diff --git a/apps/python-sdk/build/lib/firecrawl/firecrawl.py b/apps/python-sdk/build/lib/firecrawl/firecrawl.py index 98cb8ed68..6c0bc4150 100644 --- a/apps/python-sdk/build/lib/firecrawl/firecrawl.py +++ b/apps/python-sdk/build/lib/firecrawl/firecrawl.py @@ -1,25 +1,57 @@ +""" +FirecrawlApp Module + +This module provides a class `FirecrawlApp` for interacting with the Firecrawl API. +It includes methods to scrape URLs, perform searches, initiate and monitor crawl jobs, +and check the status of these jobs. The module uses requests for HTTP communication +and handles retries for certain HTTP status codes. + +Classes: + - FirecrawlApp: Main class for interacting with the Firecrawl API. +""" + import os +import time from typing import Any, Dict, Optional + import requests -import time + class FirecrawlApp: - def __init__(self, api_key=None, api_url='https://api.firecrawl.dev'): + """ + Initialize the FirecrawlApp instance. + + Args: + api_key (Optional[str]): API key for authenticating with the Firecrawl API. + api_url (Optional[str]): Base URL for the Firecrawl API. + """ + def __init__(self, api_key: Optional[str] = None, api_url: Optional[str] = None) -> None: self.api_key = api_key or os.getenv('FIRECRAWL_API_KEY') if self.api_key is None: raise ValueError('No API key provided') - self.api_url = api_url or os.getenv('FIRECRAWL_API_URL') - - - + self.api_url = api_url or os.getenv('FIRECRAWL_API_URL', 'https://api.firecrawl.dev') def scrape_url(self, url: str, params: Optional[Dict[str, Any]] = None) -> Any: + """ + Scrape the specified URL using the Firecrawl API. + + Args: + url (str): The URL to scrape. + params (Optional[Dict[str, Any]]): Additional parameters for the scrape request. + + Returns: + Any: The scraped data if the request is successful. + + Raises: + Exception: If the scrape request fails. + """ + headers = { 'Content-Type': 'application/json', 'Authorization': f'Bearer {self.api_key}' } # Prepare the base scrape parameters with the URL scrape_params = {'url': url} - + # If there are additional params, process them if params: # Initialize extractorOptions if present @@ -32,7 +64,7 @@ def scrape_url(self, url: str, params: Optional[Dict[str, Any]] = None) -> Any: extractor_options['mode'] = extractor_options.get('mode', 'llm-extraction') # Update the scrape_params with the processed extractorOptions scrape_params['extractorOptions'] = extractor_options - + # Include any other params directly at the top level of scrape_params for key, value in params.items(): if key != 'extractorOptions': @@ -41,11 +73,11 @@ def scrape_url(self, url: str, params: Optional[Dict[str, Any]] = None) -> Any: response = requests.post( f'{self.api_url}/v0/scrape', headers=headers, - json=scrape_params + json=scrape_params, ) if response.status_code == 200: response = response.json() - if response['success']: + if response['success'] and 'data' in response: return response['data'] else: raise Exception(f'Failed to scrape URL. Error: {response["error"]}') @@ -54,8 +86,21 @@ def scrape_url(self, url: str, params: Optional[Dict[str, Any]] = None) -> Any: raise Exception(f'Failed to scrape URL. Status code: {response.status_code}. Error: {error_message}') else: raise Exception(f'Failed to scrape URL. Status code: {response.status_code}') - + def search(self, query, params=None): + """ + Perform a search using the Firecrawl API. + + Args: + query (str): The search query. + params (Optional[Dict[str, Any]]): Additional parameters for the search request. + + Returns: + Any: The search results if the request is successful. + + Raises: + Exception: If the search request fails. + """ headers = { 'Content-Type': 'application/json', 'Authorization': f'Bearer {self.api_key}' @@ -70,19 +115,36 @@ def search(self, query, params=None): ) if response.status_code == 200: response = response.json() - if response['success'] == True: + + if response['success'] and 'data' in response: return response['data'] else: raise Exception(f'Failed to search. Error: {response["error"]}') - + elif response.status_code in [402, 409, 500]: error_message = response.json().get('error', 'Unknown error occurred') raise Exception(f'Failed to search. Status code: {response.status_code}. Error: {error_message}') else: raise Exception(f'Failed to search. Status code: {response.status_code}') - def crawl_url(self, url, params=None, wait_until_done=True, timeout=2): - headers = self._prepare_headers() + def crawl_url(self, url, params=None, wait_until_done=True, timeout=2, idempotency_key=None): + """ + Initiate a crawl job for the specified URL using the Firecrawl API. + + Args: + url (str): The URL to crawl. + params (Optional[Dict[str, Any]]): Additional parameters for the crawl request. + wait_until_done (bool): Whether to wait until the crawl job is completed. + timeout (int): Timeout between status checks when waiting for job completion. + idempotency_key (Optional[str]): A unique uuid key to ensure idempotency of requests. + + Returns: + Any: The crawl job ID or the crawl results if waiting until completion. + + Raises: + Exception: If the crawl job initiation or monitoring fails. + """ + headers = self._prepare_headers(idempotency_key) json_data = {'url': url} if params: json_data.update(params) @@ -97,6 +159,18 @@ def crawl_url(self, url, params=None, wait_until_done=True, timeout=2): self._handle_error(response, 'start crawl job') def check_crawl_status(self, job_id): + """ + Check the status of a crawl job using the Firecrawl API. + + Args: + job_id (str): The ID of the crawl job. + + Returns: + Any: The status of the crawl job. + + Raises: + Exception: If the status check request fails. + """ headers = self._prepare_headers() response = self._get_request(f'{self.api_url}/v0/crawl/status/{job_id}', headers) if response.status_code == 200: @@ -104,13 +178,45 @@ def check_crawl_status(self, job_id): else: self._handle_error(response, 'check crawl status') - def _prepare_headers(self): + def _prepare_headers(self, idempotency_key=None): + """ + Prepare the headers for API requests. + + Args: + idempotency_key (Optional[str]): A unique key to ensure idempotency of requests. + + Returns: + Dict[str, str]: The headers including content type, authorization, and optionally idempotency key. + """ + if idempotency_key: + return { + 'Content-Type': 'application/json', + 'Authorization': f'Bearer {self.api_key}', + 'x-idempotency-key': idempotency_key + } + return { 'Content-Type': 'application/json', - 'Authorization': f'Bearer {self.api_key}' + 'Authorization': f'Bearer {self.api_key}', } def _post_request(self, url, data, headers, retries=3, backoff_factor=0.5): + """ + Make a POST request with retries. + + Args: + url (str): The URL to send the POST request to. + data (Dict[str, Any]): The JSON data to include in the POST request. + headers (Dict[str, str]): The headers to include in the POST request. + retries (int): Number of retries for the request. + backoff_factor (float): Backoff factor for retries. + + Returns: + requests.Response: The response from the POST request. + + Raises: + requests.RequestException: If the request fails after the specified retries. + """ for attempt in range(retries): response = requests.post(url, headers=headers, json=data) if response.status_code == 502: @@ -120,6 +226,21 @@ def _post_request(self, url, data, headers, retries=3, backoff_factor=0.5): return response def _get_request(self, url, headers, retries=3, backoff_factor=0.5): + """ + Make a GET request with retries. + + Args: + url (str): The URL to send the GET request to. + headers (Dict[str, str]): The headers to include in the GET request. + retries (int): Number of retries for the request. + backoff_factor (float): Backoff factor for retries. + + Returns: + requests.Response: The response from the GET request. + + Raises: + requests.RequestException: If the request fails after the specified retries. + """ for attempt in range(retries): response = requests.get(url, headers=headers) if response.status_code == 502: @@ -129,7 +250,20 @@ def _get_request(self, url, headers, retries=3, backoff_factor=0.5): return response def _monitor_job_status(self, job_id, headers, timeout): - import time + """ + Monitor the status of a crawl job until completion. + + Args: + job_id (str): The ID of the crawl job. + headers (Dict[str, str]): The headers to include in the status check requests. + timeout (int): Timeout between status checks. + + Returns: + Any: The crawl results if the job is completed successfully. + + Raises: + Exception: If the job fails or an error occurs during status checks. + """ while True: status_response = self._get_request(f'{self.api_url}/v0/crawl/status/{job_id}', headers) if status_response.status_code == 200: @@ -140,8 +274,7 @@ def _monitor_job_status(self, job_id, headers, timeout): else: raise Exception('Crawl job completed but no data was returned') elif status_data['status'] in ['active', 'paused', 'pending', 'queued']: - if timeout < 2: - timeout = 2 + timeout=max(timeout,2) time.sleep(timeout) # Wait for the specified timeout before checking again else: raise Exception(f'Crawl job failed or was stopped. Status: {status_data["status"]}') @@ -149,6 +282,16 @@ def _monitor_job_status(self, job_id, headers, timeout): self._handle_error(status_response, 'check crawl status') def _handle_error(self, response, action): + """ + Handle errors from API responses. + + Args: + response (requests.Response): The response object from the API request. + action (str): Description of the action that was being performed. + + Raises: + Exception: An exception with a message containing the status code and error details from the response. + """ if response.status_code in [402, 408, 409, 500]: error_message = response.json().get('error', 'Unknown error occurred') raise Exception(f'Failed to {action}. Status code: {response.status_code}. Error: {error_message}') diff --git a/apps/python-sdk/dist/firecrawl-py-0.0.10.tar.gz b/apps/python-sdk/dist/firecrawl-py-0.0.10.tar.gz new file mode 100644 index 000000000..59e42d15d Binary files /dev/null and b/apps/python-sdk/dist/firecrawl-py-0.0.10.tar.gz differ diff --git a/apps/python-sdk/dist/firecrawl-py-0.0.9.tar.gz b/apps/python-sdk/dist/firecrawl-py-0.0.9.tar.gz deleted file mode 100644 index 55e0eed7b..000000000 Binary files a/apps/python-sdk/dist/firecrawl-py-0.0.9.tar.gz and /dev/null differ diff --git a/apps/python-sdk/dist/firecrawl_py-0.0.10-py3-none-any.whl b/apps/python-sdk/dist/firecrawl_py-0.0.10-py3-none-any.whl new file mode 100644 index 000000000..36676fe13 Binary files /dev/null and b/apps/python-sdk/dist/firecrawl_py-0.0.10-py3-none-any.whl differ diff --git a/apps/python-sdk/dist/firecrawl_py-0.0.9-py3-none-any.whl b/apps/python-sdk/dist/firecrawl_py-0.0.9-py3-none-any.whl deleted file mode 100644 index 83cb5b774..000000000 Binary files a/apps/python-sdk/dist/firecrawl_py-0.0.9-py3-none-any.whl and /dev/null differ diff --git a/apps/python-sdk/firecrawl/__pycache__/__init__.cpython-311.pyc b/apps/python-sdk/firecrawl/__pycache__/__init__.cpython-311.pyc deleted file mode 100644 index 605b3df71..000000000 Binary files a/apps/python-sdk/firecrawl/__pycache__/__init__.cpython-311.pyc and /dev/null differ diff --git a/apps/python-sdk/firecrawl/__pycache__/firecrawl.cpython-311.pyc b/apps/python-sdk/firecrawl/__pycache__/firecrawl.cpython-311.pyc deleted file mode 100644 index 7c98fa33c..000000000 Binary files a/apps/python-sdk/firecrawl/__pycache__/firecrawl.cpython-311.pyc and /dev/null differ diff --git a/apps/python-sdk/firecrawl_py.egg-info/PKG-INFO b/apps/python-sdk/firecrawl_py.egg-info/PKG-INFO index c1ee531a0..16af620ab 100644 --- a/apps/python-sdk/firecrawl_py.egg-info/PKG-INFO +++ b/apps/python-sdk/firecrawl_py.egg-info/PKG-INFO @@ -1,6 +1,6 @@ Metadata-Version: 2.1 Name: firecrawl-py -Version: 0.0.9 +Version: 0.0.10 Summary: Python SDK for Firecrawl API Home-page: https://github.com/mendableai/firecrawl Author: Mendable.ai diff --git a/apps/python-sdk/setup.py b/apps/python-sdk/setup.py index 7df520eb7..de49e1a5b 100644 --- a/apps/python-sdk/setup.py +++ b/apps/python-sdk/setup.py @@ -2,7 +2,7 @@ setup( name='firecrawl-py', - version='0.0.9', + version='0.0.10', url='https://github.com/mendableai/firecrawl', author='Mendable.ai', author_email='nick@mendable.ai',