firecrawl 1.7.0__py3-none-any.whl → 1.8.0__py3-none-any.whl

firecrawl/__init__.py

@@ -13,7 +13,7 @@ import os
 
 from .firecrawl import FirecrawlApp # noqa
 
-__version__ = "1.7.0"
+__version__ = "1.8.0"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

firecrawl/__tests__/v1/e2e_withAuth/test.py

@@ -371,4 +371,70 @@ def test_search_e2e():
 #     assert isinstance(llm_extraction['supports_sso'], bool)
 #     assert isinstance(llm_extraction['is_open_source'], bool)
 
+def test_search_with_string_query():
+    app = FirecrawlApp(api_url=API_URL, api_key=TEST_API_KEY)
+    response = app.search("firecrawl")
+    assert response["success"] is True
+    assert len(response["data"]) > 0
+    assert response["data"][0]["markdown"] is not None
+    assert response["data"][0]["metadata"] is not None
+    assert response["data"][0]["metadata"]["title"] is not None
+    assert response["data"][0]["metadata"]["description"] is not None
+
+def test_search_with_params_dict():
+    app = FirecrawlApp(api_url=API_URL, api_key=TEST_API_KEY)
+    response = app.search("firecrawl", {
+        "limit": 3,
+        "lang": "en",
+        "country": "us",
+        "scrapeOptions": {
+            "formats": ["markdown", "html", "links"],
+            "onlyMainContent": True
+        }
+    })
+    assert response["success"] is True
+    assert len(response["data"]) <= 3
+    for doc in response["data"]:
+        assert doc["markdown"] is not None
+        assert doc["html"] is not None
+        assert doc["links"] is not None
+        assert doc["metadata"] is not None
+        assert doc["metadata"]["title"] is not None
+        assert doc["metadata"]["description"] is not None
+
+def test_search_with_params_object():
+    app = FirecrawlApp(api_url=API_URL, api_key=TEST_API_KEY)
+    params = SearchParams(
+        query="firecrawl",
+        limit=3,
+        lang="en",
+        country="us",
+        scrapeOptions={
+            "formats": ["markdown", "html", "links"],
+            "onlyMainContent": True
+        }
+    )
+    response = app.search(params.query, params)
+    assert response["success"] is True
+    assert len(response["data"]) <= 3
+    for doc in response["data"]:
+        assert doc["markdown"] is not None
+        assert doc["html"] is not None
+        assert doc["links"] is not None
+        assert doc["metadata"] is not None
+        assert doc["metadata"]["title"] is not None
+        assert doc["metadata"]["description"] is not None
+
+def test_search_invalid_api_key():
+    app = FirecrawlApp(api_url=API_URL, api_key="invalid_api_key")
+    with pytest.raises(Exception) as e:
+        app.search("test query")
+    assert "404" in str(e.value)
+
+def test_search_with_invalid_params():
+    app = FirecrawlApp(api_url=API_URL, api_key=TEST_API_KEY)
+    with pytest.raises(Exception) as e:
+        app.search("test query", {"invalid_param": "value"})
+    assert "ValidationError" in str(e.value)
+
 

firecrawl/firecrawl.py

@@ -21,7 +21,28 @@ import websockets
 
 logger : logging.Logger = logging.getLogger("firecrawl")
 
+class SearchParams(pydantic.BaseModel):
+    query: str
+    limit: Optional[int] = 5
+    tbs: Optional[str] = None
+    filter: Optional[str] = None
+    lang: Optional[str] = "en"
+    country: Optional[str] = "us"
+    location: Optional[str] = None
+    origin: Optional[str] = "api"
+    timeout: Optional[int] = 60000
+    scrapeOptions: Optional[Dict[str, Any]] = None
+
 class FirecrawlApp:
+    class SearchResponse(pydantic.BaseModel):
+        """
+        Response from the search operation.
+        """
+        success: bool
+        data: List[Dict[str, Any]]
+        warning: Optional[str] = None
+        error: Optional[str] = None
+
     class ExtractParams(pydantic.BaseModel):
         """
         Parameters for the extract operation.
@@ -109,22 +130,36 @@ class FirecrawlApp:
         else:
             self._handle_error(response, 'scrape URL')
 
-    def search(self, query: str, params: Optional[Dict[str, Any]] = None) -> Any:
+    def search(self, query: str, params: Optional[Union[Dict[str, Any], SearchParams]] = None) -> Dict[str, Any]:
         """
-        Perform a search using the Firecrawl API.
+        Search for content using the Firecrawl API.
 
         Args:
-            query (str): The search query.
-            params (Optional[Dict[str, Any]]): Additional parameters for the search request.
+            query (str): The search query string.
+            params (Optional[Union[Dict[str, Any], SearchParams]]): Additional search parameters.
 
         Returns:
-            Any: The search results if the request is successful.
-
-        Raises:
-            NotImplementedError: If the search request is attempted on API version v1.
-            Exception: If the search request fails.
+            Dict[str, Any]: The search response containing success status and search results.
         """
-        raise NotImplementedError("Search is not supported in v1.")
+        if params is None:
+            params = {}
+
+        if isinstance(params, dict):
+            search_params = SearchParams(query=query, **params)
+        else:
+            search_params = params
+            search_params.query = query
+
+        response = requests.post(
+            f"{self.api_url}/v1/search",
+            headers={"Authorization": f"Bearer {self.api_key}"},
+            json=search_params.dict(exclude_none=True)
+        )
+
+        if response.status_code != 200:
+            raise Exception(f"Request failed with status code {response.status_code}")
+
+        return response.json()
 
     def crawl_url(self, url: str,
                   params: Optional[Dict[str, Any]] = None,
@@ -472,20 +507,24 @@ class FirecrawlApp:
         if not params or not params.get('prompt'):
             raise ValueError("Prompt is required")
 
-        if not params.get('schema'):
-            raise ValueError("Schema is required for extraction")
+        schema = params.get('schema')
+        if schema:
+            if hasattr(schema, 'model_json_schema'):
+                # Convert Pydantic model to JSON schema
+                schema = schema.model_json_schema()
+            # Otherwise assume it's already a JSON schema dict
 
         jsonData = {'urls': urls, **params}
-        jsonSchema = params['schema'].schema() if hasattr(params['schema'], 'schema') else None
+        request_data = {
+            **jsonData,
+            'allowExternalLinks': params.get('allow_external_links', False),
+            'schema': schema
+        }
 
         try:
             response = self._post_request(
                 f'{self.api_url}/v1/extract',
-                {
-                    **jsonData,
-                    'allowExternalLinks': params.get('allow_external_links', False),
-                    'schema': jsonSchema
-                },
+                request_data,
                 headers
             )
             if response.status_code == 200:

{firecrawl-1.7.0.dist-info → firecrawl-1.8.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: firecrawl
-Version: 1.7.0
+Version: 1.8.0
 Summary: Python SDK for Firecrawl API
 Home-page: https://github.com/mendableai/firecrawl
 Author: Mendable.ai

firecrawl-1.8.0.dist-info/RECORD

@@ -0,0 +1,11 @@
+firecrawl/__init__.py,sha256=3jDnDwAg-3SU8XRq2E8HWtJ0Umi4PLKGf4JEsR7ESig,2543
+firecrawl/firecrawl.py,sha256=0l5WOmiy5OxEwZvgIS0TpsFx39F3X6zADjHMzg6Q8iI,31650
+firecrawl/__tests__/e2e_withAuth/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+firecrawl/__tests__/e2e_withAuth/test.py,sha256=6OawnVF4IPeGyXg_Izi3t8U7MyT90roaJBJIG5UfllM,7935
+firecrawl/__tests__/v1/e2e_withAuth/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+firecrawl/__tests__/v1/e2e_withAuth/test.py,sha256=tL5kJJ4el37Wc-Z2TRSuSWwWG2M40h3VPxHYuWijD00,19888
+firecrawl-1.8.0.dist-info/LICENSE,sha256=nPCunEDwjRGHlmjvsiDUyIWbkqqyj3Ej84ntnh0g0zA,1084
+firecrawl-1.8.0.dist-info/METADATA,sha256=FCEt8ZVtXgyaGc2bNXssb2AC4hLDVI-LPoa-Qa_s2cM,10631
+firecrawl-1.8.0.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
+firecrawl-1.8.0.dist-info/top_level.txt,sha256=jTvz79zWhiyAezfmmHe4FQ-hR60C59UU5FrjMjijLu8,10
+firecrawl-1.8.0.dist-info/RECORD,,

firecrawl-1.7.0.dist-info/RECORD

@@ -1,11 +0,0 @@
-firecrawl/__init__.py,sha256=tgTkQWBxVVdPVtsIRHuyzQtuCpMij1NL-JNNpZY8dNM,2543
-firecrawl/firecrawl.py,sha256=XgDT01kF0t4TpGUvsA2HFgSg9-EGRsy0YDGkfo11Iyw,30496
-firecrawl/__tests__/e2e_withAuth/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-firecrawl/__tests__/e2e_withAuth/test.py,sha256=6OawnVF4IPeGyXg_Izi3t8U7MyT90roaJBJIG5UfllM,7935
-firecrawl/__tests__/v1/e2e_withAuth/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-firecrawl/__tests__/v1/e2e_withAuth/test.py,sha256=Qad0xRPboRdlH6Q5o2985b4xjpjw2jr9LCik-GbXaZ0,17470
-firecrawl-1.7.0.dist-info/LICENSE,sha256=nPCunEDwjRGHlmjvsiDUyIWbkqqyj3Ej84ntnh0g0zA,1084
-firecrawl-1.7.0.dist-info/METADATA,sha256=miIvKFogOdrN3wiPnATMJD_jmrWk-8O5FS57r9op28g,10631
-firecrawl-1.7.0.dist-info/WHEEL,sha256=2wepM1nk4DS4eFpYrW1TTqPcoGNfHhhO_i5m4cOimbo,92
-firecrawl-1.7.0.dist-info/top_level.txt,sha256=jTvz79zWhiyAezfmmHe4FQ-hR60C59UU5FrjMjijLu8,10
-firecrawl-1.7.0.dist-info/RECORD,,