#!/usr/bin/env bash
# validate-test-quality.sh — Enforce anti-pattern regression thresholds for UI tests.
#
# Usage:
#   scripts/validate-test-quality.sh [baseline_file]
#   scripts/validate-test-quality.sh --help
#
# Behavior:
# - Loads metric baselines from JSON.
# - Counts pattern matches in configured roots via ripgrep.
# - Fails if any metric exceeds baseline + allowed_growth.
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
PROJECT_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
DEFAULT_BASELINE_FILE="$PROJECT_ROOT/audit/test-quality-baseline.json"
BASELINE_FILE="${1:-$DEFAULT_BASELINE_FILE}"

usage() {
    cat <<'EOF'
validate-test-quality.sh: enforce UI test anti-pattern regression thresholds.

Usage:
  scripts/validate-test-quality.sh [baseline_file]
  scripts/validate-test-quality.sh --help
EOF
}

if [[ "${1:-}" == "--help" || "${1:-}" == "-h" ]]; then
    usage
    exit 0
fi

RED='\033[0;31m'
GREEN='\033[0;32m'
NC='\033[0m'

FAILURES=0

fail() {
    echo -e "${RED}FAIL:${NC} $1"
    FAILURES=$((FAILURES + 1))
}

pass() {
    echo -e "${GREEN}PASS:${NC} $1"
}

require_cmd() {
    if ! command -v "$1" >/dev/null 2>&1; then
        fail "Missing required command: $1"
    fi
}

relative_path() {
    local path="$1"
    if [[ "$path" == "$PROJECT_ROOT/"* ]]; then
        echo "${path#$PROJECT_ROOT/}"
    else
        echo "$path"
    fi
}

count_matches() {
    local pattern="$1"
    local root="$2"
    local glob="$3"
    local output=""
    local status=0

    set +e
    output="$(rg --count-matches --no-messages --pcre2 -N -g "$glob" "$pattern" "$root" 2>/dev/null)"
    status=$?
    set -e

    if [[ "$status" -eq 1 ]]; then
        echo "0"
        return 0
    fi

    if [[ "$status" -ne 0 ]]; then
        return "$status"
    fi

    if [[ -z "$output" ]]; then
        echo "0"
        return 0
    fi

    echo "$output" | awk -F: '{sum += $NF} END {print sum + 0}'
}

count_matches_multiline() {
    local pattern="$1"
    local root="$2"
    local glob="$3"
    local output=""
    local status=0

    set +e
    output="$(rg --count-matches --no-messages --pcre2 -U -N -g "$glob" "$pattern" "$root" 2>/dev/null)"
    status=$?
    set -e

    if [[ "$status" -eq 1 ]]; then
        echo "0"
        return 0
    fi

    if [[ "$status" -ne 0 ]]; then
        return "$status"
    fi

    if [[ -z "$output" ]]; then
        echo "0"
        return 0
    fi

    echo "$output" | awk -F: '{sum += $NF} END {print sum + 0}'
}

list_metric_files() {
    local root="$1"
    local glob="$2"
    rg --files "$root" -g "$glob" 2>/dev/null || true
}

count_swift_test_body_pattern_matches() {
    local pattern="$1"
    local root="$2"
    local glob="$3"
    local files=()

    while IFS= read -r file_path; do
        [[ -n "$file_path" ]] && files+=("$file_path")
    done < <(list_metric_files "$root" "$glob")
    if [[ "${#files[@]}" -eq 0 ]]; then
        echo "0"
        return 0
    fi

    awk -v pattern="$pattern" '
        function update_depth(line,   i, c) {
            for (i = 1; i <= length(line); i++) {
                c = substr(line, i, 1)
                if (c == "{") depth++
                else if (c == "}") depth--
            }
        }

        /^[[:space:]]*func[[:space:]]+test[[:alnum:]_]+[[:space:]]*\(.*\)[[:space:]]*(throws)?[[:space:]]*\{/ {
            in_test = 1
            depth = 0
            update_depth($0)
            if ($0 ~ pattern) count++
            if (depth <= 0) {
                in_test = 0
                depth = 0
            }
            next
        }

        {
            if (!in_test) next

            if ($0 ~ pattern) count++
            update_depth($0)
            if (depth <= 0) {
                in_test = 0
                depth = 0
            }
        }

        END { print count + 0 }
    ' "${files[@]}"
}

count_swift_empty_test_bodies() {
    local root="$1"
    local glob="$2"
    local files=()

    while IFS= read -r file_path; do
        [[ -n "$file_path" ]] && files+=("$file_path")
    done < <(list_metric_files "$root" "$glob")
    if [[ "${#files[@]}" -eq 0 ]]; then
        echo "0"
        return 0
    fi

    awk '
        function update_depth(line,   i, c) {
            for (i = 1; i <= length(line); i++) {
                c = substr(line, i, 1)
                if (c == "{") depth++
                else if (c == "}") depth--
            }
        }

        function test_body_has_code(body,   cleaned) {
            cleaned = body
            gsub(/\/\/.*/, "", cleaned)
            gsub(/[ \t\r\n{}]/, "", cleaned)
            return cleaned != ""
        }

        /^[[:space:]]*func[[:space:]]+test[[:alnum:]_]+[[:space:]]*\(.*\)[[:space:]]*(throws)?[[:space:]]*\{/ {
            in_test = 1
            depth = 0
            body = ""
            update_depth($0)
            if (depth <= 0) {
                empty_count++
                in_test = 0
                depth = 0
                body = ""
            }
            next
        }

        {
            if (!in_test) next

            body = body $0 "\n"
            update_depth($0)

            if (depth <= 0) {
                if (!test_body_has_code(body)) {
                    empty_count++
                }
                in_test = 0
                depth = 0
                body = ""
            }
        }

        END { print empty_count + 0 }
    ' "${files[@]}"
}

count_metric() {
    local mode="$1"
    local pattern="$2"
    local root="$3"
    local glob="$4"

    case "$mode" in
        rg)
            count_matches "$pattern" "$root" "$glob"
            ;;
        rg_multiline)
            count_matches_multiline "$pattern" "$root" "$glob"
            ;;
        swift_test_body_pattern)
            count_swift_test_body_pattern_matches "$pattern" "$root" "$glob"
            ;;
        swift_empty_test_bodies)
            count_swift_empty_test_bodies "$root" "$glob"
            ;;
        *)
            echo "__INVALID_MODE__:$mode"
            return 0
            ;;
    esac
}

require_cmd jq
require_cmd rg
require_cmd awk

echo "=== Test Quality Gate ==="
echo "Baseline: $(relative_path "$BASELINE_FILE")"

if [[ ! -f "$BASELINE_FILE" ]]; then
    fail "Baseline file not found: $(relative_path "$BASELINE_FILE")"
    echo ""
    echo -e "${RED}Test quality gate failed with $FAILURES issue(s).${NC}"
    exit 1
fi

if jq -e '.metrics | type == "array" and length > 0' "$BASELINE_FILE" >/dev/null; then
    pass "Baseline includes metric definitions"
else
    fail "Baseline file has no metrics"
fi

while IFS= read -r metric; do
    [[ -z "$metric" ]] && continue

    metric_id="$(jq -r '.id // empty' <<<"$metric")"
    description="$(jq -r '.description // empty' <<<"$metric")"
    mode="$(jq -r '.mode // "rg"' <<<"$metric")"
    root_rel="$(jq -r '.root // empty' <<<"$metric")"
    glob="$(jq -r '.glob // empty' <<<"$metric")"
    pattern="$(jq -r '.pattern // ""' <<<"$metric")"
    baseline="$(jq -r '.baseline // empty' <<<"$metric")"
    allowed_growth="$(jq -r '.allowed_growth // empty' <<<"$metric")"

    if [[ -z "$metric_id" || -z "$root_rel" || -z "$glob" || -z "$baseline" || -z "$allowed_growth" ]]; then
        fail "Metric entry is missing required fields: $metric"
        continue
    fi

    if [[ "$mode" != "swift_empty_test_bodies" && -z "$pattern" ]]; then
        fail "Metric '$metric_id' requires non-empty pattern for mode '$mode'"
        continue
    fi

    if ! [[ "$baseline" =~ ^[0-9]+$ ]]; then
        fail "Metric '$metric_id' has non-numeric baseline: $baseline"
        continue
    fi

    if ! [[ "$allowed_growth" =~ ^[0-9]+$ ]]; then
        fail "Metric '$metric_id' has non-numeric allowed_growth: $allowed_growth"
        continue
    fi

    if [[ "$root_rel" == /* ]]; then
        root_path="$root_rel"
    else
        root_path="$PROJECT_ROOT/$root_rel"
    fi

    if [[ ! -d "$root_path" ]]; then
        fail "Metric '$metric_id' root directory not found: $(relative_path "$root_path")"
        continue
    fi

    current_count="0"
    if ! current_count="$(count_metric "$mode" "$pattern" "$root_path" "$glob")"; then
        fail "Metric '$metric_id' failed while counting matches"
        continue
    fi

    if [[ "$current_count" == __INVALID_MODE__:* ]]; then
        fail "Metric '$metric_id' uses unsupported mode '$mode'"
        continue
    fi

    max_allowed=$((baseline + allowed_growth))
    delta=$((current_count - baseline))

    if [[ "$current_count" -le "$max_allowed" ]]; then
        pass "$metric_id ($description): current=$current_count baseline=$baseline allowed_growth=$allowed_growth threshold=$max_allowed delta=$delta"
    else
        fail "$metric_id ($description): current=$current_count exceeds threshold=$max_allowed (baseline=$baseline allowed_growth=$allowed_growth delta=$delta)"
    fi
done < <(jq -c '.metrics[]' "$BASELINE_FILE")

echo ""
if [[ "$FAILURES" -eq 0 ]]; then
    echo -e "${GREEN}Test quality gate passed.${NC}"
    exit 0
fi

echo -e "${RED}Test quality gate failed with $FAILURES issue(s).${NC}"
exit 1