Update app.py

app.py

@@ -24,7 +24,7 @@ def link_find(url):
     
     else:
         pass
-    return [node1]
+    return node1
 #https://huggingface.co/spaces/Omnibus/crawl
 
 def sitemap(url,level):
@@ -32,7 +32,7 @@ def sitemap(url,level):
     if url != "" and url != None:    
         link1=link_find(url)
         if level >=2:
-            for i,ea in enumerate(link1):
+            for i,ea in enumerate(link1['TREE']):
                 print(ea)
                 try:
                     if not ea['URL'].startswith("http"):
@@ -44,7 +44,7 @@ def sitemap(url,level):
                     out_list=link_find(f"{uri}{ea['URL']}")
                     #link1[i]['TREE']=out_list
                     if level>=3:
-                        for n,na in enumerate(link1[i]['TREE']):
+                        for n,na in enumerate(link1['TREE'][i]['TREE']):
                             print(na)
                             try:
                                 if not na['URL'].startswith("http"):
@@ -60,6 +60,42 @@ def sitemap(url,level):
                 except Exception as e:
                     print (e)
     return link1
+
+
+
+def sitemap_OG(url,level):
+    uri=""
+    if url != "" and url != None:    
+        link1=link_find(url)
+        if level >=2:
+            for i,ea in enumerate(link1):
+                print(ea)
+                try:
+                    if not ea['URL'].startswith("http"):
+                        uri1=url.split("//")[0]
+                        uri2=url.split("//")[1]
+                        uri3=uri2.split("/")[0]
+                        uri=f'{uri1}//{uri3}'
+                        print(uri)
+                    out_list=link_find(f"{uri}{ea['URL']}")
+                    link1[i]['TREE']=out_list
+                    if level>=3:
+                        for n,na in enumerate(link1[i]['TREE']):
+                            print(na)
+                            try:
+                                if not na['URL'].startswith("http"):
+                                    uri11=url.split("//")[0]
+                                    uri22=url.split("//")[1]
+                                    uri33=uri22.split("/")[0]
+                                    uri0=f'{uri11}//{uri33}'
+                                    print(uri0)
+                                out_list1=link_find(f"{uri0}{na['URL']}")
+                                link1[i]['TREE'][n]['TREE']=out_list1
+                            except Exception as e:
+                                print (e)
+                except Exception as e:
+                    print (e)
+    return link1
 with gr.Blocks() as app:
     with gr.Row():
         inp=gr.Textbox(label="URL")