────────────────────────────────────────────────────────────────────
# Step 1: Pull latest markdown from git
# ────────────────────────────────────────────────────────────────────────────

log_msg "Step 1: Fetching latest changes from git..."

if git fetch origin 2>&1 | grep -q "fatal\|error"; then
    log_err "git fetch failed"
    exit 1
fi

# Check if docs/public has changes
CHANGES=$(git diff HEAD origin/implementation --name-only -- docs/public 2>/dev/null | wc -l || echo 0)
LAST_COMMIT_LOCAL=$(git rev-parse HEAD 2>/dev/null || echo "unknown")
LAST_COMMIT_REMOTE=$(git rev-parse origin/implementation 2>/dev/null || echo "unknown")

if [ "$CHANGES" -gt 0 ]; then
    log_msg "Found $CHANGES changes in docs/public — pulling..."
    git pull origin implementation --ff-only 2>&1 | grep -E "^(Already|Fast-forward|Merge)" || true
else
    log_msg "No changes in docs/public — skipping pull"
fi

CURRENT_COMMIT=$(git rev-parse HEAD 2>/dev/null || echo "unknown")
log_ok "Git sync complete (commit: ${CURRENT_COMMIT:0:7})"

# ────────────────────────────────────────────────────────────────────────────
# Step 2: Compile markdown → HTML for each target
# ────────────────────────────────────────────────────────────────────────────

log_msg "Step 2: Compiling markdown to HTML..."

# Read targets from .sync-metadata.json
# For now, hardcode targets; eventually read from metadata
COMPILE_DIRS=""
for target in $TARGETS; do
    case "$target" in
        docs-clawdie-si)
            COMPILE_DIRS="$COMPILE_DIRS /usr/local/www/docs.clawdie.si"
            ;;
        osa-smilepowered.org)
            COMPILE_DIRS="$COMPILE_DIRS /usr/local/www/osa.smilepowered.org"
            ;;
    esac
done

SEMVER=$(git describe --tags --match 'v*' 2>/dev/null | sed 's/^v//' | cut -d'-' -f1 || echo "0.0.0")
BUILD_DATE=$(bash "$DATE_FORMAT_SCRIPT" display-date | tr 'A-Z' 'a-z')

for out_dir in $COMPILE_DIRS; do
    if [ ! -d "$out_dir" ]; then
        log_err "Output directory not found: $out_dir"
        # Non-fatal: skip this target, continue with others
        continue
    fi

    log_msg "  Compiling to: $out_dir"

    # Compile each language
    for lang in $(echo "$LANGUAGES" | tr ',' ' '); do
        log_msg "    Language: $lang"

        if [ "$lang" = "en" ]; then
            if ! "$COMPILE_SCRIPT" \
                --semver "$SEMVER" \
                --date "$BUILD_DATE" \
                --filter "$DOCS_DIR/.docignore" \
                --source "$DOCS_DIR" \
                "$out_dir" 2>&1 | tee -a "$LOG_FILE"; then
                log_err "Compilation failed for $out_dir (language: $lang)"
                exit 1
            fi
            continue
        fi

        if [ ! -d "$DOCS_DIR/$lang" ]; then
            log_msg "    Language directory missing: $DOCS_DIR/$lang (skipping)"
            continue
        fi

        if ! "$COMPILE_SCRIPT" \
            --semver "$SEMVER" \
            --date "$BUILD_DATE" \
            --filter "$DOCS_DIR/.docignore" \
            --source "$DOCS_DIR" \
            --language "$lang" \
            "$out_dir" 2>&1 | tee -a "$LOG_FILE"; then
            log_err "Compilation failed for $out_dir (language: $lang)"
            exit 1
        fi
    done
done

log_ok "Compilation complete (v${SEMVER}, ${BUILD_DATE})"

# ────────────────────────────────────────────────────────────────────────────
# Step 3: Validate compiled output
# ────────────────────────────────────────────────────────────────────────────

log_msg "Step 3: Validating output..."

VALIDATION_OK=1
for out_dir in $COMPILE_DIRS; do
    [ ! -d "$out_dir" ] && continue

    VERSION_DIR="${out_dir}/docs-v${SEMVER}_${BUILD_DATE}"

    # Check essential files exist
    [ ! -f "${VERSION_DIR}/index.html" ] && {
        log_err "Missing index.html in $VERSION_DIR"
        VALIDATION_OK=0
    }

    # Count HTML files
    HTML_COUNT=$(find "$VERSION_DIR" -name "*.html" | wc -l)
    [ "$HTML_COUNT" -lt 3 ] && {
        log_err "Too few HTML files ($HTML_COUNT) in $VERSION_DIR"
        VALIDATION_OK=0
    }

    log_msg "  $out_dir: $HTML_COUNT HTML files ✓"
done

if [ "$VALIDATION_OK" -ne 1 ]; then
    log_err "Validation failed — aborting deployment"
    exit 1
fi

log_ok "Validation passed"

# ────────────────────────────────────────────────────────────────────────────
# Step 4: Deploy via atomic symlink swap
# ────────────────────────────────────────────────────────────────────────────

log_msg "Step 4: Deploying with atomic symlink swap..."

NGINX_RELOAD_NEEDED=0

for out_dir in $COMPILE_DIRS; do
    [ ! -d "$out_dir" ] && continue

    SYMLINK_NAME="docs-current"
    case "$(basename "$out_dir")" in
        *osa*)  SYMLINK_NAME="en-current" ;;
        *)      SYMLINK_NAME="docs-current" ;;
    esac

    VERSION_DIR="docs-v${SEMVER}_${BUILD_DATE}"
    SYMLINK_PATH="${out_dir}/${SYMLINK_NAME}"

    # Atomic swap (ln -sfn is atomic on modern filesystems)
    if ln -sfn "$VERSION_DIR" "$SYMLINK_PATH" 2>&1; then
        log_ok "  Symlink swap: $SYMLINK_NAME → $VERSION_DIR"
        NGINX_RELOAD_NEEDED=1
    else
        log_err "Failed to update symlink: $SYMLINK_PATH"
        exit 1
    fi

    # Verify symlink points to correct directory
    CURRENT=$(readlink "$SYMLINK_PATH" 2>/dev/null || echo "")
    if [ "$CURRENT" = "$VERSION_DIR" ]; then
        log_ok "  Verified: $SYMLINK_PATH → $CURRENT"
    else
        log_err "Symlink verification failed: expected $VERSION_DIR, got $CURRENT"
        exit 1
    fi
done

# Reload nginx to serve new content
if [ "$NGINX_RELOAD_NEEDED" -eq 1 ]; then
    log_msg "Reloading nginx..."
    if "${SCRIPT_DIR:-$(dirname "$0")}/hostd-call.sh" service-restart '{"name":"nginx","jail":"cms"}' 2>&1 | tee -a "$LOG_FILE"; then
        log_ok "Nginx reloaded (serving v${SEMVER})"
    else
        log_err "Nginx reload failed"
        exit 1
    fi
fi

log_ok "Deployment complete (zero-downtime swap)"

# ────────────────────────────────────────────────────────────────────────────
# Step 5: Cleanup old versions (30-day retention)
# ────────────────────────────────────────────────────────────────────────────

log_msg "Step 5: Cleaning up old versions (retention: ${RETENTION_DAYS} days)..."

CLEANUP_COUNT=0
for out_dir in $COMPILE_DIRS; do
    [ ! -d "$out_dir" ] && continue

    # Find version directories older than RETENTION_DAYS
    find "$out_dir" -maxdepth 1 -name "docs-v*" -o -name "*-v*" -type d 2>/dev/null | while read old_dir; do
        # Skip current version
        SYMLINK_TARGET=$(readlink "${out_dir}/docs-current" 2>/dev/null || echo "")
        [ "$(basename "$old_dir")" = "$SYMLINK_TARGET" ] && continue

        # Check age
        AGE_DAYS=$(($(date +%s) - $(stat -f%m "$old_dir" 2>/dev/null)))
        AGE_DAYS=$((AGE_DAYS / 86400))

        if [ "$AGE_DAYS" -gt "$RETENTION_DAYS" ]; then
            log_msg "  Removing $(basename "$old_dir") (${AGE_DAYS} days old)"
            rm -rf "$old_dir"
            CLEANUP_COUNT=$((CLEANUP_COUNT + 1))
        fi
    done
done

log_ok "Cleanup complete ($CLEANUP_COUNT old versions removed)"

# ────────────────────────────────────────────────────────────────────────────
# Update metadata
# ────────────────────────────────────────────────────────────────────────────

log_msg "Updating metadata..."

NEXT_SYNC=$(date -d "+1 day" -u +"%Y-%m-%dT05:00:00Z" 2>/dev/null || date -u -v+1d +"%Y-%m-%dT05:00:00Z")

jq \
    --arg last_sync "$(date -u +%Y-%m-%dT%H:%M:%SZ)" \
    --arg last_commit "$CURRENT_COMMIT" \
    --arg next_sync "$NEXT_SYNC" \
    '.last_sync = $last_sync | .last_commit = $last_commit | .next_scheduled_sync = $next_sync' \
    "$SYNC_METADATA" > "${SYNC_METADATA}.tmp" && mv "${SYNC_METADATA}.tmp" "$SYNC_METADATA" || true

log_ok "Metadata updated"

# ────────────────────────────────────────────────────────────────────────────
# Final Summary
# ────────────────────────────────────────────────────────────────────────────

log_msg "════════════════════════════════════════════════════════════════════════"
log_msg "${GREEN}✓ SYNC COMPLETE${NC}"
log_msg "Version:  v${SEMVER}"
log_msg "Built:    ${BUILD_DATE}"
log_msg "Commit:   ${CURRENT_COMMIT:0:7}"
log_msg "Targets:  $TARGETS"
log_msg "Cleanup:  $CLEANUP_COUNT old versions removed"
log_msg ""
log_msg "Live URLs:"
log_msg "  📚 docs.clawdie.si"
log_msg "  🌐 osa.smilepowered.org"
log_msg ""

exit 0