Update app.py

app.py

@@ -1,36 +1,54 @@
-import pandas as pd
 import gradio as gr
+import requests
+import xml.etree.ElementTree as ET
 
-def compare_excel_files(new_file, old_file):
-    # Check if the file extensions are valid
-    if not new_file.name.endswith(".xlsx") or not old_file.name.endswith(".xlsx"):
-        raise ValueError("Invalid file type. Please upload Excel files with .xlsx extension.")
 
-    # Read the Excel files into pandas DataFrames
-    df_new = pd.read_excel(new_file.name)
-    df_old = pd.read_excel(old_file.name)
+def collect_urls_from_sitemap(sitemap_content):
+    root = ET.fromstring(sitemap_content)
 
-    # Compare new file to old file (Find missing data in old file)
-    missing_in_old = df_new[~df_new.isin(df_old)].dropna()
+    urls = []
+    for url_element in root.findall(".//{http://www.sitemaps.org/schemas/sitemap/0.9}url"):
+        loc_element = url_element.find("{http://www.sitemaps.org/schemas/sitemap/0.9}loc")
+        if loc_element is not None:
+            urls.append(loc_element.text)
 
-    # Compare old file to new file (Find added data in new file)
-    added_in_new = df_old[~df_old.isin(df_new)].dropna()
+    return urls
 
-    # Create Excel files for the comparison results
-    output_file_missing = 'missing_data_in_old.xlsx'
-    output_file_added = 'added_data_in_new.xlsx'
-    missing_in_old.to_excel(output_file_missing, index=False)
-    added_in_new.to_excel(output_file_added, index=False)
 
-    return output_file_missing, output_file_added
+def compare_sitemaps(old_urls, new_urls):
+    added_urls = list(set(new_urls) - set(old_urls))
+    removed_urls = list(set(old_urls) - set(new_urls))
+    return added_urls, removed_urls
+
+
+def compare_sitemaps_ui(old_sitemap_url, new_sitemap_url):
+    old_response = requests.get(old_sitemap_url)
+    new_response = requests.get(new_sitemap_url)
+
+    if old_response.status_code == 200 and new_response.status_code == 200:
+        old_sitemap_content = old_response.content
+        new_sitemap_content = new_response.content
+
+        old_urls = collect_urls_from_sitemap(old_sitemap_content)
+        new_urls = collect_urls_from_sitemap(new_sitemap_content)
+
+        added_urls, removed_urls = compare_sitemaps(old_urls, new_urls)
+
+        added_urls_text = "\n".join(added_urls)
+        removed_urls_text = "\n".join(removed_urls)
+
+        return f"Newly Added URLs:\n{added_urls_text}\n\nRemoved URLs:\n{removed_urls_text}"
+    else:
+        return "Failed to fetch sitemaps."
+
 
 iface = gr.Interface(
-    fn=compare_excel_files,
-    inputs=[gr.inputs.File(label="New Excel File"), gr.inputs.File(label="Old Excel File")],
-    outputs=[gr.outputs.File(label="Missing Data in Old"), gr.outputs.File(label="Added Data in New")],
-    title="Excel File Comparison",
-    description="Upload two Excel files: one with new data and the other with old data to find missing data in old and added data in new."
-  
+    fn=compare_sitemaps_ui,
+    inputs=["text", "text"],
+    outputs="text",
+    title="Sitemap Comparator",
+    description="Enter the URLs of the old and new sitemaps to compare.",
+    examples=[["https://example.com/old-sitemap.xml", "https://example.com/new-sitemap.xml"]]
 )
 
 iface.launch()