#!/usr/bin/env bash # request.sh t2|t2-once [idx] set -euo pipefail unset HTTP_PROXY HTTPS_PROXY ALL_PROXY http_proxy https_proxy all_proxy NO_PROXY no_proxy ROOT="$(cd "$(dirname "$0")" && pwd)" SCHEME="trafilatura-http" TARGET_URL="${T2_URL:-https://www.gov.cn/zhengce/2021-03/13/content_5592681.htm}" idx="${2:-1}" mode="${1:-t2}" raw_dir="${RAW_DIR:-$ROOT/raw}" mkdir -p "$raw_dir/bodies" export BODY_OUT="${BODY_OUT:-$raw_dir/bodies/${mode}-${idx}.json}" export SCHEME TEMPLATE="T2" IDX="$idx" METHOD="POST" MAX_TIME="${MAX_TIME:-30}" export URL="http://127.0.0.1:18881/v1/read" export TARGET_URL export DATA DATA=$(TARGET_URL="$TARGET_URL" python3 -c "import json,os; print(json.dumps({'url': os.environ['TARGET_URL'], 'formats': ['markdown'], 'max_chars': 20000}, ensure_ascii=False))") export NOTE="${NOTE:-}" bash "$ROOT/lib/curl_time.sh"