Python save cookies with Selenium, Playwright and Requests in Netscape Format

python-save-cookies-in-netscape-format.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Python save cookies with Selenium, Playwright and Requests in Netscape Format\n",
    "\n",
    "- How to save cookies to txt file?\n",
    "- How to save cookies to Netscape format?\n",
    "- How to get cookies with Selenium, Playwright and Requests?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### All cookies are something like this(list[dict]):\n",
    "\n",
    "```json\n",
    "[\n",
    "    {\n",
    "        \"domain\": \"codigofacilito.com\",\n",
    "        \"expirationDate\": 1721236171.746433,\n",
    "        \"hostOnly\": true,\n",
    "        \"httpOnly\": false,\n",
    "        \"name\": \"ahoy_visitor\",\n",
    "        \"path\": \"/\",\n",
    "        \"sameSite\": \"unspecified\",\n",
    "        \"secure\": true,\n",
    "        \"session\": false,\n",
    "        \"storeId\": \"0\",\n",
    "        \"value\": \"18a3f5d0-fb23-442b-a7b9-de1f08a14e4d\"\n",
    "    },\n",
    "    {\n",
    "        \"domain\": \".codigofacilito.com\",\n",
    "        \"expirationDate\": 1707439819.231845,\n",
    "        \"hostOnly\": false,\n",
    "        \"httpOnly\": false,\n",
    "        \"name\": \"__stripe_mid\",\n",
    "        \"path\": \"/\",\n",
    "        \"sameSite\": \"strict\",\n",
    "        \"secure\": true,\n",
    "        \"session\": false,\n",
    "        \"storeId\": \"0\",\n",
    "        \"value\": \"8f3b3562-a465-4b58-a350-27c8a87dea7465ae39\"\n",
    "    },\n",
    "    {\n",
    "        \"domain\": \".codigofacilito.com\",\n",
    "        \"hostOnly\": false,\n",
    "        \"httpOnly\": false,\n",
    "        \"name\": \"CloudFront-Key-Pair-Id\",\n",
    "        \"path\": \"/\",\n",
    "        \"sameSite\": \"unspecified\",\n",
    "        \"secure\": true,\n",
    "        \"session\": true,\n",
    "        \"storeId\": \"0\",\n",
    "        \"value\": \"APKAIAHLS7PK3GEUR2RQ\"\n",
    "    }\n",
    "]\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def to_netscape_string(cookie_data: list[dict]) -> str:\n",
    "    \"\"\"\n",
    "    Convert cookies to Netscape cookie format.\n",
    "\n",
    "    This function takes a list of cookie dictionaries and transforms them into\n",
    "    a single string in Netscape cookie file format, which is commonly used by\n",
    "    web browsers and other HTTP clients for cookie storage. The Netscape string\n",
    "    can be used to programmatically interact with websites by simulating the\n",
    "    presence of cookies that might be set during normal web browsing.\n",
    "\n",
    "    Args:\n",
    "        cookie_data (list of dict): A list of dictionaries where each dictionary\n",
    "            represents a cookie. Each dictionary should have the following keys:\n",
    "            - 'domain': The domain of the cookie.\n",
    "            - 'expires': The expiration date of the cookie as a timestamp.\n",
    "            - 'path': The path for which the cookie is valid.\n",
    "            - 'secure': A boolean indicating if the cookie is secure.\n",
    "            - 'name': The name of the cookie.\n",
    "            - 'value': The value of the cookie.\n",
    "\n",
    "    Returns:\n",
    "        str: A string representing the cookie data in Netscape cookie file format.\n",
    "\n",
    "    Example of Netscape cookie file format:\n",
    "        .example.com\tTRUE\t/\tTRUE\t0\tCloudFront-Key-Pair-Id\tAPKAIAHLS7PK3GAUR2RQ\n",
    "    \"\"\"\n",
    "    result = []\n",
    "    for cookie in cookie_data:\n",
    "        domain = cookie.get(\"domain\", \"\")\n",
    "        expiration_date = cookie.get(\"expires\", 0)\n",
    "        path = cookie.get(\"path\", \"\")\n",
    "        secure = cookie.get(\"secure\", False)\n",
    "        name = cookie.get(\"name\", \"\")\n",
    "        value = cookie.get(\"value\", \"\")\n",
    "\n",
    "        include_sub_domain = domain.startswith(\".\") if domain else False\n",
    "        expiry = str(int(expiration_date)) if expiration_date > 0 else \"0\"\n",
    "        result.append(\n",
    "            [\n",
    "                domain,\n",
    "                str(include_sub_domain).upper(),\n",
    "                path,\n",
    "                str(secure).upper(),\n",
    "                expiry,\n",
    "                name,\n",
    "                value,\n",
    "            ]\n",
    "        )\n",
    "    return \"\\n\".join(\"\\t\".join(cookie_parts) for cookie_parts in result)\n",
    "\n",
    "\n",
    "def save_cookies_to_file(\n",
    "    cookie_data: list[dict], file_path='cookies.txt'\n",
    ") -> None:\n",
    "    \"\"\"\n",
    "    Save cookies to txt file\n",
    "    \"\"\"\n",
    "    netscape_string = to_netscape_string(cookie_data)\n",
    "    with open(file_path, \"w\", encoding=\"utf-8\") as file:\n",
    "\n",
    "        header = \"\"\"\\\n",
    "        # Netscape HTTP Cookie File\n",
    "        # http://www.netscape.com/newsref/std/cookie_spec.html\n",
    "        # This is a generated file!  Do not edit.\\n\n",
    "        \"\"\"\n",
    "        file.write(header)\n",
    "        file.write(netscape_string)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get cookies with Requests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "\n",
    "url = \"https://example.com\"\n",
    "response = requests.get(url)\n",
    "cookies = response.cookies\n",
    "# save cookies to file\n",
    "netscape_string = to_netscape_string(cookies)\n",
    "save_cookies_to_file(netscape_string, 'cookies.txt')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get cookies with Playwright"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from playwright.sync_api import sync_playwright\n",
    "\n",
    "with sync_playwright() as p:\n",
    "    browser = p.chromium.launch()\n",
    "    context = browser.new_context()\n",
    "    page = context.new_page()\n",
    "    url = \"https://example.com\"\n",
    "    page.goto(url)\n",
    "    cookies = context.cookies()\n",
    "    browser.close()\n",
    "\n",
    "    # save cookies to file\n",
    "    netscape_string = to_netscape_string(cookies)\n",
    "    save_cookies_to_file(netscape_string, 'cookies.txt')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Get cookies with Selenium"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from selenium import webdriver\n",
    "\n",
    "url = 'https://example.com'\n",
    "driver = webdriver.Chrome()\n",
    "driver.get(url)\n",
    "cookies = driver.get_cookies()\n",
    "# save cookies to file\n",
    "netscape_string = to_netscape_string(cookies)\n",
    "save_cookies_to_file(netscape_string, 'cookies.txt')"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}