feat-SDK/added crawl id to ws

mendableai · Dec 18, 2024 · 19246f6 · 19246f6
1 parent e899ecb
commit 19246f6
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 8 deletions.
diff --git a/apps/js-sdk/firecrawl/package.json b/apps/js-sdk/firecrawl/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@mendable/firecrawl-js",
-  "version": "1.9.4",
+  "version": "1.9.5",
   "description": "JavaScript SDK for Firecrawl API",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

diff --git a/apps/js-sdk/firecrawl/src/index.ts b/apps/js-sdk/firecrawl/src/index.ts
@@ -934,9 +934,11 @@ export class CrawlWatcher extends TypedEventTarget<CrawlWatcherEvents> {
   private ws: WebSocket;
   public data: FirecrawlDocument<undefined>[];
   public status: CrawlStatusResponse["status"];
+  public id: string;
 
   constructor(id: string, app: FirecrawlApp) {
     super();
+    this.id = id;
     this.ws = new WebSocket(`${app.apiUrl}/v1/crawl/${id}`, app.apiKey);
     this.status = "scraping";
     this.data = [];
@@ -967,6 +969,7 @@ export class CrawlWatcher extends TypedEventTarget<CrawlWatcherEvents> {
           detail: {
             status: this.status,
             data: this.data,
+            id: this.id,
           },
         }));
       } else if (msg.type === "error") {
@@ -976,19 +979,26 @@ export class CrawlWatcher extends TypedEventTarget<CrawlWatcherEvents> {
             status: this.status,
             data: this.data,
             error: msg.error,
+            id: this.id,
           },
         }));
       } else if (msg.type === "catchup") {
         this.status = msg.data.status;
         this.data.push(...(msg.data.data ?? []));
         for (const doc of this.data) {
           this.dispatchTypedEvent("document", new CustomEvent("document", {
-            detail: doc,
+            detail: {
+              ...doc,
+              id: this.id,
+            },
           }));
         }
       } else if (msg.type === "document") {
         this.dispatchTypedEvent("document", new CustomEvent("document", {
-          detail: msg.data,
+          detail: {
+            ...msg.data,
+            id: this.id,
+          },
         }));
       }
     }
@@ -1015,6 +1025,7 @@ export class CrawlWatcher extends TypedEventTarget<CrawlWatcherEvents> {
           status: this.status,
           data: this.data,
           error: "WebSocket error",
+          id: this.id,
         },
       }));
     }).bind(this);

diff --git a/apps/python-sdk/firecrawl/__init__.py b/apps/python-sdk/firecrawl/__init__.py
@@ -13,7 +13,7 @@
 
 from .firecrawl import FirecrawlApp # noqa
 
-__version__ = "1.6.4"
+__version__ = "1.6.5"
 
 # Define the logger for the Firecrawl project
 logger: logging.Logger = logging.getLogger("firecrawl")

diff --git a/apps/python-sdk/firecrawl/firecrawl.py b/apps/python-sdk/firecrawl/firecrawl.py
@@ -704,15 +704,15 @@ def dispatch_event(self, event_type: str, detail: Dict[str, Any]):
     async def _handle_message(self, msg: Dict[str, Any]):
         if msg['type'] == 'done':
             self.status = 'completed'
-            self.dispatch_event('done', {'status': self.status, 'data': self.data})
+            self.dispatch_event('done', {'status': self.status, 'data': self.data, 'id': self.id})
         elif msg['type'] == 'error':
             self.status = 'failed'
-            self.dispatch_event('error', {'status': self.status, 'data': self.data, 'error': msg['error']})
+            self.dispatch_event('error', {'status': self.status, 'data': self.data, 'error': msg['error'], 'id': self.id})
         elif msg['type'] == 'catchup':
             self.status = msg['data']['status']
             self.data.extend(msg['data'].get('data', []))
             for doc in self.data:
-                self.dispatch_event('document', doc)
+                self.dispatch_event('document', {'data': doc, 'id': self.id})
         elif msg['type'] == 'document':
             self.data.append(msg['data'])
-            self.dispatch_event('document', msg['data'])
+            self.dispatch_event('document', {'data': msg['data'], 'id': self.id})