web seraping ai agent

{
  job_categories[]
  jobs[] {
    company_name
    role
  }
}

{
    search_box
}

  "search_box": {
    "role": "text",
    "tf623_id": "299",
    "html_tag": "span",
    "name": "Search",
    "attributes": {
      "class": "DocSearch-Button-Placeholder"
    }
  }
}

{
    headings(all the headings inside the article)[]
}

{
  "headings": [
    "AgentQL Quick Start",
    "What's an AgentQL query?",
    "Get your API key",
    "Query this page with the AgentQL Debugger Chrome extension",
    "Try it out",
    "Perform the query with the AgentQL SDK",
    "Next steps",
    "Table of contents",
    "How would you rate your experience?"
  ]
}

python -m venv agentql
source agentql/bin/activat

pip3 install agentql
agentql init

Installing dependencies...
Get your AgentQL API key at https://dev.agentql.com
Enter your AgentQL API key:

import agentql
from playwright.sync_api import sync_playwright

# Initialise the browser
with sync_playwright() as playwright, playwright.chromium.launch(headless=False) as browser:
    page = agentql.wrap(browser.new_page())
    page.goto("https://docs.agentql.com/quick-start")

    # Find "Search" button using Smart Locator
    search_button = page.get_by_prompt("search button")
    # Interact with the button
    search_button.click()

    # Define a query for modal dialog's search input
    SEARCH_BOX_QUERY = """
    {
        modal {
            search_box
        }
    }
    """

    # Get the modal's search input and fill it with "Quick Start"
    response = page.query_elements(SEARCH_BOX_QUERY)
    response.modal.search_box.type("Quick Start")

    # Define a query for the search results
    SEARCH_RESULTS_QUERY = """
    {
        modal {
            search_box
            search_results {
                items[]
            }
        }
    }
    """

    # Execute the query after the results have returned then click on the first one
    response = page.query_elements(SEARCH_RESULTS_QUERY)
    response.modal.search_results.items[0].click()

    # Used only for demo purposes. It allows you to see the effect of the script.
    page.wait_for_timeout(10000)

Traceback (most recent call last):
  File "/Users/tanabeshouhei/work/agnetql/agentql/example_script.py", line 41, in <module>
    response.modal.search_results.items[0].click()
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^
  File "/Users/tanabeshouhei/work/agnetql/agentql/lib/python3.11/site-packages/agentql/ext/playwright/sync_api/response_proxy.py", line 36, in __getitem__
    return super().__getitem__(index)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/tanabeshouhei/work/agnetql/agentql/lib/python3.11/site-packages/agentql/ext/playwright/_response_proxy_parent.py", line 31, in __getitem__
    return self._resolve_item(self._response_data[index], self._query_tree_node)  # type: ignore # returned value could be None, but to make static checker happy we ignore it
                              ~~~~~~~~~~~~~~~~~~~^^^^^^^
IndexError: list index out of range

response = page.query_elements(SEARCH_RESULTS_QUERY)

response = page.query_elements(SEARCH_RESULTS_QUERY)
print(response.modal.search_results.items)  # デバッグ用

{
    login_btn(the one in header section)
    footer {
        social_media_links(The icons that lead to Facebook, Snapchat, etc.)[]
    }
}

{
    modal {
        search_results {
            items[] {
                title
                link
            }
        }
    }
}

    # Define a query for the search results
    SEARCH_RESULTS_QUERY = """
    {
        modal {
            search_box
            search_results {
                items[]
            }
        }
    }
    """

    # Define a query for the search results
    SEARCH_RESULTS_QUERY = """
    {
        modal {
            search_results {
                items[] {
                    title
                    link
                }
            }
        }
    }
    """

[]
No search results found.

    # Wait for 3 seconds
    page.wait_for_timeout(3000)  # 3秒待機

[
  {
    "title": {
      "role": "text",
      "tf623_id": "1348",
      "html_tag": "span",
      "name": "AgentQL",
      "attributes": {}
    },
    "link": {
      "role": "link",
      "tf623_id": "1342",
      "html_tag": "a",
      "name": "",
      "attributes": {
        "class": "transition-colors duration-200 DocSearch-Hit--Child",
        "href": "/quick-start"
      }
    }
  },
  {
    "title": {
      "role": "text",
      "tf623_id": "1360",
      "html_tag": "span",
      "name": "What's an AgentQL query?",
      "attributes": {
        "class": "DocSearch-Hit-title"
      }
    },
    "link": {
      "role": "link",
      "tf623_id": "1354",
      "html_tag": "a",
      "name": "",
      "attributes": {
        "class": "transition-colors duration-200 DocSearch-Hit--Child",
        "href": "/quick-start#whats-an-agentql-query"
      }
    }
  },
  {
    "title": {
      "role": "text",
      "tf623_id": "1374",
      "html_tag": "span",
      "name": "Get your API key",
      "attributes": {
        "class": "DocSearch-Hit-title"
      }
    },
    "link": {
      "role": "link",
      "tf623_id": "1368",
      "html_tag": "a",
      "name": "",
      "attributes": {
        "class": "transition-colors duration-200 DocSearch-Hit--Child",
        "href": "/quick-start#get-your-api-key"
      }
    }
  },

    # ログインボタンを特定するためのクエリを定義
    LOGIN_BUTTON_QUERY = """
    {
      login_btn(the one in header section containing "ログイン")
    }

login_btn(ヘッダーの中にあり、テキストが「ログイン」である)

    # メール入力フォームを取得してメールアドレスを入力
    response = page.query_elements(EMAIL_INPUT_QUERY)
    response.email_input_form.type("example@gmail.com")  # メールアドレスを入力

    # 次に進むを取得するためのクエリを定義
    NEXT_BUTTON_QUERY = """
    {
      next_input_button(type=submit)
    }
    """

    # 次に進むを取得してクリック
    response = page.query_elements(NEXT_BUTTON_QUERY)
    print(response)  # デバッグ用にレスポンスを出力
    response.next_input_button.click()  # 次に進むをクリック

response.password_input_form.type("password")  # パスワードを入力

{
  "password_input_form": {
    "role": "form",
    "tf623_id": "5150",
    "html_tag": "form",
    "name": "signIn",
    "attributes": {
      "name": "signIn",
      "method": "post",
      "novalidate": true,
      "action": "https://www.amazon.co.jp/ap/signin",
      "class": "auth-validate-form auth-real-time-validation a-spacing-none",
      "data-fwcim-id": "GK7iSGYJ"
    }
  }
}
{
  "login_btn": {
    "role": "button",
    "tf623_id": "5176",
    "html_tag": "input",
    "name": "",
    "attributes": {
      "id": "signInSubmit",
      "class": "a-button-input",
      "type": "submit",
      "aria-labelledby": "auth-signin-button-announce"
    }
  }
}

with sync_playwright() as playwright, playwright.chromium.launch(headless=False) as browser:

with sync_playwright() as playwright, playwright.chromium.launch(headless=True) as browser:

{"text":"8460sd"}

{"text_options": ["8460sd", "8460sD", "8460Sd", "8460sD", "8460sd"]}

{"text_candidates": [
{
"digit_1": ["8", "0", "6", "9", "3"],
"digit_2": ["4", "5", "7", "2", "1"],
"digit_3": ["6", "8", "0", "4", "2"],
"digit_4": ["s", "5", "$", "%", "&"],
"digit_5": ["d", "#", "@", "*", "+"],
"digit_6": ["2", "3", "0", "1", "4"]
}
]}