firecrawl 4.3.6__tar.gz → 4.4.0__tar.gz

{firecrawl-4.3.6 → firecrawl-4.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: firecrawl
-Version: 4.3.6
+Version: 4.4.0
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/firecrawl/firecrawl
 Author: Mendable.ai
@@ -40,6 +40,10 @@ Requires-Dist: websockets
 Requires-Dist: nest-asyncio
 Requires-Dist: pydantic>=2.0
 Requires-Dist: aiohttp
+Dynamic: author
+Dynamic: home-page
+Dynamic: license-file
+Dynamic: requires-python
 
 # Firecrawl Python SDK
 

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl/__init__.py

@@ -17,7 +17,7 @@ from .v1 import (
     V1ChangeTrackingOptions,
 )
 
-__version__ = "4.3.6"
+__version__ = "4.4.0"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl/v2/methods/crawl.py

@@ -333,7 +333,7 @@ def wait_for_crawl_completion(
         crawl_job = get_crawl_status(client, job_id)
         
         # Check if job is complete
-        if crawl_job.status in ["completed", "failed"]:
+        if crawl_job.status in ["completed", "failed", "cancelled"]:
             return crawl_job
         
         # Check timeout

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl/v2/methods/search.py

@@ -123,7 +123,7 @@ def _validate_search_request(request: SearchRequest) -> SearchRequest:
     
     # Validate categories (if provided)
     if request.categories is not None:
-        valid_categories = {"github", "research"}
+        valid_categories = {"github", "research", "pdf"}
         for category in request.categories:
             if isinstance(category, str):
                 if category not in valid_categories:

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl/v2/types.py

@@ -186,7 +186,13 @@ class Source(BaseModel):
 SourceOption = Union[str, Source]
 
 class Category(BaseModel):
-    """Configuration for a search category."""
+    """Configuration for a search category.
+    
+    Supported categories:
+    - "github": Filter results to GitHub repositories
+    - "research": Filter results to research papers and academic sites
+    - "pdf": Filter results to PDF files (adds filetype:pdf to search)
+    """
     type: str
 
 CategoryOption = Union[str, Category]
@@ -348,7 +354,7 @@ class CrawlResponse(BaseModel):
 
 class CrawlJob(BaseModel):
     """Crawl job status and progress data."""
-    status: Literal["scraping", "completed", "failed"]
+    status: Literal["scraping", "completed", "failed", "cancelled"]
     total: int = 0
     completed: int = 0
     credits_used: int = 0

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl/v2/watcher_async.py

@@ -28,13 +28,14 @@ class AsyncWatcher:
         job_id: str,
         *,
         kind: JobKind = "crawl",
+        poll_interval: int = 2,
         timeout: Optional[int] = None,
     ) -> None:
         self._client = client
         self._job_id = job_id
         self._kind = kind
         self._timeout = timeout
-        self._poll_interval: float = 2.0
+        self._poll_interval: float = max(0.0, float(poll_interval))  # Guard against negative values
 
         http_client = getattr(client, "http_client", None)
         if http_client is not None:

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: firecrawl
-Version: 4.3.6
+Version: 4.4.0
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/firecrawl/firecrawl
 Author: Mendable.ai
@@ -40,6 +40,10 @@ Requires-Dist: websockets
 Requires-Dist: nest-asyncio
 Requires-Dist: pydantic>=2.0
 Requires-Dist: aiohttp
+Dynamic: author
+Dynamic: home-page
+Dynamic: license-file
+Dynamic: requires-python
 
 # Firecrawl Python SDK
 

{firecrawl-4.3.6 → firecrawl-4.4.0}/firecrawl.egg-info/top_level.txt

@@ -1,2 +1,3 @@
+dist
 firecrawl
 tests