Update app.py

app.py

@@ -1,37 +1,67 @@
 import streamlit as st
-import github
-import os
+from huggingface_hub import Repository
+from github import Github
 
-def find_large_files(file_name):
-    gh = github.GitHub()
-    search_results = gh.search_repos(repo_type='public', q=f"{file_name} file://*.zip")[:5] # change to [10] to show top 10 results if you wish
-    
-    zip_files = []
-    for res in search_results:
-        try:
-            repo = gh.get_repository(res['owner']['login'], res['name'])
-            for release in repo.get_releases():
-                for asset in release.get('assets'):
-                    if 'application/zip' in asset['content_type'] and f"{os.path.basename(file_name)}" in asset['name'].lower():
-                        file_size = float(asset['downloads'].split('/')[-1].split('.')[0]) / (1 << 20)
-                        zip_files.append({'repo': res, 'size': file_size})
-                        break
-            repos_by_size = sorted(zip_files, key=lambda x:x['size'], reverse=True)
-        except Exception as e:
-            print("Error while searching files in repository %s: %s" % (res['full_name'], str(e)))
-    return repos_by_size
+# Set up the Streamlit app
+st.set_page_config(page_title="GitHub File Search", page_icon=":mag_right:")
 
-st.write("Enter filename including extension: ")
-file_input = st.text_input("File Name", "test.txt")
-if ".zip" not in file_input:
-    st.write("Please enter a valid file name.")
-else:
-    st.write("Searching...")
-    repos = find_large_files(file_input)
-    if len(repos) == 0:
-        st.write("No matching files found.")
-    else:
-        table = st.table(headers=["Repo URL", "Size"], rows=[r["repo"].url for r in repos], postfix="B") + "
-<a href={}: Download Results".format("https://drive.google.com/uc?id="+os.environ['STREAMLIT_USERNAME']+"&export=download")
-        st.markdown(table)
-        st.button("Download")
+# Define a function to get the 10 repositories with the largest file of the given name
+def get_repositories_with_largest_file(filename):
+    # Search for the file on GitHub using the Hugging Face API
+    repositories = Repository.search(query=filename, provider="github")
+    # Initialize a list to store the repository information
+    repo_info = []
+    # Loop through the repositories and get the largest file size
+    for repo in repositories[:10]:
+        # Get the repository information using the PyGitHub library
+        g = Github()
+        repository = g.get_repo(repo.full_name)
+        # Get the largest file in the repository
+        largest_file = None
+        largest_file_size = 0
+        for file in repository.get_files():
+            if file.name == filename and file.size > largest_file_size:
+                largest_file = file
+                largest_file_size = file.size
+        # Add the repository information to the list
+        if largest_file:
+            repo_info.append({
+                "name": repo.name,
+                "description": repo.description,
+                "size": largest_file_size,
+                "url": repo.html_url
+            })
+    # Return the repository information
+    return repo_info
+
+# Define the Streamlit app
+def app():
+    # Set up the user interface
+    st.title("GitHub File Search")
+    st.write("Enter a file name to search for on GitHub:")
+    filename = st.text_input("File name")
+    if filename:
+        if "." not in filename:
+            st.error(f"{filename} is not a valid file name. Please enter a valid file name with the file extension.")
+        else:
+            # Get the repositories with the largest file of the given name
+            repo_info = get_repositories_with_largest_file(filename)
+            # Display the repository information
+            if repo_info:
+                st.write(f"Top 10 repositories with the largest {filename}:")
+                for index, repo in enumerate(repo_info):
+                    st.write(f"{index+1}. [{repo['name']}]({repo['url']}) - {repo['size']} bytes")
+                    st.write(f"    {repo['description']}")
+            else:
+                st.warning(f"No repositories found with the file name {filename}.")
+            # Add a button to download the results as a text file
+            st.download_button(
+                label="Download results as text file",
+                data='\n\n'.join([f"{index+1}. {repo['name']} - {repo['size']} bytes\n{repo['url']}\n{repo['description']}" for index, repo in enumerate(repo_info)]),
+                file_name=f"GitHub File Search Results - {filename}.txt",
+                mime="text/plain"
+            )
+
+# Run the Streamlit app
+if __name__ == "__main__":
+    app()