Update app.py

app.py

@@ -12,6 +12,8 @@ import os
 from io import BytesIO
 from docx import Document
 import markdown
+from github import Github
+from github import GithubException
 
 # Hugging Face variables
 GIT_TOKEN = os.environ.get('GIT_TOKEN')
@@ -28,7 +30,12 @@ def fetch_git_files(git_url, git_provider):
         parts = git_url.split('/')
         owner = parts[3]
         repo = parts[4].split('.git')[0]
-        branch = 'main'  # You might want to make this configurable
+
+        # Initialize PyGitHub with the token
+        g = Github(GIT_TOKEN)
+
+        # Get the repository
+        repo = g.get_repo(f"{owner}/{repo}")
 
         # List of common dependency files to look for
         dependency_files = [
@@ -38,48 +45,21 @@ def fetch_git_files(git_url, git_provider):
 
         all_content = ""
 
-        # Set up headers with the personal access token
-        headers = {
-            "Authorization": f"token {GIT_TOKEN}",
-            "Accept": "application/vnd.github.v3+json"
-        }
-
-        base_url = {
-            'GitHub': 'https://api.github.com',
-            'GitLab': 'https://gitlab.com/api/v4',
-            'Gitea': 'https://gitea.com/api/v1'  # Adjust this URL for your Gitea instance
-        }.get(git_provider)
-
         for file_path in dependency_files:
-            # Construct the API URL based on the git provider
-            if git_provider == 'GitHub':
-                api_url = f"{base_url}/repos/{owner}/{repo}/contents/{file_path}?ref={branch}"
-            elif git_provider == 'GitLab':
-                api_url = f"{base_url}/projects/{owner}%2F{repo}/repository/files/{file_path}/raw?ref={branch}"
-            elif git_provider == 'Gitea':
-                api_url = f"{base_url}/repos/{owner}/{repo}/contents/{file_path}?ref={branch}"
-
-            # Make the API request
-            response = requests.get(api_url, headers=headers)
-
-            if response.status_code == 200:
-                if git_provider == 'GitHub' or git_provider == 'Gitea':
-                    content = response.json()
-                    if isinstance(content, dict) and 'content' in content:
-                        file_content = base64.b64decode(content['content']).decode('utf-8')
-                        all_content += f"\n\n--- {file_path} ---\n{file_content}"
-                elif git_provider == 'GitLab':
-                    file_content = response.text
-                    all_content += f"\n\n--- {file_path} ---\n{file_content}"
+            try:
+                file_content = repo.get_contents(file_path)
+                all_content += f"\n\n--- {file_path} ---\n{file_content.decoded_content.decode('utf-8')}"
+            except GithubException:
+                # File not found, skip to next file
+                continue
 
         if not all_content:
             return "Error: No dependency files found in the repository."
 
         return all_content
-    except requests.exceptions.RequestException as e:
+
+    except GithubException as e:
         return f"Error accessing {git_provider}: {str(e)}"
-    except json.JSONDecodeError:
-        return f"Error: Unable to parse {git_provider} API response for {file_path}"
 
 def process_chunk_with_gemini(chunk, gemini_api_key):
     genai.configure(api_key=gemini_api_key)