Merge branch 'main' into nsc/semantic-index-extract

.gitignore

@@ -36,3 +36,5 @@ apps/js-sdk/firecrawl/dist
 
 /apps/api/logs/*
 /apps/api/debug/*
+
+.vscode

.vscode/settings.json

@@ -1,5 +0,0 @@
-{
-    "rust-analyzer.linkedProjects": [
-        "apps/rust-sdk/Cargo.toml"
-    ]
-}

apps/api/src/scraper/scrapeURL/lib/removeUnwantedElements.ts

@@ -39,9 +39,6 @@ const excludeNonMainTags = [
   "#nav",
   ".breadcrumbs",
   "#breadcrumbs",
-  "#search-form",
-  ".search",
-  "#search",
   ".share",
   "#share",
   ".widget",

apps/python-sdk/firecrawl/__init__.py

@@ -13,7 +13,7 @@ import os
 
 from .firecrawl import FirecrawlApp # noqa
 
-__version__ = "1.7.0"
+__version__ = "1.7.1"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

apps/python-sdk/firecrawl/firecrawl.py

@@ -472,20 +472,24 @@ class FirecrawlApp:
         if not params or not params.get('prompt'):
             raise ValueError("Prompt is required")
 
-        if not params.get('schema'):
+        schema = params.get('schema')
-            raise ValueError("Schema is required for extraction")
+        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',
-                {
+                request_data,
-                    **jsonData,
-                    'allowExternalLinks': params.get('allow_external_links', False),
-                    'schema': jsonSchema
-                },
                 headers
             )
             if response.status_code == 200: