path: root/bin/test

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

# Force GPG to use terminal-based pinentry (required for SSH sessions)
export GPG_TTY=$(tty)

# ----------------------------------------------------------------------------
# bin/test — verify a running Firecrawl deployment by hitting it from the host
#            (i.e. outside the Docker network), simulating an external client
#            such as an AI agent.
#
# Usage:
#   bin/test                     # local dev — hits http://127.0.0.1:31329
#   bin/test https://my.host     # remote   — hits the given base URL
#
# The API key is read from gopass at projects/firecrawl-dokploy/dev/api_key
# (override by exporting TEST_API_KEY before running).
# ----------------------------------------------------------------------------

BASE_URL="${1:-http://127.0.0.1:31329}"
BASE_URL="${BASE_URL%/}"

if [ -z "${TEST_API_KEY:-}" ]; then
  TEST_API_KEY="$(gopass show -o projects/firecrawl-dokploy/dev/api_key)"
fi

if [ -z "${TEST_API_KEY:-}" ]; then
  echo >&2 "ERROR: TEST_API_KEY is empty (gopass returned nothing and env was unset)"
  exit 1
fi

# ----------------------------------------------------------------------------
# Output capture: tee everything to tmp/test-<timestamp>.log AND the most recent
# run to tmp/test-latest.log, while still printing to the terminal.
# ----------------------------------------------------------------------------
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
REPO_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
TMP_DIR="$REPO_ROOT/tmp"
mkdir -p "$TMP_DIR"
LOG_FILE="$TMP_DIR/test-$(date +%Y%m%d-%H%M%S).log"
LATEST_LOG="$TMP_DIR/test-latest.log"

# Redirect all stdout+stderr through tee. Use a process substitution so the
# script keeps its normal exit code.
exec > >(tee "$LOG_FILE") 2>&1

# Mirror to test-latest.log on exit (atomic copy, not a symlink so it survives
# `rm tmp/*.log`).
trap 'cp -f "$LOG_FILE" "$LATEST_LOG" 2>/dev/null || true' EXIT

# ----------------------------------------------------------------------------
# Pretty output helpers
# ----------------------------------------------------------------------------
RED=$'\033[0;31m'
GREEN=$'\033[0;32m'
YELLOW=$'\033[0;33m'
BLUE=$'\033[0;34m'
BOLD=$'\033[1m'
RESET=$'\033[0m'

PASS=0
FAIL=0

section() {
  printf '\n%s== %s ==%s\n' "$BOLD$BLUE" "$1" "$RESET"
}

ok() {
  printf '  %s✓%s %s\n' "$GREEN" "$RESET" "$1"
  PASS=$((PASS + 1))
}

bad() {
  printf '  %s✗%s %s\n' "$RED" "$RESET" "$1"
  FAIL=$((FAIL + 1))
}

info() {
  printf '  %s…%s %s\n' "$YELLOW" "$RESET" "$1"
}

# ----------------------------------------------------------------------------
# Issue a JSON request, capture body + status
# Args: METHOD PATH [JSON_BODY]
# Sets: HTTP_STATUS, HTTP_BODY
# ----------------------------------------------------------------------------
http_call() {
  local method=$1
  local path=$2
  local body=${3:-}

  local tmp
  tmp=$(mktemp)

  local args=(
    -sS
    -o "$tmp"
    -w '%{http_code}'
    -X "$method"
    -H "Authorization: Bearer ${TEST_API_KEY}"
    -H "Content-Type: application/json"
    --connect-timeout 10
    --max-time 120
  )

  if [ -n "$body" ]; then
    args+=(-d "$body")
  fi

  HTTP_STATUS=$(curl "${args[@]}" "${BASE_URL}${path}" || echo "000")
  HTTP_BODY=$(cat "$tmp")
  rm -f "$tmp"
}

# ----------------------------------------------------------------------------
# Test runner
# ----------------------------------------------------------------------------
section "Target"
info "BASE_URL = ${BASE_URL}"
info "TEST_API_KEY = ${TEST_API_KEY:0:8}…"

# 1. Reachability ------------------------------------------------------------
section "1. Reachability"
http_call GET "/"
case "$HTTP_STATUS" in
  200|404|401)
    ok "API is reachable (HTTP $HTTP_STATUS at /)"
    ;;
  000)
    bad "Could not connect to ${BASE_URL} — is the stack running?"
    echo
    echo "Hint: run 'bin/up' first, or pass a different base URL."
    exit 1
    ;;
  *)
    bad "Unexpected HTTP $HTTP_STATUS at /"
    ;;
esac

# 2. Auth posture (informational) ------------------------------------------
# Self-hosted Firecrawl with USE_DB_AUTHENTICATION=false has no built-in
# bearer-token gate — any token (including a bogus one) is accepted by the
# API. We probe with a bogus token just to surface this fact in the log.
section "2. Auth posture (informational)"
tmp=$(mktemp)
status=$(curl -sS -o "$tmp" -w '%{http_code}' \
  -X POST "${BASE_URL}/v1/scrape" \
  -H "Authorization: Bearer fc-definitely-not-a-real-key" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com"}' \
  --connect-timeout 10 --max-time 60 || echo "000")
rm -f "$tmp"
if [ "$status" = "401" ] || [ "$status" = "403" ]; then
  ok "bogus key rejected with HTTP $status (USE_DB_AUTHENTICATION on?)"
else
  info "bogus key returned HTTP $status — self-hosted Firecrawl is open by design; restrict access at Traefik/firewall level"
fi

# 3. /v1/scrape --------------------------------------------------------------
section "3. Scrape — POST /v1/scrape https://example.com"
http_call POST "/v1/scrape" '{"url":"https://example.com","formats":["markdown"]}'
if [ "$HTTP_STATUS" = "200" ]; then
  if printf '%s' "$HTTP_BODY" | grep -qi "example domain"; then
    ok "scrape returned 200 and markdown contains 'Example Domain'"
  else
    bad "scrape returned 200 but markdown did not contain 'Example Domain'"
    echo "$HTTP_BODY" | head -c 400
    echo
  fi
else
  bad "scrape failed with HTTP $HTTP_STATUS"
  echo "$HTTP_BODY" | head -c 400
  echo
fi

# 4. /v1/search (covers SearXNG + Firecrawl scrape pipeline) -----------------
section "4. Search — POST /v1/search 'firecrawl github'"
http_call POST "/v1/search" '{"query":"firecrawl github","limit":3}'
if [ "$HTTP_STATUS" = "200" ]; then
  if printf '%s' "$HTTP_BODY" | grep -q '"success":true'; then
    ok "search returned 200 with success:true (SearXNG + scrape pipeline OK)"
  else
    bad "search returned 200 but body lacks success:true"
    echo "$HTTP_BODY" | head -c 400
    echo
  fi
else
  bad "search failed with HTTP $HTTP_STATUS"
  echo "$HTTP_BODY" | head -c 400
  echo
fi

# 5. /v1/map -----------------------------------------------------------------
section "5. Map — POST /v1/map https://example.com"
http_call POST "/v1/map" '{"url":"https://example.com"}'
if [ "$HTTP_STATUS" = "200" ]; then
  ok "map returned 200"
else
  bad "map failed with HTTP $HTTP_STATUS"
  echo "$HTTP_BODY" | head -c 400
  echo
fi

# ----------------------------------------------------------------------------
# Summary
# ----------------------------------------------------------------------------
echo
section "Summary"
printf "  %sPassed:%s %d\n" "$GREEN" "$RESET" "$PASS"
printf "  %sFailed:%s %d\n" "$RED" "$RESET" "$FAIL"
echo

if [ "$FAIL" -gt 0 ]; then
  echo "Log: $LOG_FILE"
  exit 1
fi

echo "Log: $LOG_FILE"