firecrawl 3.2.1__py3-none-any.whl → 3.3.0__py3-none-any.whl

firecrawl/__init__.py

@@ -17,7 +17,7 @@ from .v1 import (
     V1ChangeTrackingOptions,
 )
 
-__version__ = "3.2.1"
+__version__ = "3.3.0"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

firecrawl/__tests__/unit/v2/methods/aio/test_aio_crawl_request_preparation.py

@@ -14,7 +14,7 @@ class TestAsyncCrawlRequestPreparation:
             include_paths=["/docs/*"],
             exclude_paths=["/admin/*"],
             max_discovery_depth=2,
-            ignore_sitemap=True,
+            sitemap="skip",
             ignore_query_parameters=True,
             crawl_entire_domain=True,
             allow_external_links=False,
@@ -26,7 +26,7 @@ class TestAsyncCrawlRequestPreparation:
         assert payload["includePaths"] == ["/docs/*"]
         assert payload["excludePaths"] == ["/admin/*"]
         assert payload["maxDiscoveryDepth"] == 2
-        assert payload["ignoreSitemap"] is True
+        assert payload["sitemap"] == "skip"
         assert payload["ignoreQueryParameters"] is True
         assert payload["crawlEntireDomain"] is True
         assert payload["allowExternalLinks"] is False

firecrawl/__tests__/unit/v2/methods/test_crawl_request_preparation.py

@@ -24,7 +24,7 @@ class TestCrawlRequestPreparation:
             url="https://example.com",
             limit=10,
             max_discovery_depth=3,
-            ignore_sitemap=True,
+            sitemap="skip",
             crawl_entire_domain=False,
             allow_external_links=True
         )
@@ -39,8 +39,8 @@ class TestCrawlRequestPreparation:
         assert data["limit"] == 10
         assert "maxDiscoveryDepth" in data
         assert data["maxDiscoveryDepth"] == 3
-        assert "ignoreSitemap" in data
-        assert data["ignoreSitemap"] is True
+        assert "sitemap" in data
+        assert data["sitemap"] == "skip"
         assert "crawlEntireDomain" in data
         assert data["crawlEntireDomain"] is False
         assert "allowExternalLinks" in data
@@ -106,7 +106,7 @@ class TestCrawlRequestPreparation:
             include_paths=["/blog/*", "/docs/*"],
             exclude_paths=["/admin/*"],
             max_discovery_depth=3,
-            ignore_sitemap=False,
+            sitemap="include",
             limit=100,
             crawl_entire_domain=True,
             allow_external_links=False,
@@ -126,8 +126,8 @@ class TestCrawlRequestPreparation:
         assert data["excludePaths"] == ["/admin/*"]
         assert "maxDiscoveryDepth" in data
         assert data["maxDiscoveryDepth"] == 3
-        assert "ignoreSitemap" in data
-        assert data["ignoreSitemap"] is False
+        assert "sitemap" in data
+        assert data["sitemap"] == "include"
         assert "limit" in data
         assert data["limit"] == 100
         assert "crawlEntireDomain" in data

firecrawl/v2/methods/search.py

@@ -121,6 +121,17 @@ def _validate_search_request(request: SearchRequest) -> SearchRequest:
                 if source.type not in valid_sources:
                     raise ValueError(f"Invalid source type: {source.type}. Valid types: {valid_sources}")
     
+    # Validate categories (if provided)
+    if request.categories is not None:
+        valid_categories = {"github", "research"}
+        for category in request.categories:
+            if isinstance(category, str):
+                if category not in valid_categories:
+                    raise ValueError(f"Invalid category type: {category}. Valid types: {valid_categories}")
+            elif hasattr(category, 'type'):
+                if category.type not in valid_categories:
+                    raise ValueError(f"Invalid category type: {category.type}. Valid types: {valid_categories}")
+    
     # Validate location (if provided)
     if request.location is not None:
         if not isinstance(request.location, str) or len(request.location.strip()) == 0:

firecrawl/v2/types.py

@@ -174,6 +174,12 @@ class Source(BaseModel):
 
 SourceOption = Union[str, Source]
 
+class Category(BaseModel):
+    """Configuration for a search category."""
+    type: str
+
+CategoryOption = Union[str, Category]
+
 FormatString = Literal[
     # camelCase versions (API format)
     "markdown", "html", "rawHtml", "links", "screenshot", "summary", "changeTracking", "json",
@@ -331,7 +337,8 @@ class SearchResultWeb(BaseModel):
     """A web search result with URL, title, and description."""
     url: str
     title: Optional[str] = None
-    description: Optional[str] = None 
+    description: Optional[str] = None
+    category: Optional[str] = None
 
 class SearchResultNews(BaseModel):
   """A news search result with URL, title, snippet, date, image URL, and position."""
@@ -341,6 +348,7 @@ class SearchResultNews(BaseModel):
   date: Optional[str] = None
   image_url: Optional[str] = None
   position: Optional[int] = None
+  category: Optional[str] = None
 
 class SearchResultImages(BaseModel):
   """An image search result with URL, title, image URL, image width, image height, and position."""
@@ -521,6 +529,7 @@ class SearchRequest(BaseModel):
     """Request for search operations."""
     query: str
     sources: Optional[List[SourceOption]] = None
+    categories: Optional[List[CategoryOption]] = None
     limit: Optional[int] = 5
     tbs: Optional[str] = None
     location: Optional[str] = None
@@ -547,6 +556,26 @@ class SearchRequest(BaseModel):
                 raise ValueError(f"Invalid source format: {source}")
         
         return normalized_sources
+    
+    @field_validator('categories')
+    @classmethod
+    def validate_categories(cls, v):
+        """Validate and normalize categories input."""
+        if v is None:
+            return v
+        
+        normalized_categories = []
+        for category in v:
+            if isinstance(category, str):
+                normalized_categories.append(Category(type=category))
+            elif isinstance(category, dict):
+                normalized_categories.append(Category(**category))
+            elif isinstance(category, Category):
+                normalized_categories.append(category)
+            else:
+                raise ValueError(f"Invalid category format: {category}")
+        
+        return normalized_categories
 
 class LinkResult(BaseModel):
     """A generic link result with optional metadata (used by search and map)."""

{firecrawl-3.2.1.dist-info → firecrawl-3.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.1
 Name: firecrawl
-Version: 3.2.1
+Version: 3.3.0
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/firecrawl/firecrawl
 Author: Mendable.ai
@@ -38,12 +38,8 @@ Requires-Dist: httpx
 Requires-Dist: python-dotenv
 Requires-Dist: websockets
 Requires-Dist: nest-asyncio
-Requires-Dist: pydantic>=2.0
+Requires-Dist: pydantic (>=2.0)
 Requires-Dist: aiohttp
-Dynamic: author
-Dynamic: home-page
-Dynamic: license-file
-Dynamic: requires-python
 
 # Firecrawl Python SDK
 

{firecrawl-3.2.1.dist-info → firecrawl-3.3.0.dist-info}/RECORD

@@ -1,4 +1,4 @@
-firecrawl/__init__.py,sha256=SoaA5_UyKZxxrqkDlOfvoYOdHgPCsagI1bQZRameUNw,2192
+firecrawl/__init__.py,sha256=B7SJY_289oeNIVIPxklkrqFQQURmszzwEqFjYXDfd20,2192
 firecrawl/client.py,sha256=2BGIRTiW2eR6q3wu_g2s3VTQtrHYauoDeNF1YklQpHo,11089
 firecrawl/firecrawl.backup.py,sha256=v1FEN3jR4g5Aupg4xp6SLkuFvYMQuUKND2YELbYjE6c,200430
 firecrawl/types.py,sha256=W9N2pqQuevEIIjYHN9rbDf31E-nwdCECqIn11Foz2T8,2836
@@ -22,7 +22,7 @@ firecrawl/__tests__/e2e/v2/aio/test_aio_usage.py,sha256=Dh9BVo48NKSZOKgLbO7n8fpM
 firecrawl/__tests__/e2e/v2/aio/test_aio_watcher.py,sha256=hwES4Nu5c0hniZ9heIPDfvh_2JmJ2wPoX9ULTZ0Asjs,1471
 firecrawl/__tests__/unit/v2/methods/test_batch_request_preparation.py,sha256=HeOxN-sPYSssytcIRAEicJSZsFt_Oa5qGXAtdumR54c,4040
 firecrawl/__tests__/unit/v2/methods/test_crawl_params.py,sha256=p9hzg14uAs1iHKXPDSXhGU6hEzPBF_Ae34RAf5XYa10,2387
-firecrawl/__tests__/unit/v2/methods/test_crawl_request_preparation.py,sha256=9DbLkBg6tuMyg7ASGX_oaQmAy0VCV4oITCOrfeR2UkY,8806
+firecrawl/__tests__/unit/v2/methods/test_crawl_request_preparation.py,sha256=PEKbooNXfQwPpvcPHXABJnveztgAA-RFBhtlSs8uPro,8780
 firecrawl/__tests__/unit/v2/methods/test_crawl_validation.py,sha256=kErOmHSD01eMjXiMd4rgsMVGd_aU2G9uVymBjbAFoGw,3918
 firecrawl/__tests__/unit/v2/methods/test_map_request_preparation.py,sha256=toVcgnMp_cFeYsIUuyKGEWZGp0nAAkzaeFGUbY0zY0o,1868
 firecrawl/__tests__/unit/v2/methods/test_scrape_request_preparation.py,sha256=wDOslsA5BN4kyezlaT5GeMv_Ifn8f461EaA7i5ujnaQ,3482
@@ -31,7 +31,7 @@ firecrawl/__tests__/unit/v2/methods/test_search_validation.py,sha256=7UGcNHpQzCp
 firecrawl/__tests__/unit/v2/methods/test_usage_types.py,sha256=cCHHfa6agSjD0brQ9rcAcw2kaI9riUH5C0dXV-fqktg,591
 firecrawl/__tests__/unit/v2/methods/test_webhook.py,sha256=AvvW-bKpUA--Lvtif2bmUIp-AxiaMJ29ie1i9dk8WbI,4586
 firecrawl/__tests__/unit/v2/methods/aio/test_aio_crawl_params.py,sha256=9azJxVvDOBqUevLp-wBF9gF7Ptj-7nN6LOkPQncFX2M,456
-firecrawl/__tests__/unit/v2/methods/aio/test_aio_crawl_request_preparation.py,sha256=vFbitMhH92JS5AtU78KQPi6kbT2fv68i9-rBrY5hVss,2574
+firecrawl/__tests__/unit/v2/methods/aio/test_aio_crawl_request_preparation.py,sha256=RkIKt7uxBzVhAkrLQwXYjmC-9sj32SUNQrJZgF2WEMs,2565
 firecrawl/__tests__/unit/v2/methods/aio/test_aio_crawl_validation.py,sha256=WMgltdrrT2HOflqGyahC4v-Wb29_8sypN0hwS9lYXe8,403
 firecrawl/__tests__/unit/v2/methods/aio/test_aio_map_request_preparation.py,sha256=PdUJrR0JLWqrithAnRXwuRrnsIN2h_DTu6-xvTOn_UU,725
 firecrawl/__tests__/unit/v2/methods/aio/test_aio_scrape_request_preparation.py,sha256=A5DT4wpH4vrIPvFxKVHrtDH5A3bgJ_ad4fmVQ8LN1t0,1993
@@ -45,7 +45,7 @@ firecrawl/v1/client.py,sha256=sydurfEFTsXyowyaGryA1lkPxN_r9Nf6iQpM43OwJyM,201672
 firecrawl/v2/__init__.py,sha256=Jc6a8tBjYG5OPkjDM5pl-notyys-7DEj7PLEfepv3fc,137
 firecrawl/v2/client.py,sha256=P6WAzwYGLLIANTrqAM-K4EUdGWQoFsi-zCjBibbxKQw,30507
 firecrawl/v2/client_async.py,sha256=zwxHis1bSh0tSF1480ze-4XDQEDJ5yDur1ZqtL94dwc,10127
-firecrawl/v2/types.py,sha256=bbHXPWJp6Kvjx9rKkTPyWZwdqVTErS4VYZKfHsb7ZQc,21137
+firecrawl/v2/types.py,sha256=rBdTaTQmb1SmdR8O0GvA_gCfBG-QCtBOrMsFgA9Usms,22114
 firecrawl/v2/watcher.py,sha256=FOU71tqSKxgeuGycu4ye0SLc2dw7clIcoQjPsi-4Csc,14229
 firecrawl/v2/watcher_async.py,sha256=AVjW2mgABniolSsauK4u0FW8ya6WzRUdyEg2R-8vGCw,10278
 firecrawl/v2/methods/batch.py,sha256=us7zUGl7u9ZDIEk2J3rNqj87bkaNjXU27SMFW_fdcg8,11932
@@ -53,7 +53,7 @@ firecrawl/v2/methods/crawl.py,sha256=4ZUmanHNuNtq9wbKMAZ3lenuPcNdOaV0kYXqMI5XJJ8
 firecrawl/v2/methods/extract.py,sha256=-Jr4BtraU3b7hd3JIY73V-S69rUclxyXyUpoQb6DCQk,4274
 firecrawl/v2/methods/map.py,sha256=4SADb0-lkbdOWDmO6k8_TzK0yRti5xsN40N45nUl9uA,2592
 firecrawl/v2/methods/scrape.py,sha256=CSHBwC-P91UfrW3zHirjNAs2h899FKcWvd1DY_4fJdo,1921
-firecrawl/v2/methods/search.py,sha256=c6tkDQGYZeLsPABPVfzhjalsasnhlien3w80aoe89t0,7077
+firecrawl/v2/methods/search.py,sha256=6BKiQ1aKJjWBKm9BBtKxFKGD74kCKBeMIp_OgjcDFAw,7673
 firecrawl/v2/methods/usage.py,sha256=OJlkxwaB-AAtgO3WLr9QiqBRmjdh6GVhroCgleegupQ,1460
 firecrawl/v2/methods/aio/__init__.py,sha256=RocMJnGwnLIvGu3G8ZvY8INkipC7WHZiu2bE31eSyJs,35
 firecrawl/v2/methods/aio/batch.py,sha256=GS_xsd_Uib1fxFITBK1sH88VGzFMrIcqJVQqOvMQ540,3735
@@ -70,10 +70,10 @@ firecrawl/v2/utils/http_client.py,sha256=_n8mp4xi6GGihg662Lsv6TSlvw9zykyADwEk0fg
 firecrawl/v2/utils/http_client_async.py,sha256=iy89_bk2HS3afSRHZ8016eMCa9Fk-5MFTntcOHfbPgE,1936
 firecrawl/v2/utils/normalize.py,sha256=nlTU6QRghT1YKZzNZlIQj4STSRuSUGrS9cCErZIcY5w,3636
 firecrawl/v2/utils/validation.py,sha256=L8by7z-t6GuMGIYkK7il1BM8d-4_-sAdG9hDMF_LeG4,14518
-firecrawl-3.2.1.dist-info/licenses/LICENSE,sha256=nPCunEDwjRGHlmjvsiDUyIWbkqqyj3Ej84ntnh0g0zA,1084
 tests/test_change_tracking.py,sha256=_IJ5ShLcoj2fHDBaw-nE4I4lHdmDB617ocK_XMHhXps,4177
 tests/test_timeout_conversion.py,sha256=PWlIEMASQNhu4cp1OW_ebklnE9NCiigPnEFCtI5N3w0,3996
-firecrawl-3.2.1.dist-info/METADATA,sha256=k1UNlt3XP09k-9i3oAkm4ElYkZOflYHkFIWuOHD4YsU,7392
-firecrawl-3.2.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-firecrawl-3.2.1.dist-info/top_level.txt,sha256=8T3jOaSN5mtLghO-R3MQ8KO290gIX8hmfxQmglBPdLE,16
-firecrawl-3.2.1.dist-info/RECORD,,
+firecrawl-3.3.0.dist-info/LICENSE,sha256=nPCunEDwjRGHlmjvsiDUyIWbkqqyj3Ej84ntnh0g0zA,1084
+firecrawl-3.3.0.dist-info/METADATA,sha256=Yk4-i0M6LdZdXoOQgdXyaclEX695WhU0JjhfdPePg9w,7313
+firecrawl-3.3.0.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+firecrawl-3.3.0.dist-info/top_level.txt,sha256=8T3jOaSN5mtLghO-R3MQ8KO290gIX8hmfxQmglBPdLE,16
+firecrawl-3.3.0.dist-info/RECORD,,

{firecrawl-3.2.1.dist-info → firecrawl-3.3.0.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.9.0)
+Generator: bdist_wheel (0.38.4)
 Root-Is-Purelib: true
 Tag: py3-none-any