mind2web clean notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Benchmarking LaVague on Mind2Web"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will be benchmarking LaVague on Mind2Web, a dataset for agents."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/daniel/miniconda3/envs/lavague/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from lavague.defaults import default_get_driver\n",
    "import inspect\n",
    "\n",
    "get_driver = default_get_driver\n",
    "driver = get_driver()\n",
    "    \n",
    "# Gets the original source code of the get_driver method\n",
    "source_code = inspect.getsource(get_driver)\n",
    "\n",
    "# Split the source code into lines and remove the first line (method definition)\n",
    "source_code_lines = source_code.splitlines()[1:]\n",
    "source_code_lines = [line.strip() for line in source_code_lines[:-1]]\n",
    "import_lines = [line for line in source_code_lines if line.startswith(\"from\") or line.startswith(\"import\")] \n",
    "exec(\"\\n\".join(import_lines))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use a GPT-3.5 from Azure OpenAI"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from llama_index.llms.azure_openai import AzureOpenAI\n",
    "from lavague.prompts import DEFAULT_PROMPT\n",
    "\n",
    "api_key=os.getenv(\"AZURE_OPENAI_KEY\")\n",
    "api_version=\"2024-02-15-preview\"\n",
    "azure_endpoint = os.getenv(\"AZURE_OPENAI_ENDPOINT\")\n",
    "model = \"gpt-35-turbo\"\n",
    "deployment_name = \"gpt-35-turbo\"\n",
    "\n",
    "class LLM(AzureOpenAI):\n",
    "    def __init__(self):\n",
    "        super().__init__(\n",
    "            model=deployment_name,\n",
    "            deployment_name=deployment_name,\n",
    "            api_key=api_key,\n",
    "            azure_endpoint=azure_endpoint,\n",
    "            api_version=api_version,\n",
    "            temperature=0.0\n",
    "        )\n",
    "llm = LLM()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>split</th>\n",
       "      <th>annotation_uid</th>\n",
       "      <th>confirmed_task</th>\n",
       "      <th>raw_html</th>\n",
       "      <th>cleaned_html</th>\n",
       "      <th>action_uid</th>\n",
       "      <th>operation</th>\n",
       "      <th>code</th>\n",
       "      <th>cur_actions_desc</th>\n",
       "      <th>cur_actions_reprs</th>\n",
       "      <th>pos_candidates</th>\n",
       "      <th>prev_actions_desc</th>\n",
       "      <th>prev_actions_reprs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6033</td>\n",
       "      <td>test_task</td>\n",
       "      <td>640e0425-bceb-45ff-ba4d-dbc5b62e31d5</td>\n",
       "      <td>Find the \"Rock And Roll Over\" reviews</td>\n",
       "      <td>&lt;!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...</td>\n",
       "      <td>&lt;html backend_node_id=\"338\"&gt;\\n  &lt;body&gt;\\n    &lt;d...</td>\n",
       "      <td>96238fb3-bc46-4dff-95c0-4b2d4ecc70ea</td>\n",
       "      <td>{'op': 'TYPE', 'original_op': 'TYPE', 'value':...</td>\n",
       "      <td>```python\\nelement = driver.find_element(By.XP...</td>\n",
       "      <td>Enter \"Rock And Roll Over\" in the text box to ...</td>\n",
       "      <td>[textbox]  Enter artist name or song title -&gt; ...</td>\n",
       "      <td>[{'attributes': '{\"backend_node_id\": \"248\", \"b...</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3394</td>\n",
       "      <td>test_domain</td>\n",
       "      <td>34e0bf85-6441-40cb-b7f6-d107e5bcb049</td>\n",
       "      <td>Look up the visitors trend for Apple stock</td>\n",
       "      <td>&lt;!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...</td>\n",
       "      <td>&lt;html backend_node_id=\"14307\"&gt;\\n  &lt;div backend...</td>\n",
       "      <td>8043e8e8-bb7a-4521-80bf-246bc16e883c</td>\n",
       "      <td>{'op': 'CLICK', 'original_op': 'CLICK', 'value...</td>\n",
       "      <td>```python\\nelement = driver.find_element(By.XP...</td>\n",
       "      <td>Click on \"AAPL\" to look up the visitors trend ...</td>\n",
       "      <td>[div]  AAPL -&gt; CLICK</td>\n",
       "      <td>[{'attributes': '{\"backend_node_id\": \"14603\", ...</td>\n",
       "      <td>['Enter \"apple\" in the search box to look up t...</td>\n",
       "      <td>['[textbox]  Search for news, symbols or compa...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>320</td>\n",
       "      <td>test_domain</td>\n",
       "      <td>77269ea5-70a4-4cfa-a2f9-9937a1c55096</td>\n",
       "      <td>Search for early care and education programs f...</td>\n",
       "      <td>&lt;!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...</td>\n",
       "      <td>&lt;html backend_node_id=\"44287\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>b416d08b-90c4-43e1-b3d6-9d4a05e56a06</td>\n",
       "      <td>{'op': 'CLICK', 'original_op': 'CLICK', 'value...</td>\n",
       "      <td>```python\\nelement = driver.find_element(By.XP...</td>\n",
       "      <td>Select the After School Care checkbox.</td>\n",
       "      <td>[checkbox]  After School Care -&gt; CLICK</td>\n",
       "      <td>[{'attributes': '{\"backend_node_id\": \"44285\", ...</td>\n",
       "      <td>['Click on the \"services for RESIDENTS\" link.'...</td>\n",
       "      <td>['[link]  services for RESIDENTS -&gt; CLICK', '[...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6669</td>\n",
       "      <td>test_task</td>\n",
       "      <td>7f90a191-9dbe-478a-8ae2-8aa45b790158</td>\n",
       "      <td>Find more films from the director of Smile.</td>\n",
       "      <td>&lt;!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...</td>\n",
       "      <td>&lt;html backend_node_id=\"23177\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>fd713700-3876-44a8-80ab-4da898beab42</td>\n",
       "      <td>{'op': 'CLICK', 'original_op': 'CLICK', 'value...</td>\n",
       "      <td>```python\\nelement = driver.find_element(By.XP...</td>\n",
       "      <td>Click on \"Smile.\"</td>\n",
       "      <td>[div]  Smile -&gt; CLICK</td>\n",
       "      <td>[{'attributes': '{\"backend_node_id\": \"23642\", ...</td>\n",
       "      <td>['Search for \"Smile\" in the TV Shows and Movie...</td>\n",
       "      <td>['[textbox]  Search TV Shows and Movies... -&gt; ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>226</td>\n",
       "      <td>test_domain</td>\n",
       "      <td>332ed50d-4772-4eb3-9de9-27ff39abc161</td>\n",
       "      <td>Create a Fitness board.</td>\n",
       "      <td>&lt;!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...</td>\n",
       "      <td>&lt;html backend_node_id=\"123\"&gt;\\n  &lt;body backend_...</td>\n",
       "      <td>01342a5e-5cc3-45e7-bd9d-b522611fc7bf</td>\n",
       "      <td>{'op': 'CLICK', 'original_op': 'CLICK', 'value...</td>\n",
       "      <td>```python\\nelement = driver.find_element(By.XP...</td>\n",
       "      <td>Click on James Smith's profile image.</td>\n",
       "      <td>[img]  James Smith -&gt; CLICK</td>\n",
       "      <td>[{'attributes': '{\"backend_node_id\": \"117\", \"b...</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0        split                        annotation_uid  \\\n",
       "0        6033    test_task  640e0425-bceb-45ff-ba4d-dbc5b62e31d5   \n",
       "1        3394  test_domain  34e0bf85-6441-40cb-b7f6-d107e5bcb049   \n",
       "2         320  test_domain  77269ea5-70a4-4cfa-a2f9-9937a1c55096   \n",
       "3        6669    test_task  7f90a191-9dbe-478a-8ae2-8aa45b790158   \n",
       "4         226  test_domain  332ed50d-4772-4eb3-9de9-27ff39abc161   \n",
       "\n",
       "                                      confirmed_task  \\\n",
       "0              Find the \"Rock And Roll Over\" reviews   \n",
       "1         Look up the visitors trend for Apple stock   \n",
       "2  Search for early care and education programs f...   \n",
       "3        Find more films from the director of Smile.   \n",
       "4                            Create a Fitness board.   \n",
       "\n",
       "                                            raw_html  \\\n",
       "0  <!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...   \n",
       "1  <!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...   \n",
       "2  <!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...   \n",
       "3  <!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...   \n",
       "4  <!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Tr...   \n",
       "\n",
       "                                        cleaned_html  \\\n",
       "0  <html backend_node_id=\"338\">\\n  <body>\\n    <d...   \n",
       "1  <html backend_node_id=\"14307\">\\n  <div backend...   \n",
       "2  <html backend_node_id=\"44287\">\\n  <body>\\n    ...   \n",
       "3  <html backend_node_id=\"23177\">\\n  <body>\\n    ...   \n",
       "4  <html backend_node_id=\"123\">\\n  <body backend_...   \n",
       "\n",
       "                             action_uid  \\\n",
       "0  96238fb3-bc46-4dff-95c0-4b2d4ecc70ea   \n",
       "1  8043e8e8-bb7a-4521-80bf-246bc16e883c   \n",
       "2  b416d08b-90c4-43e1-b3d6-9d4a05e56a06   \n",
       "3  fd713700-3876-44a8-80ab-4da898beab42   \n",
       "4  01342a5e-5cc3-45e7-bd9d-b522611fc7bf   \n",
       "\n",
       "                                           operation  \\\n",
       "0  {'op': 'TYPE', 'original_op': 'TYPE', 'value':...   \n",
       "1  {'op': 'CLICK', 'original_op': 'CLICK', 'value...   \n",
       "2  {'op': 'CLICK', 'original_op': 'CLICK', 'value...   \n",
       "3  {'op': 'CLICK', 'original_op': 'CLICK', 'value...   \n",
       "4  {'op': 'CLICK', 'original_op': 'CLICK', 'value...   \n",
       "\n",
       "                                                code  \\\n",
       "0  ```python\\nelement = driver.find_element(By.XP...   \n",
       "1  ```python\\nelement = driver.find_element(By.XP...   \n",
       "2  ```python\\nelement = driver.find_element(By.XP...   \n",
       "3  ```python\\nelement = driver.find_element(By.XP...   \n",
       "4  ```python\\nelement = driver.find_element(By.XP...   \n",
       "\n",
       "                                    cur_actions_desc  \\\n",
       "0  Enter \"Rock And Roll Over\" in the text box to ...   \n",
       "1  Click on \"AAPL\" to look up the visitors trend ...   \n",
       "2             Select the After School Care checkbox.   \n",
       "3                                  Click on \"Smile.\"   \n",
       "4              Click on James Smith's profile image.   \n",
       "\n",
       "                                   cur_actions_reprs  \\\n",
       "0  [textbox]  Enter artist name or song title -> ...   \n",
       "1                               [div]  AAPL -> CLICK   \n",
       "2             [checkbox]  After School Care -> CLICK   \n",
       "3                              [div]  Smile -> CLICK   \n",
       "4                        [img]  James Smith -> CLICK   \n",
       "\n",
       "                                      pos_candidates  \\\n",
       "0  [{'attributes': '{\"backend_node_id\": \"248\", \"b...   \n",
       "1  [{'attributes': '{\"backend_node_id\": \"14603\", ...   \n",
       "2  [{'attributes': '{\"backend_node_id\": \"44285\", ...   \n",
       "3  [{'attributes': '{\"backend_node_id\": \"23642\", ...   \n",
       "4  [{'attributes': '{\"backend_node_id\": \"117\", \"b...   \n",
       "\n",
       "                                   prev_actions_desc  \\\n",
       "0                                               None   \n",
       "1  ['Enter \"apple\" in the search box to look up t...   \n",
       "2  ['Click on the \"services for RESIDENTS\" link.'...   \n",
       "3  ['Search for \"Smile\" in the TV Shows and Movie...   \n",
       "4                                               None   \n",
       "\n",
       "                                  prev_actions_reprs  \n",
       "0                                               None  \n",
       "1  ['[textbox]  Search for news, symbols or compa...  \n",
       "2  ['[link]  services for RESIDENTS -> CLICK', '[...  \n",
       "3  ['[textbox]  Search TV Shows and Movies... -> ...  \n",
       "4                                               None  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd \n",
    "from datasets import load_dataset\n",
    "\n",
    "dataset = load_dataset(\"BigAction/mind2web_clean\")\n",
    "\n",
    "df = dataset[\"train\"].to_pandas()\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will define some utility functions for this benchmark.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "import ast\n",
    "\n",
    "def keep_assignments(code_snippet):\n",
    "    # Regex to match variable assignments. This pattern assumes variable names are valid Python identifiers\n",
    "    # and captures typical assignment statements, excluding those that might appear in comments or strings.\n",
    "    pattern = r'^\\s*[a-zA-Z_][a-zA-Z0-9_]*\\s*=\\s*.+'\n",
    "\n",
    "    # Filter and keep only lines with variable assignments\n",
    "    assignments = [line for line in code_snippet.split('\\n') if re.match(pattern, line)]\n",
    "\n",
    "    # Join the filtered lines back into a string\n",
    "    return \"\\n\".join(assignments)\n",
    "\n",
    "# This function will be used to visit each node in the AST\n",
    "class VariableVisitor(ast.NodeVisitor):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.output = []\n",
    "    \n",
    "    def visit_Assign(self, node):\n",
    "        \n",
    "        # For each assignment, print the targets (variables)\n",
    "        for target in node.targets:\n",
    "            if isinstance(target, ast.Name):  # Ensure it's a variable assignment\n",
    "                self.output.append(target.id)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from bs4 import BeautifulSoup\n",
    "\n",
    "def beautify_html(html_string):\n",
    "    # Use BeautifulSoup to prettify the HTML\n",
    "    soup = BeautifulSoup(html_string, 'html.parser')\n",
    "    pretty_html_string = soup.prettify()\n",
    "    return pretty_html_string\n",
    "\n",
    "# Function to extract all 'backend_node_id' attributes\n",
    "def extract_backend_node_ids(html_content):\n",
    "    soup = BeautifulSoup(html_content, 'html.parser')\n",
    "    return set([tag['backend_node_id'] for tag in soup.find_all(attrs={\"backend_node_id\": True})])\n",
    "\n",
    "def intersection_score(set1, set2):\n",
    "    intersection_set = set1 & set2\n",
    "    intersection_len = len(intersection_set)\n",
    "    ratio1 = intersection_len / len(set1)\n",
    "    ratio2 = intersection_len / len(set2) \n",
    "    return max(ratio1, ratio2), min(ratio1, ratio2)\n",
    "\n",
    "def intersection_backend_node_id(ground_truth_outer_html, context_str):\n",
    "    ground_truth_ids = extract_backend_node_ids(ground_truth_outer_html)\n",
    "    context_ids = extract_backend_node_ids(context_str)\n",
    "    recall, precision = intersection_score(ground_truth_ids, context_ids)\n",
    "    return recall, precision\n",
    "\n",
    "def load_html(html, driver):\n",
    "    \"\"\"Loads a specific HTML content into the browser.\"\"\"\n",
    "    file_path = 'sample_page.html'\n",
    "\n",
    "    with open(file_path, 'w', encoding='utf-8') as file:\n",
    "        file.write(html)\n",
    "        \n",
    "    abs_file_path = os.path.abspath(\"sample_page.html\")\n",
    "\n",
    "    # Use the file:/// protocol to load the local HTML file\n",
    "    driver.get(f\"file:///{abs_file_path}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "config.json: 100%|██████████| 743/743 [00:00<00:00, 4.57MB/s]\n",
      "model.safetensors: 100%|██████████| 133M/133M [00:00<00:00, 494MB/s] \n",
      "tokenizer_config.json: 100%|██████████| 366/366 [00:00<00:00, 2.34MB/s]\n",
      "vocab.txt: 100%|██████████| 232k/232k [00:00<00:00, 41.4MB/s]\n",
      "tokenizer.json: 100%|██████████| 711k/711k [00:00<00:00, 35.5MB/s]\n",
      "special_tokens_map.json: 100%|██████████| 125/125 [00:00<00:00, 871kB/s]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "from langchain_text_splitters import RecursiveCharacterTextSplitter\n",
    "from llama_index.core.node_parser import LangchainNodeParser, CodeSplitter\n",
    "from llama_index.retrievers.bm25 import BM25Retriever\n",
    "from llama_index.core import VectorStoreIndex, Document\n",
    "from lavague.defaults import DefaultEmbedder\n",
    "\n",
    "def get_retriever_recursive(embed, html):\n",
    "    \n",
    "    K = 3\n",
    "    \n",
    "    text_list = [html]\n",
    "    documents = [Document(text=t) for t in text_list]\n",
    "    \n",
    "    splitter = LangchainNodeParser(lc_splitter=RecursiveCharacterTextSplitter.from_language(\n",
    "        language=\"html\",\n",
    "    ))\n",
    "    \n",
    "    nodes = splitter.get_nodes_from_documents(documents)\n",
    "    nodes = [node for node in nodes if node.text]\n",
    "\n",
    "    index = VectorStoreIndex(nodes, embed_model=embed)\n",
    "    retriever_recursive = BM25Retriever.from_defaults(index = index, similarity_top_k=K)\n",
    "    return retriever_recursive\n",
    "\n",
    "def get_retriever_code(embed, html):\n",
    "    \n",
    "    K = 3\n",
    "    \n",
    "    text_list = [html]\n",
    "    documents = [Document(text=t) for t in text_list]\n",
    "    \n",
    "    splitter = CodeSplitter(\n",
    "        language=\"html\",\n",
    "        chunk_lines=50,  # lines per chunk\n",
    "        chunk_lines_overlap=15,  # lines overlap between chunks\n",
    "        max_chars=2000,  # max chars per chunk\n",
    "    )\n",
    "    \n",
    "    nodes = splitter.get_nodes_from_documents(documents)\n",
    "    nodes = [node for node in nodes if node.text]\n",
    "\n",
    "    index = VectorStoreIndex(nodes, embed_model=embed)\n",
    "    retriever_code = BM25Retriever.from_defaults(index = index, similarity_top_k=K)\n",
    "    return retriever_code\n",
    "\n",
    "def get_nodes(query, html, get_retriever, embedder):    \n",
    "    retriever = get_retriever(embedder, html)\n",
    "    source_nodes = retriever.retrieve(query)\n",
    "    source_nodes = [node.text for node in source_nodes]\n",
    "    return source_nodes\n",
    "\n",
    "embedder = DefaultEmbedder()\n",
    "\n",
    "import re\n",
    "\n",
    "def extract_first_python_code(markdown_text: str):\n",
    "    # Pattern to match the first ```python ``` code block\n",
    "    pattern = r\"```python(.*?)```\"\n",
    "\n",
    "    # Using re.DOTALL to make '.' match also newlines\n",
    "    match = re.search(pattern, markdown_text, re.DOTALL)\n",
    "    if match:\n",
    "        # Return the first matched group, which is the code inside the ```python ```\n",
    "        return match.group(1).strip()\n",
    "    else:\n",
    "        # Return None if no match is found\n",
    "        return None\n",
    "\n",
    "decontaminate_html = lambda x: re.sub(r' backend_node_id=\"\\d+\"', '', x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "192it [06:13,  1.94s/it]\n"
     ]
    },
    {
     "ename": "AttributeError",
     "evalue": "'NoneType' object has no attribute 'split'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[47], line 42\u001b[0m\n\u001b[1;32m     39\u001b[0m generated_code \u001b[38;5;241m=\u001b[39m extract_first_python_code(response)\n\u001b[1;32m     41\u001b[0m \u001b[38;5;66;03m# Keep only the variable assignments in the generated code\u001b[39;00m\n\u001b[0;32m---> 42\u001b[0m code \u001b[38;5;241m=\u001b[39m \u001b[43mkeep_assignments\u001b[49m\u001b[43m(\u001b[49m\u001b[43mgenerated_code\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     44\u001b[0m \u001b[38;5;66;03m# Split the code into lines and keep only the first assignment\u001b[39;00m\n\u001b[1;32m     45\u001b[0m code \u001b[38;5;241m=\u001b[39m code\u001b[38;5;241m.\u001b[39msplit(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m\"\u001b[39m)[\u001b[38;5;241m0\u001b[39m]\n",
      "Cell \u001b[0;32mIn[5], line 10\u001b[0m, in \u001b[0;36mkeep_assignments\u001b[0;34m(code_snippet)\u001b[0m\n\u001b[1;32m      7\u001b[0m pattern \u001b[38;5;241m=\u001b[39m \u001b[38;5;124mr\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m^\u001b[39m\u001b[38;5;124m\\\u001b[39m\u001b[38;5;124ms*[a-zA-Z_][a-zA-Z0-9_]*\u001b[39m\u001b[38;5;124m\\\u001b[39m\u001b[38;5;124ms*=\u001b[39m\u001b[38;5;124m\\\u001b[39m\u001b[38;5;124ms*.+\u001b[39m\u001b[38;5;124m'\u001b[39m\n\u001b[1;32m      9\u001b[0m \u001b[38;5;66;03m# Filter and keep only lines with variable assignments\u001b[39;00m\n\u001b[0;32m---> 10\u001b[0m assignments \u001b[38;5;241m=\u001b[39m [line \u001b[38;5;28;01mfor\u001b[39;00m line \u001b[38;5;129;01min\u001b[39;00m \u001b[43mcode_snippet\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43msplit\u001b[49m(\u001b[38;5;124m'\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m'\u001b[39m) \u001b[38;5;28;01mif\u001b[39;00m re\u001b[38;5;241m.\u001b[39mmatch(pattern, line)]\n\u001b[1;32m     12\u001b[0m \u001b[38;5;66;03m# Join the filtered lines back into a string\u001b[39;00m\n\u001b[1;32m     13\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;241m.\u001b[39mjoin(assignments)\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'NoneType' object has no attribute 'split'"
     ]
    }
   ],
   "source": [
    "from tqdm import tqdm\n",
    "\n",
    "outputs = []\n",
    "\n",
    "get_retriever = get_retriever_recursive\n",
    "retriever_name = get_retriever.__name__\n",
    "\n",
    "for index, row in tqdm(df.iterrows()):\n",
    "    \n",
    "    if row[\"action_uid\"] in results[\"action_uid\"].values:\n",
    "        continue\n",
    "    \n",
    "    ground_truth_code = row[\"code\"]\n",
    "    html = row[\"cleaned_html\"]\n",
    "    query = row[\"cur_actions_desc\"]\n",
    "    action_uid = row[\"action_uid\"]\n",
    "\n",
    "    backend_node_id = ground_truth_code.split(\"backend_node_id\")[1].split(\"']\")[0].replace(\"='\",\"\")\n",
    "\n",
    "    load_html(html, driver)\n",
    "\n",
    "    source_nodes = get_nodes(query, html, get_retriever, embedder)\n",
    "    backend_node_id = ground_truth_code.split(\"backend_node_id\")[1].split(\"']\")[0].replace(\"='\",\"\")\n",
    "\n",
    "    ground_truth_element = driver.find_element(By.XPATH, f'//*[@backend_node_id=\"{backend_node_id}\"]')\n",
    "    ground_truth_outer_html = driver.execute_script(\"return arguments[0].outerHTML;\", ground_truth_element)\n",
    "    \n",
    "    context_str = \"\\n\".join(source_nodes)\n",
    "    recall_retriever, precision_retriever = intersection_backend_node_id(ground_truth_outer_html, context_str)\n",
    "    \n",
    "    # We remove the backend node ids to ensure the LLM does not use them\n",
    "    decontaminated_context_str = decontaminate_html(context_str)\n",
    "    prompt = DEFAULT_PROMPT.format(context_str=decontaminated_context_str, query_str=query)\n",
    "\n",
    "    response = llm.complete(prompt).text\n",
    "\n",
    "    generated_code = extract_first_python_code(response)\n",
    "\n",
    "    # Keep only the variable assignments in the generated code\n",
    "    code = keep_assignments(generated_code)\n",
    "\n",
    "    # Split the code into lines and keep only the first assignment\n",
    "    code = code.split(\"\\n\")[0]\n",
    "    parsed_code = ast.parse(code)\n",
    "\n",
    "    recall_llm, precision_llm = None, None\n",
    "    execution_error = \"\"\n",
    "\n",
    "    try:\n",
    "    # Create a visitor instance and use it to visit the nodes in the parsed AST\n",
    "        visitor = VariableVisitor()\n",
    "        visitor.visit(parsed_code)\n",
    "        variable_name = visitor.output[0]\n",
    "\n",
    "        # Execute the code to define the first variable\n",
    "        exec(code)\n",
    "\n",
    "        # Assign the variable to the target_element variable which will be used afterwards to compute score\n",
    "        target_element = None\n",
    "        exec(f\"\"\"target_element = {variable_name}\"\"\")\n",
    "\n",
    "        target_outer_html = driver.execute_script(\"return arguments[0].outerHTML;\", target_element)\n",
    "        recall_llm, precision_llm = intersection_backend_node_id(ground_truth_outer_html, target_outer_html)\n",
    "    except Exception as e:\n",
    "        target_outer_html = \"\"\n",
    "        execution_error = str(e)\n",
    "\n",
    "    output = {\n",
    "        \"query\": query,\n",
    "        \"recall_retriever\": recall_retriever,\n",
    "        \"precision_retriever\": precision_retriever,\n",
    "        \"recall_llm\": recall_llm,\n",
    "        \"precision_llm\": precision_llm,\n",
    "        \"html\": html,\n",
    "        \"action_uid\": action_uid,\n",
    "        \"source_nodes\": (\"-\"*10).join(source_nodes),\n",
    "        \"ground_truth_code\": f\"\"\"ground_truth_element = driver.find_element(By.XPATH, f'//*[@backend_node_id=\"{backend_node_id}\"]')\"\"\",\n",
    "        \"ground_truth_outer_html\": ground_truth_outer_html,\n",
    "        \"generated_code\": generated_code,\n",
    "        \"target_outer_html\": target_outer_html,\n",
    "        \"execution_error\": execution_error,\n",
    "        \"retriever\": retriever_name,\n",
    "        \"model\": model,\n",
    "    }\n",
    "    outputs.append(output)\n",
    "    \n",
    "    results = pd.DataFrame(outputs)\n",
    "\n",
    "    results.to_csv(f\"benchmark_retriever_{retriever_name}_model_{model}.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = pd.read_csv(\"benchmark_retriever_get_retriever_recursive_model_gpt-35-turbo.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>query</th>\n",
       "      <th>recall_retriever</th>\n",
       "      <th>precision_retriever</th>\n",
       "      <th>recall_llm</th>\n",
       "      <th>precision_llm</th>\n",
       "      <th>html</th>\n",
       "      <th>action_uid</th>\n",
       "      <th>source_nodes</th>\n",
       "      <th>ground_truth_code</th>\n",
       "      <th>ground_truth_outer_html</th>\n",
       "      <th>generated_code</th>\n",
       "      <th>target_outer_html</th>\n",
       "      <th>execution_error</th>\n",
       "      <th>retriever</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Enter \"Rock And Roll Over\" in the text box to ...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"338\"&gt;\\n  &lt;body&gt;\\n    &lt;d...</td>\n",
       "      <td>96238fb3-bc46-4dff-95c0-4b2d4ecc70ea</td>\n",
       "      <td>&lt;div backend_node_id=\"2662\"&gt;\\n                ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;input backend_node_id=\"248\" type=\"text\" name=...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"2663\"&gt;Hi ! I have a Iba...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Click on \"AAPL\" to look up the visitors trend ...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.114943</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>&lt;html backend_node_id=\"14307\"&gt;\\n  &lt;div backend...</td>\n",
       "      <td>8043e8e8-bb7a-4521-80bf-246bc16e883c</td>\n",
       "      <td>&lt;div backend_node_id=\"14708\"&gt;\\n               ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;li backend_node_id=\"14603\" role=\"option\" titl...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"14606\"&gt;AAPL&lt;/text&gt;</td>\n",
       "      <td>NaN</td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Select the After School Care checkbox.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.008772</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"44287\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>b416d08b-90c4-43e1-b3d6-9d4a05e56a06</td>\n",
       "      <td>&lt;span backend_node_id=\"44871\"&gt;\\n              ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;input backend_node_id=\"44285\" type=\"checkbox\"...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;input backend_node_id=\"44285\" type=\"checkbox\"...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Click on \"Smile.\"</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.050000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>&lt;html backend_node_id=\"23177\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>fd713700-3876-44a8-80ab-4da898beab42</td>\n",
       "      <td>&lt;div backend_node_id=\"23101\"&gt;\\n               ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;li backend_node_id=\"23642\" role=\"option\"&gt;\\n  ...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"23648\"&gt;Smile&lt;/text&gt;</td>\n",
       "      <td>NaN</td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Click on James Smith's profile image.</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.011236</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"123\"&gt;\\n  &lt;body backend_...</td>\n",
       "      <td>01342a5e-5cc3-45e7-bd9d-b522611fc7bf</td>\n",
       "      <td>&lt;body backend_node_id=\"187\"&gt;\\n    &lt;div backend...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;img backend_node_id=\"117\" alt=\"James Smith\"&gt;</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;img backend_node_id=\"117\" alt=\"James Smith\"&gt;</td>\n",
       "      <td>NaN</td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               query  recall_retriever  \\\n",
       "0  Enter \"Rock And Roll Over\" in the text box to ...               0.0   \n",
       "1  Click on \"AAPL\" to look up the visitors trend ...               1.0   \n",
       "2             Select the After School Care checkbox.               1.0   \n",
       "3                                  Click on \"Smile.\"               1.0   \n",
       "4              Click on James Smith's profile image.               1.0   \n",
       "\n",
       "   precision_retriever  recall_llm  precision_llm  \\\n",
       "0             0.000000         0.0       0.000000   \n",
       "1             0.114943         1.0       0.100000   \n",
       "2             0.008772         1.0       1.000000   \n",
       "3             0.050000         1.0       0.333333   \n",
       "4             0.011236         1.0       1.000000   \n",
       "\n",
       "                                                html  \\\n",
       "0  <html backend_node_id=\"338\">\\n  <body>\\n    <d...   \n",
       "1  <html backend_node_id=\"14307\">\\n  <div backend...   \n",
       "2  <html backend_node_id=\"44287\">\\n  <body>\\n    ...   \n",
       "3  <html backend_node_id=\"23177\">\\n  <body>\\n    ...   \n",
       "4  <html backend_node_id=\"123\">\\n  <body backend_...   \n",
       "\n",
       "                             action_uid  \\\n",
       "0  96238fb3-bc46-4dff-95c0-4b2d4ecc70ea   \n",
       "1  8043e8e8-bb7a-4521-80bf-246bc16e883c   \n",
       "2  b416d08b-90c4-43e1-b3d6-9d4a05e56a06   \n",
       "3  fd713700-3876-44a8-80ab-4da898beab42   \n",
       "4  01342a5e-5cc3-45e7-bd9d-b522611fc7bf   \n",
       "\n",
       "                                        source_nodes  \\\n",
       "0  <div backend_node_id=\"2662\">\\n                ...   \n",
       "1  <div backend_node_id=\"14708\">\\n               ...   \n",
       "2  <span backend_node_id=\"44871\">\\n              ...   \n",
       "3  <div backend_node_id=\"23101\">\\n               ...   \n",
       "4  <body backend_node_id=\"187\">\\n    <div backend...   \n",
       "\n",
       "                                   ground_truth_code  \\\n",
       "0  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "1  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "2  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "3  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "4  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "\n",
       "                             ground_truth_outer_html  \\\n",
       "0  <input backend_node_id=\"248\" type=\"text\" name=...   \n",
       "1  <li backend_node_id=\"14603\" role=\"option\" titl...   \n",
       "2  <input backend_node_id=\"44285\" type=\"checkbox\"...   \n",
       "3  <li backend_node_id=\"23642\" role=\"option\">\\n  ...   \n",
       "4      <img backend_node_id=\"117\" alt=\"James Smith\">   \n",
       "\n",
       "                                      generated_code  \\\n",
       "0  # Let's proceed step by step.\\n# First we need...   \n",
       "1  # Let's proceed step by step.\\n# First we need...   \n",
       "2  # Let's proceed step by step.\\n# First we need...   \n",
       "3  # Let's proceed step by step.\\n# First we need...   \n",
       "4  # Let's proceed step by step.\\n# First we need...   \n",
       "\n",
       "                                   target_outer_html execution_error  \\\n",
       "0  <text backend_node_id=\"2663\">Hi ! I have a Iba...             NaN   \n",
       "1          <text backend_node_id=\"14606\">AAPL</text>             NaN   \n",
       "2  <input backend_node_id=\"44285\" type=\"checkbox\"...             NaN   \n",
       "3         <text backend_node_id=\"23648\">Smile</text>             NaN   \n",
       "4      <img backend_node_id=\"117\" alt=\"James Smith\">             NaN   \n",
       "\n",
       "                 retriever  model  \n",
       "0  get_retriever_recursive    NaN  \n",
       "1  get_retriever_recursive    NaN  \n",
       "2  get_retriever_recursive    NaN  \n",
       "3  get_retriever_recursive    NaN  \n",
       "4  get_retriever_recursive    NaN  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "x = results['recall_retriever']\n",
    "y = results['recall_llm']\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Filter results on recall retriever == 1.0\n",
    "filtered_results = results[results['recall_retriever'] == 1.0]\n",
    "\n",
    "# Plot distribution of recall_llm\n",
    "plt.hist(filtered_results['recall_llm'], bins=10)\n",
    "plt.xlabel('recall_llm')\n",
    "plt.ylabel('Frequency')\n",
    "plt.title('Distribution of recall_llm when retriever has perfect recall')\n",
    "plt.show()\n",
    "\n",
    "filtered_results = results[results['recall_retriever'] < 1.0]\n",
    "\n",
    "# Plot distribution of recall_llm\n",
    "plt.hist(filtered_results['recall_llm'], bins=10)\n",
    "plt.xlabel('recall_llm')\n",
    "plt.ylabel('Frequency')\n",
    "plt.title('Distribution of recall_llm when retriever has imperfect recall')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Retriever performance:  0.8177083333333334\n",
      "LLM performance:  0.5572916666666666\n",
      "Deperdition rate:  0.6815286624203821\n"
     ]
    }
   ],
   "source": [
    "print(\"Retriever performance: \", (results.recall_retriever == 1.0).mean())\n",
    "print(\"LLM performance: \", (results.recall_llm == 1.0).mean())\n",
    "print(\"Deperdition rate: \", (results.recall_llm == 1.0).mean() / (results.recall_retriever == 1.0).mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6242038216560509"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "perfect_retriever = results.loc[results.recall_retriever == 1.0]\n",
    "(perfect_retriever.recall_llm == 1.0).mean()\n",
    "# perfect_retriever = perfect_retriever.loc[~perfect_retriever.recall_llm.isna()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>query</th>\n",
       "      <th>recall_retriever</th>\n",
       "      <th>precision_retriever</th>\n",
       "      <th>recall_llm</th>\n",
       "      <th>precision_llm</th>\n",
       "      <th>html</th>\n",
       "      <th>action_uid</th>\n",
       "      <th>source_nodes</th>\n",
       "      <th>ground_truth_code</th>\n",
       "      <th>ground_truth_outer_html</th>\n",
       "      <th>generated_code</th>\n",
       "      <th>target_outer_html</th>\n",
       "      <th>execution_error</th>\n",
       "      <th>retriever</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Enter \"Rock And Roll Over\" in the text box to ...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"338\"&gt;\\n  &lt;body&gt;\\n    &lt;d...</td>\n",
       "      <td>96238fb3-bc46-4dff-95c0-4b2d4ecc70ea</td>\n",
       "      <td>&lt;div backend_node_id=\"2662\"&gt;\\n                ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;input backend_node_id=\"248\" type=\"text\" name=...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"2663\"&gt;Hi ! I have a Iba...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Click on \"AAPL\" to look up the visitors trend ...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.114943</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.100000</td>\n",
       "      <td>&lt;html backend_node_id=\"14307\"&gt;\\n  &lt;div backend...</td>\n",
       "      <td>8043e8e8-bb7a-4521-80bf-246bc16e883c</td>\n",
       "      <td>&lt;div backend_node_id=\"14708\"&gt;\\n               ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;li backend_node_id=\"14603\" role=\"option\" titl...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"14606\"&gt;AAPL&lt;/text&gt;</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Select the After School Care checkbox.</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.008772</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"44287\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>b416d08b-90c4-43e1-b3d6-9d4a05e56a06</td>\n",
       "      <td>&lt;span backend_node_id=\"44871\"&gt;\\n              ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;input backend_node_id=\"44285\" type=\"checkbox\"...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;input backend_node_id=\"44285\" type=\"checkbox\"...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Click on \"Smile.\"</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.050000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>&lt;html backend_node_id=\"23177\"&gt;\\n  &lt;body&gt;\\n    ...</td>\n",
       "      <td>fd713700-3876-44a8-80ab-4da898beab42</td>\n",
       "      <td>&lt;div backend_node_id=\"23101\"&gt;\\n               ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;li backend_node_id=\"23642\" role=\"option\"&gt;\\n  ...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"23648\"&gt;Smile&lt;/text&gt;</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Click on James Smith's profile image.</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.011236</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"123\"&gt;\\n  &lt;body backend_...</td>\n",
       "      <td>01342a5e-5cc3-45e7-bd9d-b522611fc7bf</td>\n",
       "      <td>&lt;body backend_node_id=\"187\"&gt;\\n    &lt;div backend...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;img backend_node_id=\"117\" alt=\"James Smith\"&gt;</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;img backend_node_id=\"117\" alt=\"James Smith\"&gt;</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>Click on the search option to get the best rat...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"15457\"&gt;\\n  &lt;body backen...</td>\n",
       "      <td>5036b341-d933-4f21-9bde-af116b2f78e2</td>\n",
       "      <td>&lt;div backend_node_id=\"16067\"&gt;\\n               ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;svg backend_node_id=\"15442\" class=\"icon\"&gt;&lt;/svg&gt;</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;form backend_node_id=\"15635\" name=\"search-for...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>188</th>\n",
       "      <td>Click on the \"Benefits and financial support f...</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.016260</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>&lt;html backend_node_id=\"4953\"&gt;\\n  &lt;body&gt;\\n    &lt;...</td>\n",
       "      <td>bf65fe94-df6e-4817-8530-93bf6b9800a3</td>\n",
       "      <td>&lt;div backend_node_id=\"5365\"&gt;\\n              &lt;f...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;a backend_node_id=\"4947\"&gt;\\n                  ...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"5568\"&gt;Benefits and fina...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>Click on the link labeled \"Thrill Rides.\"</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"64176\"&gt;\\n  &lt;body backen...</td>\n",
       "      <td>c7cd25d6-0721-488d-9a01-975f6b1bc76d</td>\n",
       "      <td>&lt;p backend_node_id=\"67582\"&gt;\\n                 ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;a backend_node_id=\"64159\"&gt;\\n                 ...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;text backend_node_id=\"66373\"&gt;Thrill Rides&lt;/text&gt;</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>Enter \"Ohio\" in the search box for the location.</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.012346</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>&lt;html backend_node_id=\"1531\"&gt;\\n  &lt;body&gt;\\n    &lt;...</td>\n",
       "      <td>370ca9f4-e6ee-4576-bcf1-ea418704debf</td>\n",
       "      <td>&lt;div backend_node_id=\"1948\"&gt;\\n                ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;input backend_node_id=\"1528\" type=\"text\" name...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;input backend_node_id=\"1528\" type=\"text\" name...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191</th>\n",
       "      <td>Click on the \"Price\" heading.</td>\n",
       "      <td>0.916667</td>\n",
       "      <td>0.277311</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.055556</td>\n",
       "      <td>&lt;html backend_node_id=\"329709\"&gt;\\n  &lt;body backe...</td>\n",
       "      <td>3bb445bc-1794-4365-8d2e-7eaea29db4e3</td>\n",
       "      <td>&lt;div backend_node_id=\"344655\"&gt;\\n              ...</td>\n",
       "      <td>driver.find_element(By.XPATH, f'//*[@backend_n...</td>\n",
       "      <td>&lt;li backend_node_id=\"335800\"&gt;\\n               ...</td>\n",
       "      <td># Let's proceed step by step.\\n# First we need...</td>\n",
       "      <td>&lt;h2 backend_node_id=\"329691\" title=\"Expand Pri...</td>\n",
       "      <td></td>\n",
       "      <td>get_retriever_recursive</td>\n",
       "      <td>gpt-35-turbo</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>192 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                 query  recall_retriever  \\\n",
       "0    Enter \"Rock And Roll Over\" in the text box to ...          0.000000   \n",
       "1    Click on \"AAPL\" to look up the visitors trend ...          1.000000   \n",
       "2               Select the After School Care checkbox.          1.000000   \n",
       "3                                    Click on \"Smile.\"          1.000000   \n",
       "4                Click on James Smith's profile image.          1.000000   \n",
       "..                                                 ...               ...   \n",
       "187  Click on the search option to get the best rat...          0.000000   \n",
       "188  Click on the \"Benefits and financial support f...          1.000000   \n",
       "189          Click on the link labeled \"Thrill Rides.\"          0.000000   \n",
       "190   Enter \"Ohio\" in the search box for the location.          1.000000   \n",
       "191                      Click on the \"Price\" heading.          0.916667   \n",
       "\n",
       "     precision_retriever  recall_llm  precision_llm  \\\n",
       "0               0.000000         0.0       0.000000   \n",
       "1               0.114943         1.0       0.100000   \n",
       "2               0.008772         1.0       1.000000   \n",
       "3               0.050000         1.0       0.333333   \n",
       "4               0.011236         1.0       1.000000   \n",
       "..                   ...         ...            ...   \n",
       "187             0.000000         0.0       0.000000   \n",
       "188             0.016260         1.0       0.500000   \n",
       "189             0.000000         0.0       0.000000   \n",
       "190             0.012346         1.0       1.000000   \n",
       "191             0.277311         1.0       0.055556   \n",
       "\n",
       "                                                  html  \\\n",
       "0    <html backend_node_id=\"338\">\\n  <body>\\n    <d...   \n",
       "1    <html backend_node_id=\"14307\">\\n  <div backend...   \n",
       "2    <html backend_node_id=\"44287\">\\n  <body>\\n    ...   \n",
       "3    <html backend_node_id=\"23177\">\\n  <body>\\n    ...   \n",
       "4    <html backend_node_id=\"123\">\\n  <body backend_...   \n",
       "..                                                 ...   \n",
       "187  <html backend_node_id=\"15457\">\\n  <body backen...   \n",
       "188  <html backend_node_id=\"4953\">\\n  <body>\\n    <...   \n",
       "189  <html backend_node_id=\"64176\">\\n  <body backen...   \n",
       "190  <html backend_node_id=\"1531\">\\n  <body>\\n    <...   \n",
       "191  <html backend_node_id=\"329709\">\\n  <body backe...   \n",
       "\n",
       "                               action_uid  \\\n",
       "0    96238fb3-bc46-4dff-95c0-4b2d4ecc70ea   \n",
       "1    8043e8e8-bb7a-4521-80bf-246bc16e883c   \n",
       "2    b416d08b-90c4-43e1-b3d6-9d4a05e56a06   \n",
       "3    fd713700-3876-44a8-80ab-4da898beab42   \n",
       "4    01342a5e-5cc3-45e7-bd9d-b522611fc7bf   \n",
       "..                                    ...   \n",
       "187  5036b341-d933-4f21-9bde-af116b2f78e2   \n",
       "188  bf65fe94-df6e-4817-8530-93bf6b9800a3   \n",
       "189  c7cd25d6-0721-488d-9a01-975f6b1bc76d   \n",
       "190  370ca9f4-e6ee-4576-bcf1-ea418704debf   \n",
       "191  3bb445bc-1794-4365-8d2e-7eaea29db4e3   \n",
       "\n",
       "                                          source_nodes  \\\n",
       "0    <div backend_node_id=\"2662\">\\n                ...   \n",
       "1    <div backend_node_id=\"14708\">\\n               ...   \n",
       "2    <span backend_node_id=\"44871\">\\n              ...   \n",
       "3    <div backend_node_id=\"23101\">\\n               ...   \n",
       "4    <body backend_node_id=\"187\">\\n    <div backend...   \n",
       "..                                                 ...   \n",
       "187  <div backend_node_id=\"16067\">\\n               ...   \n",
       "188  <div backend_node_id=\"5365\">\\n              <f...   \n",
       "189  <p backend_node_id=\"67582\">\\n                 ...   \n",
       "190  <div backend_node_id=\"1948\">\\n                ...   \n",
       "191  <div backend_node_id=\"344655\">\\n              ...   \n",
       "\n",
       "                                     ground_truth_code  \\\n",
       "0    driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "1    driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "2    driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "3    driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "4    driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "..                                                 ...   \n",
       "187  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "188  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "189  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "190  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "191  driver.find_element(By.XPATH, f'//*[@backend_n...   \n",
       "\n",
       "                               ground_truth_outer_html  \\\n",
       "0    <input backend_node_id=\"248\" type=\"text\" name=...   \n",
       "1    <li backend_node_id=\"14603\" role=\"option\" titl...   \n",
       "2    <input backend_node_id=\"44285\" type=\"checkbox\"...   \n",
       "3    <li backend_node_id=\"23642\" role=\"option\">\\n  ...   \n",
       "4        <img backend_node_id=\"117\" alt=\"James Smith\">   \n",
       "..                                                 ...   \n",
       "187   <svg backend_node_id=\"15442\" class=\"icon\"></svg>   \n",
       "188  <a backend_node_id=\"4947\">\\n                  ...   \n",
       "189  <a backend_node_id=\"64159\">\\n                 ...   \n",
       "190  <input backend_node_id=\"1528\" type=\"text\" name...   \n",
       "191  <li backend_node_id=\"335800\">\\n               ...   \n",
       "\n",
       "                                        generated_code  \\\n",
       "0    # Let's proceed step by step.\\n# First we need...   \n",
       "1    # Let's proceed step by step.\\n# First we need...   \n",
       "2    # Let's proceed step by step.\\n# First we need...   \n",
       "3    # Let's proceed step by step.\\n# First we need...   \n",
       "4    # Let's proceed step by step.\\n# First we need...   \n",
       "..                                                 ...   \n",
       "187  # Let's proceed step by step.\\n# First we need...   \n",
       "188  # Let's proceed step by step.\\n# First we need...   \n",
       "189  # Let's proceed step by step.\\n# First we need...   \n",
       "190  # Let's proceed step by step.\\n# First we need...   \n",
       "191  # Let's proceed step by step.\\n# First we need...   \n",
       "\n",
       "                                     target_outer_html execution_error  \\\n",
       "0    <text backend_node_id=\"2663\">Hi ! I have a Iba...                   \n",
       "1            <text backend_node_id=\"14606\">AAPL</text>                   \n",
       "2    <input backend_node_id=\"44285\" type=\"checkbox\"...                   \n",
       "3           <text backend_node_id=\"23648\">Smile</text>                   \n",
       "4        <img backend_node_id=\"117\" alt=\"James Smith\">                   \n",
       "..                                                 ...             ...   \n",
       "187  <form backend_node_id=\"15635\" name=\"search-for...                   \n",
       "188  <text backend_node_id=\"5568\">Benefits and fina...                   \n",
       "189  <text backend_node_id=\"66373\">Thrill Rides</text>                   \n",
       "190  <input backend_node_id=\"1528\" type=\"text\" name...                   \n",
       "191  <h2 backend_node_id=\"329691\" title=\"Expand Pri...                   \n",
       "\n",
       "                   retriever         model  \n",
       "0    get_retriever_recursive  gpt-35-turbo  \n",
       "1    get_retriever_recursive  gpt-35-turbo  \n",
       "2    get_retriever_recursive  gpt-35-turbo  \n",
       "3    get_retriever_recursive  gpt-35-turbo  \n",
       "4    get_retriever_recursive  gpt-35-turbo  \n",
       "..                       ...           ...  \n",
       "187  get_retriever_recursive  gpt-35-turbo  \n",
       "188  get_retriever_recursive  gpt-35-turbo  \n",
       "189  get_retriever_recursive  gpt-35-turbo  \n",
       "190  get_retriever_recursive  gpt-35-turbo  \n",
       "191  get_retriever_recursive  gpt-35-turbo  \n",
       "\n",
       "[192 rows x 15 columns]"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "row = results.iloc[187]\n",
    "\n",
    "html = row[\"html\"]\n",
    "load_html(html, driver)\n",
    "\n",
    "ground_truth_code = row[\"ground_truth_code\"]\n",
    "exec(f\"ground_truth_element = {ground_truth_code}\")\n",
    "generated_code = row[\"generated_code\"]\n",
    "\n",
    "code = keep_assignments(generated_code)\n",
    "\n",
    "# Split the code into lines and keep only the first assignment\n",
    "code = code.split(\"\\n\")[0]\n",
    "parsed_code = ast.parse(code)\n",
    "\n",
    "# Create a visitor instance and use it to visit the nodes in the parsed AST\n",
    "visitor = VariableVisitor()\n",
    "visitor.visit(parsed_code)\n",
    "variable_name = visitor.output[0]\n",
    "\n",
    "# Execute the code to define the first variable\n",
    "exec(code)\n",
    "\n",
    "# Assign the variable to the target_element variable which will be used afterwards to compute score\n",
    "target_element = None\n",
    "exec(f\"\"\"target_element = {variable_name}\"\"\")\n",
    "\n",
    "def highlight_element(element, driver):\n",
    "    driver.execute_script(\"\"\"\n",
    "    arguments[0].style.outline = \"10px dashed blue\";\n",
    "    arguments[0].style.backgroundColor = \"rgba(135, 206, 250, 0.3)\";\n",
    "    arguments[0].style.borderRadius = \"50px\";\"\"\", element)\n",
    "    \n",
    "def unhighlight_element(element, driver):\n",
    "    driver.execute_script(\"\"\"\n",
    "    arguments[0].style.outline = \"\";\n",
    "    arguments[0].style.backgroundColor = \"\";\n",
    "    arguments[0].style.borderRadius = \"\";\"\"\", element)\n",
    "\n",
    "highlight_element(ground_truth_element, driver)\n",
    "driver.save_screenshot(\"screenshot1.png\")\n",
    "unhighlight_element(ground_truth_element, driver)\n",
    "\n",
    "highlight_element(target_element, driver)\n",
    "driver.save_screenshot(\"screenshot2.png\")\n",
    "\n",
    "def combine_screenshots(output_path=\"combined_screenshot.png\"):\n",
    "    from PIL import Image, ImageDraw, ImageFont\n",
    "\n",
    "    # Load the images\n",
    "    image1 = Image.open(\"screenshot1.png\")\n",
    "    image2 = Image.open(\"screenshot2.png\")\n",
    "    \n",
    "    # Calculate dimensions for the combined image\n",
    "    total_width = image1.width + image2.width\n",
    "    max_height = max(image1.height, image2.height)\n",
    "    \n",
    "    # Create a new blank image with the correct size\n",
    "    combined_image = Image.new('RGB', (total_width, max_height))\n",
    "    \n",
    "    # Paste the original images into the combined image\n",
    "    combined_image.paste(image1, (0, 0))\n",
    "    combined_image.paste(image2, (image1.width, 0))\n",
    "    \n",
    "    draw = ImageDraw.Draw(combined_image)\n",
    "    font = ImageFont.load_default()  # Load default font\n",
    "    draw.text((10, 10), \"Targeted element\", fill=\"black\", font=font)\n",
    "    draw.text((image1.width + 10, 10), \"Ground truth\", fill=\"black\", font=font)\n",
    "\n",
    "    \n",
    "    # Save the combined image\n",
    "    combined_image.save(output_path)\n",
    "\n",
    "from IPython.display import Image, display\n",
    "combine_screenshots()\n",
    "display(Image(\"combined_screenshot.png\"))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "myenv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}