Update app.py

app.py

@@ -1,37 +1,93 @@
 import os
+import logging
 import requests
 import gradio as gr
 from groq import Groq
+from urllib.parse import urlparse
 
-def generate_docs(repo_id: str) -> str:
+# Configure logging
+logging.basicConfig(level=logging.INFO)
+
+
+def extract_repo_id(input_str: str) -> str:
     """
-    Fetches repository details from GitHub and generates documentation using the Groq API.
+    Extract the repository ID in the format 'user/repo' from a URL or plain string.
     """
-    # Fetch basic repo info from GitHub
-    if repo_id.endswith(".git"):
-        repo_id = repo_id[:-4]
-    prefix = "github.com/"
-    start = repo_id.find(prefix)
-    if start != -1:
-        # Offset start by length of prefix
-        repo_path = repo_id[start + len(prefix):]
-        print(repo_path)
+    input_str = input_str.strip()
+
+    # Remove trailing '.git' if present
+    if input_str.endswith(".git"):
+        input_str = input_str[:-4]
+
+    # If input is a URL, parse it to extract the repo path.
+    if input_str.startswith("http"):
+        try:
+            parsed = urlparse(input_str)
+            path_parts = parsed.path.strip("/").split("/")
+            if len(path_parts) >= 2:
+                repo_id = "/".join(path_parts[:2])
+                logging.info("Extracted repo_id: %s", repo_id)
+                return repo_id
+            else:
+                raise ValueError("Invalid GitHub URL format.")
+        except Exception as e:
+            raise ValueError(f"Error parsing URL: {e}")
     else:
-        print("Invalid URL format")
-    response = requests.get(repo_path)
+        # Assume it's already in the correct format.
+        return input_str
+
+
+def fetch_repo_details(repo_id: str) -> dict:
+    """
+    Fetch repository details from GitHub using its API.
+    """
+    api_url = f"https://api.github.com/repos/{repo_id}"
+    response = requests.get(api_url)
     if response.status_code != 200:
-        return f"Error: Unable to fetch repository details for '{repo_id}'. Please check the repository ID."
-    url = f"https://api.github.com/repos/{repo_path}"
-    repo_data = response.json()
-    repo_description = repo_data.get("description", "No description provided.")
-    readme = repo_data.get("README.md", "No description provided.")
-  
+        raise ValueError(
+            f"Error: Unable to fetch repository details for '{repo_id}'. "
+            f"Status code: {response.status_code}"
+        )
+    return response.json()
+
+
+def fetch_readme(repo_id: str) -> str:
+    """
+    Attempt to fetch the README.md from the repository's main branch.
+    """
+    readme_url = f"https://raw.githubusercontent.com/{repo_id}/main/README.md"
+    response = requests.get(readme_url)
+    if response.status_code == 200:
+        return response.text
+    else:
+        return "No README found."
+
+
+def generate_docs(repo_input: str) -> str:
+    """
+    Fetch repository details from GitHub and generate documentation using the Groq API.
+    """
+    try:
+        # Extract a clean repository ID.
+        repo_id = extract_repo_id(repo_input)
+    except ValueError as e:
+        return str(e)
+
+    try:
+        repo_data = fetch_repo_details(repo_id)
+    except ValueError as e:
+        return str(e)
+
+    # Extract repository information.
     repo_name = repo_data.get("name", repo_id)
+    repo_description = repo_data.get("description", "No description provided.")
+    readme_content = fetch_readme(repo_id)
 
-    # Construct a prompt that includes some of the repository details.
+    # Construct the prompt for the Groq API.
     prompt = (
         f"Generate comprehensive documentation for the GitHub repository '{repo_id}'. "
-        f"The repository is named '{repo_name}' and its description is: {repo_description}, you should use {readme} for the base docs.\n\n"
+        f"The repository is named '{repo_name}' and its description is: {repo_description}. "
+        f"Use the following README content as reference:\n\n{readme_content}\n\n"
         "Please include sections such as an introduction, installation instructions, usage examples, "
         "and any relevant details that would help a new user understand and work with this repository."
     )
@@ -47,9 +103,7 @@ def generate_docs(repo_id: str) -> str:
     try:
         # Call the Groq API with the generated prompt.
         chat_completion = client.chat.completions.create(
-            messages=[
-                {"role": "user", "content": prompt}
-            ],
+            messages=[{"role": "user", "content": prompt}],
             model="llama-3.3-70b-versatile",
             stream=False,
         )
@@ -60,15 +114,19 @@ def generate_docs(repo_id: str) -> str:
     documentation = chat_completion.choices[0].message.content
     return documentation
 
+
 # Define the Gradio Blocks interface.
 with gr.Blocks() as demo:
     gr.Markdown("# GitHub Repository Documentation Generator")
     gr.Markdown(
-        "Enter a GitHub repository ID (in the format `user/repo`) below. "
+        "Enter a GitHub repository URL or ID (in the format `user/repo`) below. "
         "This tool fetches repository details and uses the Groq API to generate documentation."
     )
-    
-    repo_id_input = gr.Textbox(label="GitHub Repository ID (e.g., user/repo)", placeholder="octocat/Hello-World")
+
+    repo_id_input = gr.Textbox(
+        label="GitHub Repository URL or ID",
+        placeholder="https://github.com/octocat/Hello-World or octocat/Hello-World",
+    )
     output_box = gr.Textbox(label="Generated Documentation", lines=20)
     generate_button = gr.Button("Generate Documentation")
 
@@ -76,4 +134,4 @@ with gr.Blocks() as demo:
     generate_button.click(fn=generate_docs, inputs=repo_id_input, outputs=output_box)
 
 # Launch the Gradio app.
-demo.launch()
+demo.launch(ssr=False, debug=True)