llm-intelligence/scripts/collector_stats_window_audit.sh

#!/usr/bin/env bash

set -euo pipefail

LIMIT=7
DB_URL="${DATABASE_URL:-}"
INPUT_PATH=""
THRESHOLD=""
FIELD_SEP=$'\x1f'

usage() {
    cat <<'EOF'
用法:
  bash scripts/collector_stats_window_audit.sh --db <DATABASE_URL> [--limit N] [--assert-success-rate PCT]
  bash scripts/collector_stats_window_audit.sh --input <tsv-file> [--limit N] [--assert-success-rate PCT]

输入 TSV 列顺序:
  source<TAB>success<TAB>error_message<TAB>created_at
EOF
}

classify_failure() {
    local message normalized
    message="${1:-}"
    normalized="$(printf '%s' "$message" | tr '[:upper:]' '[:lower:]')"

    if [[ -z "${normalized// }" ]]; then
        printf '%s\n' "collector_runtime_failure"
        return
    fi

    case "$normalized" in
        *"api key"*|*"strict real mode"*|*"database_url"*|*"password authentication failed"*|*"permission denied"*|*"role does not exist"*|*"relation does not exist"*|*"must provide"*)
            printf '%s\n' "precondition_missing"
            ;;
        *"429"*|*"rate limit"*|*"too many requests"*|*"timeout"*|*"temporarily unavailable"*|*"transport closed"*|*"connection reset"*|*"connection refused"*|*"eof"*|*"tls handshake timeout"*|*"no such host"*|*"i/o timeout"*)
            printf '%s\n' "external_provider_failure"
            ;;
        *)
            printf '%s\n' "collector_runtime_failure"
            ;;
    esac
}

fetch_rows_from_db() {
    if [[ -z "${DB_URL:-}" ]]; then
        echo "missing --db / DATABASE_URL" >&2
        return 1
    fi
    psql "$DB_URL" -F "$FIELD_SEP" -Atqc "
        SELECT
            COALESCE(source, ''),
            CASE WHEN success THEN 't' ELSE 'f' END,
            COALESCE(error_message, ''),
            TO_CHAR(created_at, 'YYYY-MM-DD HH24:MI:SS')
        FROM collector_stats
        ORDER BY created_at DESC
        LIMIT ${LIMIT};
    "
}

fetch_rows_from_file() {
    if [[ -z "${INPUT_PATH:-}" ]]; then
        echo "missing --input" >&2
        return 1
    fi
    head -n "$LIMIT" "$INPUT_PATH"
}

while [[ $# -gt 0 ]]; do
    case "$1" in
        --db)
            DB_URL="$2"
            shift 2
            ;;
        --input)
            INPUT_PATH="$2"
            shift 2
            ;;
        --limit)
            LIMIT="$2"
            shift 2
            ;;
        --assert-success-rate)
            THRESHOLD="$2"
            shift 2
            ;;
        --help|-h)
            usage
            exit 0
            ;;
        *)
            echo "unknown arg: $1" >&2
            usage >&2
            exit 1
            ;;
    esac
done

if [[ -n "$INPUT_PATH" ]]; then
    ROWS="$(fetch_rows_from_file)"
else
    ROWS="$(fetch_rows_from_db)"
fi

SUCCESS_COUNT=0
FAILURE_COUNT=0
PRECONDITION_COUNT=0
EXTERNAL_COUNT=0
RUNTIME_COUNT=0
UNKNOWN_COUNT=0
ROW_COUNT=0
DETAIL_LINES=""

while IFS= read -r raw_line; do
    [[ -z "${raw_line}" ]] && continue
    normalized_line="${raw_line//$'\t'/$FIELD_SEP}"
    IFS="$FIELD_SEP" read -r source success error_message created_at <<< "$normalized_line"
    ROW_COUNT=$((ROW_COUNT + 1))
    if [[ "$success" == "t" || "$success" == "true" ]]; then
        SUCCESS_COUNT=$((SUCCESS_COUNT + 1))
        category="success"
        rendered_error="-"
    else
        FAILURE_COUNT=$((FAILURE_COUNT + 1))
        category="$(classify_failure "$error_message")"
        rendered_error="${error_message:-unknown}"
        case "$category" in
            precondition_missing)
                PRECONDITION_COUNT=$((PRECONDITION_COUNT + 1))
                ;;
            external_provider_failure)
                EXTERNAL_COUNT=$((EXTERNAL_COUNT + 1))
                ;;
            collector_runtime_failure)
                RUNTIME_COUNT=$((RUNTIME_COUNT + 1))
                ;;
            *)
                UNKNOWN_COUNT=$((UNKNOWN_COUNT + 1))
                ;;
        esac
    fi
    DETAIL_LINES+=$'sample_'"${ROW_COUNT}"$' created_at='"${created_at:-unknown}"$' source='"${source:-unknown}"$' outcome='"$([[ "$category" == "success" ]] && printf '%s' "success" || printf '%s' "failure")"$' category='"${category}"$' error='"${rendered_error}"$'\n'
done <<< "$ROWS"

if [[ "$ROW_COUNT" -eq 0 ]]; then
    echo "window_size=0 success_count=0 failure_count=0 success_rate=0.00 threshold=${THRESHOLD:-n/a} precondition_missing=0 external_provider_failure=0 collector_runtime_failure=0 unknown_failure=0"
    echo "sample_window=empty"
    if [[ -n "$THRESHOLD" ]]; then
        exit 1
    fi
    exit 0
fi

SUCCESS_RATE="$(awk -v success="$SUCCESS_COUNT" -v total="$ROW_COUNT" 'BEGIN { printf "%.2f", (success * 100) / total }')"
echo "window_size=${ROW_COUNT} success_count=${SUCCESS_COUNT} failure_count=${FAILURE_COUNT} success_rate=${SUCCESS_RATE} threshold=${THRESHOLD:-n/a} precondition_missing=${PRECONDITION_COUNT} external_provider_failure=${EXTERNAL_COUNT} collector_runtime_failure=${RUNTIME_COUNT} unknown_failure=${UNKNOWN_COUNT}"
printf '%s' "$DETAIL_LINES"

if [[ -n "$THRESHOLD" ]]; then
    if awk -v actual="$SUCCESS_RATE" -v threshold="$THRESHOLD" 'BEGIN { exit !(actual >= threshold) }'; then
        exit 0
    fi
    exit 1
fi