chore: update firecrawl scrape to V1 api (#8367)

api/core/tools/provider/builtin/firecrawl/firecrawl_appx.py

@@ -37,9 +37,8 @@ class FirecrawlApp:
         for i in range(retries):
             try:
                 response = requests.request(method, url, json=data, headers=headers)
-                response.raise_for_status()
                 return response.json()
-            except requests.exceptions.RequestException as e:
+            except requests.exceptions.RequestException:
                 if i < retries - 1:
                     time.sleep(backoff_factor * (2**i))
                 else:
@@ -47,7 +46,7 @@ class FirecrawlApp:
         return None
 
     def scrape_url(self, url: str, **kwargs):
-        endpoint = f"{self.base_url}/v0/scrape"
+        endpoint = f"{self.base_url}/v1/scrape"
         data = {"url": url, **kwargs}
         logger.debug(f"Sent request to {endpoint=} body={data}")
         response = self._request("POST", endpoint, data)
@@ -55,39 +54,41 @@ class FirecrawlApp:
             raise HTTPError("Failed to scrape URL after multiple retries")
         return response
 
-    def search(self, query: str, **kwargs):
+    def map(self, url: str, **kwargs):
-        endpoint = f"{self.base_url}/v0/search"
+        endpoint = f"{self.base_url}/v1/map"
-        data = {"query": query, **kwargs}
+        data = {"url": url, **kwargs}
         logger.debug(f"Sent request to {endpoint=} body={data}")
         response = self._request("POST", endpoint, data)
         if response is None:
-            raise HTTPError("Failed to perform search after multiple retries")
+            raise HTTPError("Failed to perform map after multiple retries")
         return response
 
     def crawl_url(
         self, url: str, wait: bool = True, poll_interval: int = 5, idempotency_key: str | None = None, **kwargs
     ):
-        endpoint = f"{self.base_url}/v0/crawl"
+        endpoint = f"{self.base_url}/v1/crawl"
         headers = self._prepare_headers(idempotency_key)
         data = {"url": url, **kwargs}
         logger.debug(f"Sent request to {endpoint=} body={data}")
         response = self._request("POST", endpoint, data, headers)
         if response is None:
             raise HTTPError("Failed to initiate crawl after multiple retries")
-        job_id: str = response["jobId"]
+        elif response.get("success") == False:
+            raise HTTPError(f'Failed to crawl: {response.get("error")}')
+        job_id: str = response["id"]
         if wait:
             return self._monitor_job_status(job_id=job_id, poll_interval=poll_interval)
         return response
 
     def check_crawl_status(self, job_id: str):
-        endpoint = f"{self.base_url}/v0/crawl/status/{job_id}"
+        endpoint = f"{self.base_url}/v1/crawl/{job_id}"
         response = self._request("GET", endpoint)
         if response is None:
             raise HTTPError(f"Failed to check status for job {job_id} after multiple retries")
         return response
 
     def cancel_crawl_job(self, job_id: str):
-        endpoint = f"{self.base_url}/v0/crawl/cancel/{job_id}"
+        endpoint = f"{self.base_url}/v1/crawl/{job_id}"
         response = self._request("DELETE", endpoint)
         if response is None:
             raise HTTPError(f"Failed to cancel job {job_id} after multiple retries")
@@ -116,6 +117,6 @@ def get_json_params(tool_parameters: dict[str, Any], key):
             # support both single quotes and double quotes
             param = param.replace("'", '"')
             param = json.loads(param)
-        except:
+        except Exception:
             raise ValueError(f"Invalid {key} format.")
         return param

api/core/tools/provider/builtin/firecrawl/tools/crawl.py

@@ -8,39 +8,38 @@ from core.tools.tool.builtin_tool import BuiltinTool
 class CrawlTool(BuiltinTool):
     def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
         """
-        the crawlerOptions and pageOptions comes from doc here:
+        the api doc:
         https://docs.firecrawl.dev/api-reference/endpoint/crawl
         """
         app = FirecrawlApp(
             api_key=self.runtime.credentials["firecrawl_api_key"], base_url=self.runtime.credentials["base_url"]
         )
-        crawlerOptions = {}
+
-        pageOptions = {}
+        scrapeOptions = {}
+        payload = {}
 
         wait_for_results = tool_parameters.get("wait_for_results", True)
 
-        crawlerOptions["excludes"] = get_array_params(tool_parameters, "excludes")
+        payload["excludePaths"] = get_array_params(tool_parameters, "excludePaths")
-        crawlerOptions["includes"] = get_array_params(tool_parameters, "includes")
+        payload["includePaths"] = get_array_params(tool_parameters, "includePaths")
-        crawlerOptions["returnOnlyUrls"] = tool_parameters.get("returnOnlyUrls", False)
+        payload["maxDepth"] = tool_parameters.get("maxDepth")
-        crawlerOptions["maxDepth"] = tool_parameters.get("maxDepth")
+        payload["ignoreSitemap"] = tool_parameters.get("ignoreSitemap", False)
-        crawlerOptions["mode"] = tool_parameters.get("mode")
+        payload["limit"] = tool_parameters.get("limit", 5)
-        crawlerOptions["ignoreSitemap"] = tool_parameters.get("ignoreSitemap", False)
+        payload["allowBackwardLinks"] = tool_parameters.get("allowBackwardLinks", False)
-        crawlerOptions["limit"] = tool_parameters.get("limit", 5)
+        payload["allowExternalLinks"] = tool_parameters.get("allowExternalLinks", False)
-        crawlerOptions["allowBackwardCrawling"] = tool_parameters.get("allowBackwardCrawling", False)
+        payload["webhook"] = tool_parameters.get("webhook")
-        crawlerOptions["allowExternalContentLinks"] = tool_parameters.get("allowExternalContentLinks", False)
+
-
+        scrapeOptions["formats"] = get_array_params(tool_parameters, "formats")
-        pageOptions["headers"] = get_json_params(tool_parameters, "headers")
+        scrapeOptions["headers"] = get_json_params(tool_parameters, "headers")
-        pageOptions["includeHtml"] = tool_parameters.get("includeHtml", False)
+        scrapeOptions["includeTags"] = get_array_params(tool_parameters, "includeTags")
-        pageOptions["includeRawHtml"] = tool_parameters.get("includeRawHtml", False)
+        scrapeOptions["excludeTags"] = get_array_params(tool_parameters, "excludeTags")
-        pageOptions["onlyIncludeTags"] = get_array_params(tool_parameters, "onlyIncludeTags")
+        scrapeOptions["onlyMainContent"] = tool_parameters.get("onlyMainContent", False)
-        pageOptions["removeTags"] = get_array_params(tool_parameters, "removeTags")
+        scrapeOptions["waitFor"] = tool_parameters.get("waitFor", 0)
-        pageOptions["onlyMainContent"] = tool_parameters.get("onlyMainContent", False)
+        scrapeOptions = {k: v for k, v in scrapeOptions.items() if v not in (None, "")}
-        pageOptions["replaceAllPathsWithAbsolutePaths"] = tool_parameters.get("replaceAllPathsWithAbsolutePaths", False)
+        payload["scrapeOptions"] = scrapeOptions or None
-        pageOptions["screenshot"] = tool_parameters.get("screenshot", False)
+
-        pageOptions["waitFor"] = tool_parameters.get("waitFor", 0)
+        payload = {k: v for k, v in payload.items() if v not in (None, "")}
-
+
-        crawl_result = app.crawl_url(
+        crawl_result = app.crawl_url(url=tool_parameters["url"], wait=wait_for_results, **payload)
-            url=tool_parameters["url"], wait=wait_for_results, crawlerOptions=crawlerOptions, pageOptions=pageOptions
-        )
 
         return self.create_json_message(crawl_result)

api/core/tools/provider/builtin/firecrawl/tools/crawl.yaml

@@ -31,44 +31,33 @@ parameters:
       en_US: If you choose not to wait, it will directly return a job ID. You can use this job ID to check the crawling results or cancel the crawling task, which is usually very useful for a large-scale crawling task.
       zh_Hans: 如果选择不等待，则会直接返回一个job_id，可以通过job_id查询爬取结果或取消爬取任务，这通常对于一个大型爬取任务来说非常有用。
     form: form
-############## Crawl Options #######################
+############## Payload #######################
-  - name: includes
+  - name: excludePaths
     type: string
-    required: false
     label:
-      en_US: URL patterns to include
+      en_US: URL patterns to exclude
-      zh_Hans: 要包含的URL模式
+      zh_Hans: 要排除的URL模式
     placeholder:
       en_US: Use commas to separate multiple tags
       zh_Hans: 多个标签时使用半角逗号分隔
     human_description:
       en_US: |
-       Only pages matching these patterns will be crawled. Example: blog/*, about/*
+        Pages matching these patterns will be skipped. Example: blog/*, about/*
-      zh_Hans: 只有与这些模式匹配的页面才会被爬取。示例：blog/*, about/*
+      zh_Hans: 匹配这些模式的页面将被跳过。示例：blog/*, about/*
     form: form
-  - name: excludes
+  - name: includePaths
     type: string
+    required: false
     label:
-      en_US: URL patterns to exclude
+      en_US: URL patterns to include
-      zh_Hans: 要排除的URL模式
+      zh_Hans: 要包含的URL模式
     placeholder:
       en_US: Use commas to separate multiple tags
       zh_Hans: 多个标签时使用半角逗号分隔
     human_description:
       en_US: |
-       Pages matching these patterns will be skipped. Example: blog/*, about/*
+       Only pages matching these patterns will be crawled. Example: blog/*, about/*
-      zh_Hans: 匹配这些模式的页面将被跳过。示例：blog/*, about/*
+      zh_Hans: 只有与这些模式匹配的页面才会被爬取。示例：blog/*, about/*
-    form: form
-  - name: returnOnlyUrls
-    type: boolean
-    default: false
-    label:
-      en_US: return Only Urls
-      zh_Hans: 仅返回URL
-    human_description:
-      en_US: |
-       If true, returns only the URLs as a list on the crawl status. Attention: the return response will be a list of URLs inside the data, not a list of documents.
-      zh_Hans: 只返回爬取到的网页链接，而不是网页内容本身。
     form: form
   - name: maxDepth
     type: number
@@ -80,27 +69,10 @@ parameters:
       zh_Hans: 相对于输入的URL，爬取的最大深度。maxDepth为0时，仅抓取输入的URL。maxDepth为1时，抓取输入的URL以及所有一级深层页面。maxDepth为2时，抓取输入的URL以及所有两级深层页面。更高值遵循相同模式。
     form: form
     min: 0
-  - name: mode
+    default: 2
-    type: select
-    required: false
-    form: form
-    options:
-      - value: default
-        label:
-          en_US: default
-      - value: fast
-        label:
-          en_US: fast
-    default: default
-    label:
-      en_US: Crawl Mode
-      zh_Hans: 爬取模式
-    human_description:
-      en_US: The crawling mode to use. Fast mode crawls 4x faster websites without sitemap, but may not be as accurate and shouldn't be used in heavy js-rendered websites.
-      zh_Hans: 使用fast模式将不会使用其站点地图，比普通模式快4倍，但是可能不够准确，也不适用于大量js渲染的网站。
   - name: ignoreSitemap
     type: boolean
-    default: false
+    default: true
     label:
       en_US: ignore Sitemap
       zh_Hans: 忽略站点地图
@@ -120,7 +92,7 @@ parameters:
     form: form
     min: 1
     default: 5
-  - name: allowBackwardCrawling
+  - name: allowBackwardLinks
     type: boolean
     default: false
     label:
@@ -130,7 +102,7 @@ parameters:
       en_US: Enables the crawler to navigate from a specific URL to previously linked pages. For instance, from 'example.com/product/123' back to 'example.com/product'
       zh_Hans: 使爬虫能够从特定URL导航到之前链接的页面。例如，从'example.com/product/123'返回到'example.com/product'
     form: form
-  - name: allowExternalContentLinks
+  - name: allowExternalLinks
     type: boolean
     default: false
     label:
@@ -140,7 +112,30 @@ parameters:
       en_US: Allows the crawler to follow links to external websites.
       zh_Hans:
     form: form
-############## Page Options #######################
+  - name: webhook
+    type: string
+    label:
+      en_US: webhook
+    human_description:
+      en_US: |
+       The URL to send the webhook to. This will trigger for crawl started (crawl.started) ,every page crawled (crawl.page) and when the crawl is completed (crawl.completed or crawl.failed). The response will be the same as the /scrape endpoint.
+      zh_Hans: 发送Webhook的URL。这将在开始爬取（crawl.started）、每爬取一个页面（crawl.page）以及爬取完成（crawl.completed或crawl.failed）时触发。响应将与/scrape端点相同。
+    form: form
+############## Scrape Options #######################
+  - name: formats
+    type: string
+    label:
+      en_US: Formats
+      zh_Hans: 结果的格式
+    placeholder:
+      en_US: Use commas to separate multiple tags
+      zh_Hans: 多个标签时使用半角逗号分隔
+    human_description:
+      en_US: |
+        Formats to include in the output. Available options: markdown, html, rawHtml, links, screenshot
+      zh_Hans: |
+        输出中应包含的格式。可以填入: markdown, html, rawHtml, links, screenshot
+    form: form
   - name: headers
     type: string
     label:
@@ -155,30 +150,10 @@ parameters:
       en_US: Please enter an object that can be serialized in JSON
       zh_Hans: 请输入可以json序列化的对象
     form: form
-  - name: includeHtml
+  - name: includeTags
-    type: boolean
-    default: false
-    label:
-      en_US: include Html
-      zh_Hans: 包含HTML
-    human_description:
-      en_US: Include the HTML version of the content on page. Will output a html key in the response.
-      zh_Hans: 返回中包含一个HTML版本的内容，将以html键返回。
-    form: form
-  - name: includeRawHtml
-    type: boolean
-    default: false
-    label:
-      en_US: include Raw Html
-      zh_Hans: 包含原始HTML
-    human_description:
-      en_US: Include the raw HTML content of the page. Will output a rawHtml key in the response.
-      zh_Hans: 返回中包含一个原始HTML版本的内容，将以rawHtml键返回。
-    form: form
-  - name: onlyIncludeTags
     type: string
     label:
-      en_US: only Include Tags
+      en_US: Include Tags
       zh_Hans: 仅抓取这些标签
     placeholder:
       en_US: Use commas to separate multiple tags
@@ -189,20 +164,10 @@ parameters:
       zh_Hans: |
         仅在最终输出中包含HTML页面的这些标签，可以通过标签名、类或ID来设定，使用逗号分隔值。示例：script, .ad, #footer
     form: form
-  - name: onlyMainContent
+  - name: excludeTags
-    type: boolean
-    default: false
-    label:
-      en_US: only Main Content
-      zh_Hans: 仅抓取主要内容
-    human_description:
-      en_US: Only return the main content of the page excluding headers, navs, footers, etc.
-      zh_Hans: 只返回页面的主要内容，不包括头部、导航栏、尾部等。
-    form: form
-  - name: removeTags
     type: string
     label:
-      en_US: remove Tags
+      en_US: Exclude Tags
       zh_Hans: 要移除这些标签
     human_description:
       en_US: |
@@ -213,25 +178,15 @@ parameters:
       en_US: Use commas to separate multiple tags
       zh_Hans: 多个标签时使用半角逗号分隔
     form: form
-  - name: replaceAllPathsWithAbsolutePaths
+  - name: onlyMainContent
-    type: boolean
-    default: false
-    label:
-      en_US: All AbsolutePaths
-      zh_Hans: 使用绝对路径
-    human_description:
-      en_US: Replace all relative paths with absolute paths for images and links.
-      zh_Hans: 将所有图片和链接的相对路径替换为绝对路径。
-    form: form
-  - name: screenshot
     type: boolean
     default: false
     label:
-      en_US: screenshot
+      en_US: only Main Content
-      zh_Hans: 截图
+      zh_Hans: 仅抓取主要内容
     human_description:
-      en_US: Include a screenshot of the top of the page that you are scraping.
+      en_US: Only return the main content of the page excluding headers, navs, footers, etc.
-      zh_Hans: 提供正在抓取的页面的顶部的截图。
+      zh_Hans: 只返回页面的主要内容，不包括头部、导航栏、尾部等。
     form: form
   - name: waitFor
     type: number

api/core/tools/provider/builtin/firecrawl/tools/map.py

@@ -0,0 +1,25 @@
+from typing import Any
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.provider.builtin.firecrawl.firecrawl_appx import FirecrawlApp
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class MapTool(BuiltinTool):
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
+        """
+        the api doc:
+        https://docs.firecrawl.dev/api-reference/endpoint/map
+        """
+        app = FirecrawlApp(
+            api_key=self.runtime.credentials["firecrawl_api_key"], base_url=self.runtime.credentials["base_url"]
+        )
+        payload = {}
+        payload["search"] = tool_parameters.get("search")
+        payload["ignoreSitemap"] = tool_parameters.get("ignoreSitemap", True)
+        payload["includeSubdomains"] = tool_parameters.get("includeSubdomains", False)
+        payload["limit"] = tool_parameters.get("limit", 5000)
+
+        map_result = app.map(url=tool_parameters["url"], **payload)
+
+        return self.create_json_message(map_result)

api/core/tools/provider/builtin/firecrawl/tools/map.yaml

@@ -0,0 +1,59 @@
+identity:
+  name: map
+  author: hjlarry
+  label:
+    en_US: Map
+    zh_Hans: 地图式快爬
+description:
+  human:
+    en_US: Input a website and get all the urls on the website - extremly fast
+    zh_Hans: 输入一个网站，快速获取网站上的所有网址。
+  llm: Input a website and get all the urls on the website - extremly fast
+parameters:
+  - name: url
+    type: string
+    required: true
+    label:
+      en_US: Start URL
+      zh_Hans: 起始URL
+    human_description:
+      en_US: The base URL to start crawling from.
+      zh_Hans: 要爬取网站的起始URL。
+    llm_description: The URL of the website that needs to be crawled. This is a required parameter.
+    form: llm
+  - name: search
+    type: string
+    label:
+      en_US: search
+      zh_Hans: 搜索查询
+    human_description:
+      en_US: Search query to use for mapping. During the Alpha phase, the 'smart' part of the search functionality is limited to 100 search results. However, if map finds more results, there is no limit applied.
+      zh_Hans: 用于映射的搜索查询。在Alpha阶段，搜索功能的“智能”部分限制为最多100个搜索结果。然而，如果地图找到了更多结果，则不施加任何限制。
+    llm_description: Search query to use for mapping. During the Alpha phase, the 'smart' part of the search functionality is limited to 100 search results. However, if map finds more results, there is no limit applied.
+    form: llm
+############## Page Options #######################
+  - name: ignoreSitemap
+    type: boolean
+    default: true
+    label:
+      en_US: ignore Sitemap
+      zh_Hans: 忽略站点地图
+    human_description:
+      en_US: Ignore the website sitemap when crawling.
+      zh_Hans: 爬取时忽略网站站点地图。
+    form: form
+  - name: includeSubdomains
+    type: boolean
+    default: false
+    label:
+      en_US: include Subdomains
+      zh_Hans: 包含子域名
+    form: form
+  - name: limit
+    type: number
+    min: 0
+    default: 5000
+    label:
+      en_US: Maximum results
+      zh_Hans: 最大结果数量
+    form: form

api/core/tools/provider/builtin/firecrawl/tools/scrape.py

@@ -6,34 +6,34 @@ from core.tools.tool.builtin_tool import BuiltinTool
 
 
 class ScrapeTool(BuiltinTool):
-    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> list[ToolInvokeMessage]:
         """
-        the pageOptions and extractorOptions comes from doc here:
+        the api doc:
         https://docs.firecrawl.dev/api-reference/endpoint/scrape
         """
         app = FirecrawlApp(
             api_key=self.runtime.credentials["firecrawl_api_key"], base_url=self.runtime.credentials["base_url"]
         )
 
-        pageOptions = {}
+        payload = {}
-        extractorOptions = {}
+        extract = {}
-
+
-        pageOptions["headers"] = get_json_params(tool_parameters, "headers")
+        payload["formats"] = get_array_params(tool_parameters, "formats")
-        pageOptions["includeHtml"] = tool_parameters.get("includeHtml", False)
+        payload["onlyMainContent"] = tool_parameters.get("onlyMainContent", True)
-        pageOptions["includeRawHtml"] = tool_parameters.get("includeRawHtml", False)
+        payload["includeTags"] = get_array_params(tool_parameters, "includeTags")
-        pageOptions["onlyIncludeTags"] = get_array_params(tool_parameters, "onlyIncludeTags")
+        payload["excludeTags"] = get_array_params(tool_parameters, "excludeTags")
-        pageOptions["removeTags"] = get_array_params(tool_parameters, "removeTags")
+        payload["headers"] = get_json_params(tool_parameters, "headers")
-        pageOptions["onlyMainContent"] = tool_parameters.get("onlyMainContent", False)
+        payload["waitFor"] = tool_parameters.get("waitFor", 0)
-        pageOptions["replaceAllPathsWithAbsolutePaths"] = tool_parameters.get("replaceAllPathsWithAbsolutePaths", False)
+        payload["timeout"] = tool_parameters.get("timeout", 30000)
-        pageOptions["screenshot"] = tool_parameters.get("screenshot", False)
+
-        pageOptions["waitFor"] = tool_parameters.get("waitFor", 0)
+        extract["schema"] = get_json_params(tool_parameters, "schema")
-
+        extract["systemPrompt"] = tool_parameters.get("systemPrompt")
-        extractorOptions["mode"] = tool_parameters.get("mode", "")
+        extract["prompt"] = tool_parameters.get("prompt")
-        extractorOptions["extractionPrompt"] = tool_parameters.get("extractionPrompt", "")
+        extract = {k: v for k, v in extract.items() if v not in (None, "")}
-        extractorOptions["extractionSchema"] = get_json_params(tool_parameters, "extractionSchema")
+        payload["extract"] = extract or None
-
+
-        crawl_result = app.scrape_url(
+        payload = {k: v for k, v in payload.items() if v not in (None, "")}
-            url=tool_parameters["url"], pageOptions=pageOptions, extractorOptions=extractorOptions
-        )
 
-        return self.create_json_message(crawl_result)
+        crawl_result = app.scrape_url(url=tool_parameters["url"], **payload)
+        markdown_result = crawl_result.get("data", {}).get("markdown", "")
+        return [self.create_text_message(markdown_result), self.create_json_message(crawl_result)]

api/core/tools/provider/builtin/firecrawl/tools/scrape.yaml

@@ -6,8 +6,8 @@ identity:
     zh_Hans: 单页面抓取
 description:
   human:
-    en_US: Extract data from a single URL.
+    en_US: Turn any url into clean data.
-    zh_Hans: 从单个URL抓取数据。
+    zh_Hans: 将任何网址转换为干净的数据。
   llm: This tool is designed to scrape URL and output the content in Markdown format.
 parameters:
   - name: url
@@ -21,45 +21,35 @@ parameters:
       zh_Hans: 要抓取并提取数据的网站URL。
     llm_description: The URL of the website that needs to be crawled. This is a required parameter.
     form: llm
-############## Page Options #######################
+############## Payload #######################
-  - name: headers
+  - name: formats
     type: string
     label:
-      en_US: headers
+      en_US: Formats
-      zh_Hans: 请求头
+      zh_Hans: 结果的格式
+    placeholder:
+      en_US: Use commas to separate multiple tags
+      zh_Hans: 多个标签时使用半角逗号分隔
     human_description:
       en_US: |
-        Headers to send with the request. Can be used to send cookies, user-agent, etc. Example: {"cookies": "testcookies"}
+        Formats to include in the output. Available options: markdown, html, rawHtml, links, screenshot, extract, screenshot@fullPage
       zh_Hans: |
-        随请求发送的头部。可以用来发送cookies、用户代理等。示例：{"cookies": "testcookies"}
+        输出中应包含的格式。可以填入: markdown, html, rawHtml, links, screenshot, extract, screenshot@fullPage
-    placeholder:
-      en_US: Please enter an object that can be serialized in JSON
-      zh_Hans: 请输入可以json序列化的对象
     form: form
-  - name: includeHtml
+  - name: onlyMainContent
-    type: boolean
-    default: false
-    label:
-      en_US: include Html
-      zh_Hans: 包含HTML
-    human_description:
-      en_US: Include the HTML version of the content on page. Will output a html key in the response.
-      zh_Hans: 返回中包含一个HTML版本的内容，将以html键返回。
-    form: form
-  - name: includeRawHtml
     type: boolean
     default: false
     label:
-      en_US: include Raw Html
+      en_US: only Main Content
-      zh_Hans: 包含原始HTML
+      zh_Hans: 仅抓取主要内容
     human_description:
-      en_US: Include the raw HTML content of the page. Will output a rawHtml key in the response.
+      en_US: Only return the main content of the page excluding headers, navs, footers, etc.
-      zh_Hans: 返回中包含一个原始HTML版本的内容，将以rawHtml键返回。
+      zh_Hans: 只返回页面的主要内容，不包括头部、导航栏、尾部等。
     form: form
-  - name: onlyIncludeTags
+  - name: includeTags
     type: string
     label:
-      en_US: only Include Tags
+      en_US: Include Tags
       zh_Hans: 仅抓取这些标签
     placeholder:
       en_US: Use commas to separate multiple tags
@@ -70,20 +60,10 @@ parameters:
       zh_Hans: |
         仅在最终输出中包含HTML页面的这些标签，可以通过标签名、类或ID来设定，使用逗号分隔值。示例：script, .ad, #footer
     form: form
-  - name: onlyMainContent
+  - name: excludeTags
-    type: boolean
-    default: false
-    label:
-      en_US: only Main Content
-      zh_Hans: 仅抓取主要内容
-    human_description:
-      en_US: Only return the main content of the page excluding headers, navs, footers, etc.
-      zh_Hans: 只返回页面的主要内容，不包括头部、导航栏、尾部等。
-    form: form
-  - name: removeTags
     type: string
     label:
-      en_US: remove Tags
+      en_US: Exclude Tags
       zh_Hans: 要移除这些标签
     human_description:
       en_US: |
@@ -94,29 +74,24 @@ parameters:
       en_US: Use commas to separate multiple tags
       zh_Hans: 多个标签时使用半角逗号分隔
     form: form
-  - name: replaceAllPathsWithAbsolutePaths
+  - name: headers
-    type: boolean
+    type: string
-    default: false
-    label:
-      en_US: All AbsolutePaths
-      zh_Hans: 使用绝对路径
-    human_description:
-      en_US: Replace all relative paths with absolute paths for images and links.
-      zh_Hans: 将所有图片和链接的相对路径替换为绝对路径。
-    form: form
-  - name: screenshot
-    type: boolean
-    default: false
     label:
-      en_US: screenshot
+      en_US: headers
-      zh_Hans: 截图
+      zh_Hans: 请求头
     human_description:
-      en_US: Include a screenshot of the top of the page that you are scraping.
+      en_US: |
-      zh_Hans: 提供正在抓取的页面的顶部的截图。
+        Headers to send with the request. Can be used to send cookies, user-agent, etc. Example: {"cookies": "testcookies"}
+      zh_Hans: |
+        随请求发送的头部。可以用来发送cookies、用户代理等。示例：{"cookies": "testcookies"}
+    placeholder:
+      en_US: Please enter an object that can be serialized in JSON
+      zh_Hans: 请输入可以json序列化的对象
     form: form
   - name: waitFor
     type: number
     min: 0
+    default: 0
     label:
       en_US: wait For
       zh_Hans: 等待时间
@@ -124,57 +99,54 @@ parameters:
       en_US: Wait x amount of milliseconds for the page to load to fetch content.
       zh_Hans: 等待x毫秒以使页面加载并获取内容。
     form: form
-############## Extractor Options #######################
+  - name: timeout
-  - name: mode
+    type: number
-    type: select
+    min: 0
-    options:
+    default: 30000
-      - value: markdown
-        label:
-          en_US: markdown
-      - value: llm-extraction
-        label:
-          en_US: llm-extraction
-      - value: llm-extraction-from-raw-html
-        label:
-          en_US: llm-extraction-from-raw-html
-      - value: llm-extraction-from-markdown
-        label:
-          en_US: llm-extraction-from-markdown
-    label:
-      en_US: Extractor Mode
-      zh_Hans: 提取模式
-    human_description:
-      en_US: |
-       The extraction mode to use. 'markdown': Returns the scraped markdown content, does not perform LLM extraction. 'llm-extraction': Extracts information from the cleaned and parsed content using LLM.
-      zh_Hans: 使用的提取模式。“markdown”：返回抓取的markdown内容，不执行LLM提取。“llm-extractioin”：使用LLM按Extractor Schema从内容中提取信息。
-    form: form
-  - name: extractionPrompt
-    type: string
     label:
-      en_US: Extractor Prompt
+      en_US: Timeout
-      zh_Hans: 提取时的提示词
     human_description:
-      en_US: A prompt describing what information to extract from the page, applicable for LLM extraction modes.
+      en_US: Timeout in milliseconds for the request.
-      zh_Hans: 当使用LLM提取模式时，用于给LLM描述提取规则。
+      zh_Hans: 请求的超时时间（以毫秒为单位）。
     form: form
-  - name: extractionSchema
+############## Extractor Options #######################
+  - name: schema
     type: string
     label:
       en_US: Extractor Schema
       zh_Hans: 提取时的结构
     placeholder:
       en_US: Please enter an object that can be serialized in JSON
+      zh_Hans: 请输入可以json序列化的对象
     human_description:
       en_US: |
-        The schema for the data to be extracted, required only for LLM extraction modes. Example: {
+        The schema for the data to be extracted. Example: {
             "type": "object",
             "properties": {"company_mission": {"type": "string"}},
             "required": ["company_mission"]
         }
       zh_Hans: |
-       当使用LLM提取模式时，使用该结构去提取，示例：{
+       使用该结构去提取，示例：{
             "type": "object",
             "properties": {"company_mission": {"type": "string"}},
             "required": ["company_mission"]
         }
     form: form
+  - name: systemPrompt
+    type: string
+    label:
+      en_US: Extractor System Prompt
+      zh_Hans: 提取时的系统提示词
+    human_description:
+      en_US: The system prompt to use for the extraction.
+      zh_Hans: 用于提取的系统提示。
+    form: form
+  - name: prompt
+    type: string
+    label:
+      en_US: Extractor Prompt
+      zh_Hans: 提取时的提示词
+    human_description:
+      en_US: The prompt to use for the extraction without a schema.
+      zh_Hans: 用于无schema时提取的提示词
+    form: form

api/core/tools/provider/builtin/firecrawl/tools/search.py

@@ -1,27 +0,0 @@
-from typing import Any
-
-from core.tools.entities.tool_entities import ToolInvokeMessage
-from core.tools.provider.builtin.firecrawl.firecrawl_appx import FirecrawlApp
-from core.tools.tool.builtin_tool import BuiltinTool
-
-
-class SearchTool(BuiltinTool):
-    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
-        """
-        the pageOptions and searchOptions comes from doc here:
-        https://docs.firecrawl.dev/api-reference/endpoint/search
-        """
-        app = FirecrawlApp(
-            api_key=self.runtime.credentials["firecrawl_api_key"], base_url=self.runtime.credentials["base_url"]
-        )
-        pageOptions = {}
-        pageOptions["onlyMainContent"] = tool_parameters.get("onlyMainContent", False)
-        pageOptions["fetchPageContent"] = tool_parameters.get("fetchPageContent", True)
-        pageOptions["includeHtml"] = tool_parameters.get("includeHtml", False)
-        pageOptions["includeRawHtml"] = tool_parameters.get("includeRawHtml", False)
-        searchOptions = {"limit": tool_parameters.get("limit")}
-        search_result = app.search(
-            query=tool_parameters["keyword"], pageOptions=pageOptions, searchOptions=searchOptions
-        )
-
-        return self.create_json_message(search_result)

api/core/tools/provider/builtin/firecrawl/tools/search.yaml

@@ -1,75 +0,0 @@
-identity:
-  name: search
-  author: ahasasjeb
-  label:
-    en_US: Search
-    zh_Hans: 搜索
-description:
-  human:
-    en_US: Search, and output in Markdown format
-    zh_Hans: 搜索，并且以Markdown格式输出
-  llm: This tool can perform online searches and convert the results to Markdown format.
-parameters:
-  - name: keyword
-    type: string
-    required: true
-    label:
-      en_US: keyword
-      zh_Hans: 关键词
-    human_description:
-      en_US: Input keywords to use Firecrawl API for search.
-      zh_Hans: 输入关键词即可使用Firecrawl API进行搜索。
-    llm_description: Efficiently extract keywords from user text.
-    form: llm
-############## Page Options #######################
-  - name: onlyMainContent
-    type: boolean
-    default: false
-    label:
-      en_US: only Main Content
-      zh_Hans: 仅抓取主要内容
-    human_description:
-      en_US: Only return the main content of the page excluding headers, navs, footers, etc.
-      zh_Hans: 只返回页面的主要内容，不包括头部、导航栏、尾部等。
-    form: form
-  - name: fetchPageContent
-    type: boolean
-    default: true
-    label:
-      en_US: fetch Page Content
-      zh_Hans: 抓取页面内容
-    human_description:
-      en_US: Fetch the content of each page. If false, defaults to a basic fast serp API.
-      zh_Hans: 获取每个页面的内容。如果为否，则使用基本的快速搜索结果页面API。
-    form: form
-  - name: includeHtml
-    type: boolean
-    default: false
-    label:
-      en_US: include Html
-      zh_Hans: 包含HTML
-    human_description:
-      en_US: Include the HTML version of the content on page. Will output a html key in the response.
-      zh_Hans: 返回中包含一个HTML版本的内容，将以html键返回。
-    form: form
-  - name: includeRawHtml
-    type: boolean
-    default: false
-    label:
-      en_US: include Raw Html
-      zh_Hans: 包含原始HTML
-    human_description:
-      en_US: Include the raw HTML content of the page. Will output a rawHtml key in the response.
-      zh_Hans: 返回中包含一个原始HTML版本的内容，将以rawHtml键返回。
-    form: form
-############## Search Options #######################
-  - name: limit
-    type: number
-    min: 0
-    label:
-      en_US: Maximum results
-      zh_Hans: 最大结果数量
-    human_description:
-      en_US: Maximum number of results. Max is 20 during beta.
-      zh_Hans: 最大结果数量。在测试阶段，最大为20。
-    form: form