firecrawl 2.5.2__tar.gz → 2.5.4__tar.gz

{firecrawl-2.5.2 → firecrawl-2.5.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: firecrawl
-Version: 2.5.2
+Version: 2.5.4
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/mendableai/firecrawl
 Author: Mendable.ai

{firecrawl-2.5.2 → firecrawl-2.5.4}/firecrawl/__init__.py

@@ -13,7 +13,7 @@ import os
 
 from .firecrawl import FirecrawlApp, AsyncFirecrawlApp, JsonConfig, ScrapeOptions, ChangeTrackingOptions # noqa
 
-__version__ = "2.5.2"
+__version__ = "2.5.4"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

{firecrawl-2.5.2 → firecrawl-2.5.4}/firecrawl/firecrawl.py

@@ -161,7 +161,7 @@ class ScrapeOptions(pydantic.BaseModel):
 class WaitAction(pydantic.BaseModel):
     """Wait action to perform during scraping."""
     type: Literal["wait"]
-    milliseconds: int
+    milliseconds: Optional[int] = None
     selector: Optional[str] = None
 
 class ScreenshotAction(pydantic.BaseModel):
@@ -259,6 +259,7 @@ class CrawlParams(pydantic.BaseModel):
     deduplicateSimilarURLs: Optional[bool] = None
     ignoreQueryParameters: Optional[bool] = None
     regexOnFullURL: Optional[bool] = None
+    delay: Optional[int] = None  # Delay in seconds between scrapes
 
 class CrawlResponse(pydantic.BaseModel):
     """Response from crawling operations."""
@@ -540,9 +541,9 @@ class FirecrawlApp:
                 json_options["schema"] = self._ensure_schema_dict(json_options["schema"])
             scrape_params['jsonOptions'] = json_options if isinstance(json_options, dict) else json_options.dict(exclude_none=True)
         if actions:
-            scrape_params['actions'] = [action.dict(exclude_none=True) for action in actions]
+            scrape_params['actions'] = [action if isinstance(action, dict) else action.dict(exclude_none=True) for action in actions]
         if change_tracking_options:
-            scrape_params['changeTrackingOptions'] = change_tracking_options.dict(exclude_none=True)
+            scrape_params['changeTrackingOptions'] = change_tracking_options if isinstance(change_tracking_options, dict) else change_tracking_options.dict(exclude_none=True)
         
         scrape_params.update(kwargs)
 
@@ -681,6 +682,7 @@ class FirecrawlApp:
         deduplicate_similar_urls: Optional[bool] = None,
         ignore_query_parameters: Optional[bool] = None,
         regex_on_full_url: Optional[bool] = None,
+        delay: Optional[int] = None,
         poll_interval: Optional[int] = 2,
         idempotency_key: Optional[str] = None,
         **kwargs
@@ -703,6 +705,7 @@ class FirecrawlApp:
             deduplicate_similar_urls (Optional[bool]): Remove similar URLs
             ignore_query_parameters (Optional[bool]): Ignore URL parameters
             regex_on_full_url (Optional[bool]): Apply regex to full URLs
+            delay (Optional[int]): Delay in seconds between scrapes
             poll_interval (Optional[int]): Seconds between status checks (default: 2)
             idempotency_key (Optional[str]): Unique key to prevent duplicate requests
             **kwargs: Additional parameters to pass to the API
@@ -748,6 +751,8 @@ class FirecrawlApp:
             crawl_params['ignoreQueryParameters'] = ignore_query_parameters
         if regex_on_full_url is not None:
             crawl_params['regexOnFullURL'] = regex_on_full_url
+        if delay is not None:
+            crawl_params['delay'] = delay
 
         # Add any additional kwargs
         crawl_params.update(kwargs)
@@ -788,6 +793,7 @@ class FirecrawlApp:
         deduplicate_similar_urls: Optional[bool] = None,
         ignore_query_parameters: Optional[bool] = None,
         regex_on_full_url: Optional[bool] = None,
+        delay: Optional[int] = None,
         idempotency_key: Optional[str] = None,
         **kwargs
     ) -> CrawlResponse:
@@ -854,6 +860,8 @@ class FirecrawlApp:
             crawl_params['ignoreQueryParameters'] = ignore_query_parameters
         if regex_on_full_url is not None:
             crawl_params['regexOnFullURL'] = regex_on_full_url
+        if delay is not None:
+            crawl_params['delay'] = delay
 
         # Add any additional kwargs
         crawl_params.update(kwargs)
@@ -2924,7 +2932,7 @@ class AsyncFirecrawlApp(FirecrawlApp):
                 json_options["schema"] = self._ensure_schema_dict(json_options["schema"])
             scrape_params['jsonOptions'] = json_options if isinstance(json_options, dict) else json_options.dict(exclude_none=True)
         if actions:
-            scrape_params['actions'] = [action.dict(exclude_none=True) for action in actions]
+            scrape_params['actions'] = [action if isinstance(action, dict) else action.dict(exclude_none=True) for action in actions]
 
         if 'extract' in scrape_params and scrape_params['extract'] and 'schema' in scrape_params['extract']:
             scrape_params['extract']['schema'] = self._ensure_schema_dict(scrape_params['extract']['schema'])
@@ -3240,6 +3248,7 @@ class AsyncFirecrawlApp(FirecrawlApp):
         deduplicate_similar_urls: Optional[bool] = None,
         ignore_query_parameters: Optional[bool] = None,
         regex_on_full_url: Optional[bool] = None,
+        delay: Optional[int] = None,
         poll_interval: Optional[int] = 2,
         idempotency_key: Optional[str] = None,
         **kwargs
@@ -3262,6 +3271,7 @@ class AsyncFirecrawlApp(FirecrawlApp):
             deduplicate_similar_urls (Optional[bool]): Remove similar URLs
             ignore_query_parameters (Optional[bool]): Ignore URL parameters
             regex_on_full_url (Optional[bool]): Apply regex to full URLs
+            delay (Optional[int]): Delay in seconds between scrapes
             poll_interval (Optional[int]): Seconds between status checks (default: 2)
             idempotency_key (Optional[str]): Unique key to prevent duplicate requests
             **kwargs: Additional parameters to pass to the API
@@ -3307,6 +3317,8 @@ class AsyncFirecrawlApp(FirecrawlApp):
             crawl_params['ignoreQueryParameters'] = ignore_query_parameters
         if regex_on_full_url is not None:
             crawl_params['regexOnFullURL'] = regex_on_full_url
+        if delay is not None:
+            crawl_params['delay'] = delay
 
         # Add any additional kwargs
         crawl_params.update(kwargs)
@@ -3348,6 +3360,7 @@ class AsyncFirecrawlApp(FirecrawlApp):
         deduplicate_similar_urls: Optional[bool] = None,
         ignore_query_parameters: Optional[bool] = None,
         regex_on_full_url: Optional[bool] = None,
+        delay: Optional[int] = None,
         poll_interval: Optional[int] = 2,
         idempotency_key: Optional[str] = None,
         **kwargs
@@ -3412,6 +3425,8 @@ class AsyncFirecrawlApp(FirecrawlApp):
             crawl_params['ignoreQueryParameters'] = ignore_query_parameters
         if regex_on_full_url is not None:
             crawl_params['regexOnFullURL'] = regex_on_full_url
+        if delay is not None:
+            crawl_params['delay'] = delay
 
         # Add any additional kwargs
         crawl_params.update(kwargs)

{firecrawl-2.5.2 → firecrawl-2.5.4}/firecrawl.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: firecrawl
-Version: 2.5.2
+Version: 2.5.4
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/mendableai/firecrawl
 Author: Mendable.ai