Find stale PRs where the author hasn't commented in 2 weeks

stale-prs.sh

#!/usr/bin/env bash
set -uo pipefail

# =============================================================================
# Kubernetes PR Wrangler Helper Script
# Finds PRs that should be closed according to:
# https://kubernetes.io/docs/contribute/participate/pr-wranglers/#when-to-close-pull-requests
# =============================================================================

help_message() {
    cat <<EOF
Usage: $0 [-r <repository>] [-d <inactivity days>] [-c <cla days>] [-v]
  -r   Repository in the format owner/repo (default: kubernetes/website)
  -d   Inactivity period in days for no response (default: 14)
  -c   CLA unsigned period in days (default: 14)
  -v   Enable verbose output for troubleshooting
  -h   Show this help message

Criteria for closing PRs (per Kubernetes docs):
  1. Author hasn't signed the CLA for two weeks
  2. Author has not responded to comments or feedback in 2 or more weeks

Note: Authors listed in OWNERS_ALIASES are skipped (trusted contributors).
EOF
}

# Default values
REPO="kubernetes/website"
INACTIVE_DAYS=14
CLA_DAYS=14
VERBOSE=0

while getopts "r:d:c:vh" opt; do
    case $opt in
        r) REPO="$OPTARG" ;;
        d) INACTIVE_DAYS="$OPTARG" ;;
        c) CLA_DAYS="$OPTARG" ;;
        v) VERBOSE=1 ;;
        h) help_message; exit 0 ;;
        *) help_message; exit 1 ;;
    esac
done

log_verbose() {
    if [[ $VERBOSE -eq 1 ]]; then
        echo "[DEBUG] $*" >&2
    fi
}

log_info() {
    echo "[INFO] $*"
}

log_flag() {
    echo "[FLAG] $*"
}

# Validate dependencies
for cmd in gh jq date; do
    if ! command -v "$cmd" &> /dev/null; then
        echo "Error: Required command '$cmd' not found" >&2
        exit 1
    fi
done

log_info "Repository: $REPO"
log_info "Inactivity period (days): $INACTIVE_DAYS"
log_info "CLA unsigned period (days): $CLA_DAYS"

# Calculate threshold dates (portable for macOS and Linux)
get_threshold_date() {
    local days=$1
    if date -v-1d +%s &> /dev/null; then
        date -v-"${days}"d +%Y-%m-%dT%H:%M:%SZ
    else
        date -d "${days} days ago" +%Y-%m-%dT%H:%M:%SZ
    fi
}

date_to_seconds() {
    local date_str=$1
    if date -v-1d +%s &> /dev/null; then
        date -jf "%Y-%m-%dT%H:%M:%SZ" "$date_str" +%s 2>/dev/null || echo 0
    else
        date -d "$date_str" +%s 2>/dev/null || echo 0
    fi
}

format_timestamp() {
    local ts=$1
    date -r "$ts" +%Y-%m-%d 2>/dev/null || date -d "@$ts" +%Y-%m-%d 2>/dev/null || echo "unknown"
}

INACTIVE_DATE=$(get_threshold_date "$INACTIVE_DAYS")
INACTIVE_DATE_SEC=$(date_to_seconds "$INACTIVE_DATE")
CLA_DATE=$(get_threshold_date "$CLA_DAYS")
CLA_DATE_SEC=$(date_to_seconds "$CLA_DATE")

log_verbose "Inactive threshold date: $INACTIVE_DATE ($INACTIVE_DATE_SEC)"
log_verbose "CLA threshold date: $CLA_DATE ($CLA_DATE_SEC)"

# =============================================================================
# Fetch and parse OWNERS_ALIASES to skip trusted contributors
# =============================================================================

TRUSTED_AUTHORS_LIST=""

fetch_trusted_authors() {
    log_info "Fetching OWNERS_ALIASES from $REPO..."
    
    local owners_content
    owners_content=$(gh api "repos/$REPO/contents/OWNERS_ALIASES" --jq '.content' 2>/dev/null | base64 -d 2>/dev/null || echo "")
    
    if [[ -z "$owners_content" ]]; then
        log_verbose "Could not fetch OWNERS_ALIASES file, skipping trusted author check"
        return
    fi
    
    # Extract all unique usernames from lines like "    - username" or "    - username # comment"
    TRUSTED_AUTHORS_LIST=$(echo "$owners_content" \
        | grep -E '^\s*-\s+[a-zA-Z0-9]' \
        | sed 's/^[^-]*-\s*//' \
        | sed 's/\s*#.*//' \
        | tr -d ' ' \
        | tr '[:upper:]' '[:lower:]' \
        | sort -u || true)
    
    local count
    count=$(echo "$TRUSTED_AUTHORS_LIST" | grep -c . || echo 0)
    
    log_info "Loaded $count trusted authors from OWNERS_ALIASES"
    log_verbose "Trusted authors: $(echo $TRUSTED_AUTHORS_LIST | tr '\n' ' ')"
}

is_trusted_author() {
    local author=$1
    local author_lower="${author,,}"
    
    if echo "$TRUSTED_AUTHORS_LIST" | grep -qix "$author_lower"; then
        return 0  # trusted
    fi
    return 1  # not trusted
}

fetch_trusted_authors

# =============================================================================
# Arrays to store flagged PR details
# Format: "timestamp|pr_number|url|author|reason|has_lgtm"
# =============================================================================

declare -a cla_flagged_prs=()
declare -a inactive_flagged_prs=()

# =============================================================================
# Check author activity
# =============================================================================

check_author_activity() {
    local pr_number=$1
    local author=$2
    local threshold_sec=$3
    
    log_verbose "Checking activity for PR #$pr_number by $author"
    
    local issue_comments
    issue_comments=$(gh api --paginate "repos/$REPO/issues/$pr_number/comments" \
        --jq "[.[] | select(.user.login == \"$author\") | .created_at] | join(\"\n\")" 2>/dev/null || echo "")
    
    local review_comments
    review_comments=$(gh api --paginate "repos/$REPO/pulls/$pr_number/comments" \
        --jq "[.[] | select(.user.login == \"$author\") | .created_at] | join(\"\n\")" 2>/dev/null || echo "")
    
    local reviews
    reviews=$(gh api --paginate "repos/$REPO/pulls/$pr_number/reviews" \
        --jq "[.[] | select(.user.login == \"$author\") | .submitted_at] | join(\"\n\")" 2>/dev/null || echo "")
    
    local commits
    commits=$(gh api --paginate "repos/$REPO/pulls/$pr_number/commits" \
        --jq "[.[] | select(.author.login == \"$author\" or .committer.login == \"$author\") | .commit.committer.date] | join(\"\n\")" 2>/dev/null || echo "")
    
    local all_timestamps
    all_timestamps=$(printf '%s\n%s\n%s\n%s' "$issue_comments" "$review_comments" "$reviews" "$commits" | grep -v '^$' | sort -r || true)
    
    log_verbose "All activity timestamps for PR #$pr_number: $all_timestamps"
    
    if [[ -z "$all_timestamps" ]]; then
        echo "none"
        return
    fi
    
    local last_activity
    last_activity=$(echo "$all_timestamps" | head -n 1)
    local last_activity_sec
    last_activity_sec=$(date_to_seconds "$last_activity")
    
    if (( last_activity_sec > threshold_sec )); then
        echo "recent:$last_activity"
    else
        echo "stale:$last_activity"
    fi
}

# =============================================================================
# Process a single PR
# =============================================================================

process_pr() {
    local pr_json=$1
    
    local pr_number url author created_at labels_json
    pr_number=$(echo "$pr_json" | jq -r '.number')
    url=$(echo "$pr_json" | jq -r '.url')
    author=$(echo "$pr_json" | jq -r '.author.login')
    created_at=$(echo "$pr_json" | jq -r '.createdAt')
    labels_json=$(echo "$pr_json" | jq -c '[.labels[].name]')
    
    local created_sec
    created_sec=$(date_to_seconds "$created_at")
    
    log_verbose "Processing PR #$pr_number by $author (created: $created_at)"
    log_verbose "Labels: $labels_json"
    
    # Skip PRs from trusted authors (OWNERS_ALIASES members)
    if is_trusted_author "$author"; then
        log_verbose "Skipping PR #$pr_number: author ($author) is in OWNERS_ALIASES"
        return
    fi
    
    # Skip PRs that are too new to evaluate
    if (( created_sec > INACTIVE_DATE_SEC )); then
        log_verbose "Skipping PR #$pr_number: created within the inactivity period"
        return
    fi
    
    # Skip work-in-progress PRs
    if echo "$labels_json" | jq -e 'map(select(startswith("do-not-merge") or . == "wip")) | length > 0' > /dev/null; then
        log_verbose "Skipping PR #$pr_number: marked as work in progress or do-not-merge"
        return
    fi
    
    # Filter: only process PRs with the "language/en" label (for kubernetes/website)
    if [[ "$REPO" == "kubernetes/website" ]]; then
        if ! echo "$labels_json" | jq -e 'map(select(. == "language/en")) | length > 0' > /dev/null; then
            log_verbose "Skipping PR #$pr_number: not labeled as language/en"
            return
        fi
    fi
    
    # Check if PR has lgtm label
    local has_lgtm="no"
    if echo "$labels_json" | jq -e 'map(select(. == "lgtm")) | length > 0' > /dev/null; then
        has_lgtm="yes"
    fi
    
    # Condition 1: CLA not signed for CLA_DAYS
    if echo "$labels_json" | jq -e 'map(select(. == "cncf-cla: no")) | length > 0' > /dev/null; then
        if (( created_sec < CLA_DATE_SEC )); then
            log_flag "PR #$pr_number: $url"
            log_flag "  → Author ($author) hasn't signed the CLA (PR older than $CLA_DAYS days)"
            # Store with timestamp for sorting (oldest first) and lgtm status
            cla_flagged_prs+=("$created_sec|$pr_number|$url|$author|CLA not signed since $created_at|$has_lgtm")
        else
            log_verbose "PR #$pr_number has unsigned CLA but is within grace period"
        fi
    fi
    
    # Condition 2: Author hasn't responded in INACTIVE_DAYS
    local activity_status
    activity_status=$(check_author_activity "$pr_number" "$author" "$INACTIVE_DATE_SEC")
    
    case "$activity_status" in
        none)
            log_flag "PR #$pr_number: $url"
            log_flag "  → Author ($author) has never commented or committed since PR creation ($created_at)"
            inactive_flagged_prs+=("$created_sec|$pr_number|$url|$author|No activity since creation|$has_lgtm")
            ;;
        stale:*)
            local last_date=${activity_status#stale:}
            log_flag "PR #$pr_number: $url"
            log_flag "  → Author ($author) last active on $last_date (over $INACTIVE_DAYS days ago)"
            inactive_flagged_prs+=("$created_sec|$pr_number|$url|$author|Last active: $last_date|$has_lgtm")
            ;;
        recent:*)
            log_verbose "PR #$pr_number: Author has recent activity"
            ;;
    esac
}

# =============================================================================
# Open URLs in Firefox (oldest first)
# =============================================================================

open_in_firefox() {
    local -a urls=("$@")
    
    if [[ ${#urls[@]} -eq 0 ]]; then
        echo "No URLs to open."
        return
    fi
    
    echo ""
    echo "Opening ${#urls[@]} PRs in Firefox (oldest first)..."
    
    for url in "${urls[@]}"; do
        log_verbose "Opening: $url"
        if [[ "$OSTYPE" == "darwin"* ]]; then
            open -a "Firefox" "$url" 2>/dev/null || open "$url" 2>/dev/null
        elif command -v firefox &> /dev/null; then
            firefox "$url" &>/dev/null &
        elif command -v xdg-open &> /dev/null; then
            xdg-open "$url" &>/dev/null &
        else
            echo "Warning: Could not open $url - no suitable browser found"
        fi
        # Small delay to maintain tab order
        sleep 0.3
    done
    
    echo "Done! Opened ${#urls[@]} tabs."
}

# =============================================================================
# Print summary for a category
# =============================================================================

print_category_summary() {
    local title=$1
    shift
    local -a entries=("$@")
    
    if [[ ${#entries[@]} -eq 0 ]]; then
        return
    fi
    
    echo ""
    echo "$title (${#entries[@]}):"
    echo "----------------------------------------------"
    
    # Sort by timestamp (first field) - oldest first
    while IFS= read -r entry; do
        IFS='|' read -r ts num url auth reason has_lgtm <<< "$entry"
        lgtm_indicator=""
        if [[ "$has_lgtm" == "yes" ]]; then
            lgtm_indicator=" [LGTM]"
        fi
        echo "  PR #$num - $auth (created: $(format_timestamp "$ts"))$lgtm_indicator"
        echo "    $url"
        echo "    Reason: $reason"
    done < <(printf '%s\n' "${entries[@]}" | sort -t'|' -k1 -n)
}

# =============================================================================
# Main execution
# =============================================================================

log_info "Fetching open PRs from $REPO..."

PR_JSON=$(gh pr list -R "$REPO" \
    --state open \
    --limit 1000 \
    --json number,url,author,createdAt,labels)

PR_COUNT=$(echo "$PR_JSON" | jq 'length')
log_info "Found $PR_COUNT open PRs to evaluate"

if [[ "$PR_COUNT" -eq 0 ]]; then
    log_info "No open PRs found."
    exit 0
fi

PROCESSED=0
while IFS= read -r pr; do
    process_pr "$pr"
    ((PROCESSED++)) || true
    
    if (( PROCESSED % 25 == 0 )); then
        log_info "Processed $PROCESSED/$PR_COUNT PRs..."
    fi
done <<< "$(echo "$PR_JSON" | jq -c '.[]')"

log_info "Processed $PROCESSED/$PR_COUNT PRs."

# =============================================================================
# Summary
# =============================================================================

echo ""
echo "=============================================="
echo "                  SUMMARY"
echo "=============================================="

print_category_summary "PRs with unsigned CLA" "${cla_flagged_prs[@]}"
print_category_summary "PRs with inactive authors" "${inactive_flagged_prs[@]}"

TOTAL_FLAGGED=$(( ${#cla_flagged_prs[@]} + ${#inactive_flagged_prs[@]} ))

if [[ $TOTAL_FLAGGED -eq 0 ]]; then
    echo ""
    echo "✅ No PRs found meeting the closure criteria."
else
    echo ""
    echo "⚠️  Total PRs to review for closure: $TOTAL_FLAGGED"
    echo ""
    echo "Remember: Leave a /close comment on PRs that should be closed."
    echo "Contributors can reopen PRs if they resume work."
    
    # Build URL arrays for each category (sorted oldest first)
    declare -a all_urls=()
    declare -a cla_urls=()
    declare -a inactive_urls=()
    declare -a lgtm_urls=()
    
    # Process CLA flagged PRs
    if [[ ${#cla_flagged_prs[@]} -gt 0 ]]; then
        while IFS= read -r entry; do
            IFS='|' read -r ts num url auth reason has_lgtm <<< "$entry"
            cla_urls+=("$url")
            all_urls+=("$url")
            if [[ "$has_lgtm" == "yes" ]]; then
                lgtm_urls+=("$url")
            fi
        done < <(printf '%s\n' "${cla_flagged_prs[@]}" | sort -t'|' -k1 -n)
    fi
    
    # Process inactive flagged PRs
    if [[ ${#inactive_flagged_prs[@]} -gt 0 ]]; then
        while IFS= read -r entry; do
            IFS='|' read -r ts num url auth reason has_lgtm <<< "$entry"
            inactive_urls+=("$url")
            all_urls+=("$url")
            if [[ "$has_lgtm" == "yes" ]]; then
                lgtm_urls+=("$url")
            fi
        done < <(printf '%s\n' "${inactive_flagged_prs[@]}" | sort -t'|' -k1 -n)
    fi
    
    # Deduplicate all_urls and lgtm_urls (a PR might be in both CLA and inactive)
    readarray -t all_urls < <(printf '%s\n' "${all_urls[@]}" | awk '!seen[$0]++')
    readarray -t lgtm_urls < <(printf '%s\n' "${lgtm_urls[@]}" | awk '!seen[$0]++')
    
    # Count for display
    cla_count=${#cla_urls[@]}
    inactive_count=${#inactive_urls[@]}
    lgtm_count=${#lgtm_urls[@]}
    all_count=${#all_urls[@]}
    
    # Interactive menu
    echo ""
    echo "Open PRs in Firefox (oldest first):"
    echo "------------------------------------"
    echo "  [a] ALL flagged PRs ($all_count)"
    
    if [[ $cla_count -gt 0 ]]; then
        echo "  [c] CLA unsigned PRs only ($cla_count)"
    else
        echo "  [c] CLA unsigned PRs only (none)"
    fi
    
    if [[ $inactive_count -gt 0 ]]; then
        echo "  [i] Inactive author PRs only ($inactive_count)"
    else
        echo "  [i] Inactive author PRs only (none)"
    fi
    
    if [[ $lgtm_count -gt 0 ]]; then
        echo "  [l] PRs with LGTM label only ($lgtm_count)"
    else
        echo "  [l] PRs with LGTM label only (none)"
    fi
    
    echo "  [n] Don't open any PRs"
    echo ""
    read -p "Your choice [a/c/i/l/n]: " response
    
    case "$response" in
        [aA])
            open_in_firefox "${all_urls[@]}"
            ;;
        [cC])
            if [[ $cla_count -gt 0 ]]; then
                open_in_firefox "${cla_urls[@]}"
            else
                echo "No CLA unsigned PRs to open."
            fi
            ;;
        [iI])
            if [[ $inactive_count -gt 0 ]]; then
                open_in_firefox "${inactive_urls[@]}"
            else
                echo "No inactive author PRs to open."
            fi
            ;;
        [lL])
            if [[ $lgtm_count -gt 0 ]]; then
                open_in_firefox "${lgtm_urls[@]}"
            else
                echo "No PRs with LGTM label to open."
            fi
            ;;
        *)
            echo "Skipping browser opening."
            ;;
    esac
fi

exit 0

Added check for CLA

Written with the help of ChatGPT.

Update assisted by Copilot & Claude Opus 4.5

Usage: ./pr-wrangler.sh [-r ] [-d ] [-c ] [-v]
-r Repository in the format owner/repo (default: kubernetes/website)
-d Inactivity period in days for no response (default: 14)
-c CLA unsigned period in days (default: 14)
-v Enable verbose output for troubleshooting
-h Show this help message

Criteria for closing PRs (per Kubernetes docs):

1. Author hasn't signed the CLA for two weeks
2. Author has not responded to comments or feedback in 2 or more weeks

Note: Authors listed in OWNERS_ALIASES are skipped (trusted contributors).