Update app.py

app.py

@@ -5,7 +5,57 @@ from datetime import datetime, timedelta
 import time
 import os
 
-def fetch_github_users(github_token=None, max_users=200, min_followers=10):
+def get_contribution_count(username, github_token, days=30):
+    """
+    Get accurate contribution count using GitHub GraphQL API
+    """
+    if not github_token:
+        return 0
+    
+    # Calculate date range
+    end_date = datetime.now()
+    start_date = end_date - timedelta(days=days)
+    
+    query = """
+    query($username: String!, $from: DateTime!, $to: DateTime!) {
+      user(login: $username) {
+        contributionsCollection(from: $from, to: $to) {
+          contributionCalendar {
+            totalContributions
+          }
+        }
+      }
+    }
+    """
+    
+    variables = {
+        "username": username,
+        "from": start_date.isoformat(),
+        "to": end_date.isoformat()
+    }
+    
+    headers = {
+        'Authorization': f'bearer {github_token}',
+        'Content-Type': 'application/json',
+    }
+    
+    try:
+        response = requests.post(
+            'https://api.github.com/graphql',
+            json={'query': query, 'variables': variables},
+            headers=headers
+        )
+        
+        if response.status_code == 200:
+            data = response.json()
+            if 'data' in data and data['data']['user']:
+                return data['data']['user']['contributionsCollection']['contributionCalendar']['totalContributions']
+    except Exception as e:
+        print(f"Error fetching contributions for {username}: {e}")
+    
+    return 0
+
+def fetch_github_users(github_token=None, max_users=200, min_followers=10, days=30):
     """
     Fetch GitHub users from UAE directly using GitHub API
     Sorted by contributions (primary) and followers (secondary)
@@ -13,14 +63,15 @@ def fetch_github_users(github_token=None, max_users=200, min_followers=10):
     # Read token from environment variable if not provided
     if not github_token:
         github_token = os.getenv('GITHUB_TOKEN')
+    
+    if not github_token:
+        return pd.DataFrame(), "❌ GitHub token is required for accurate contribution counts. Please provide a token."
 
     headers = {
         'Accept': 'application/vnd.github.v3+json',
+        'Authorization': f'token {github_token}'
     }
 
-    if github_token:
-        headers['Authorization'] = f'token {github_token}'
-
     all_users = []
 
     # Search locations in UAE
@@ -54,30 +105,8 @@ def fetch_github_users(github_token=None, max_users=200, min_followers=10):
                         if user_response.status_code == 200:
                             user_data = user_response.json()
 
-                            # Fetch user's contribution data for last 30 days
-                            events_url = f"https://api.github.com/users/{user['login']}/events/public?per_page=100"
-                            events_response = requests.get(events_url, headers=headers)
-
-                            contributions = 0
-                            thirty_days_ago = datetime.now() - timedelta(days=30)
-                            
-                            if events_response.status_code == 200:
-                                events = events_response.json()
-                                # Count various contribution events from last 30 days
-                                for event in events:
-                                    event_date = datetime.strptime(event.get('created_at', ''), '%Y-%m-%dT%H:%M:%SZ')
-                                    if event_date >= thirty_days_ago:
-                                        event_type = event.get('type', '')
-                                        # Count different types of contributions
-                                        if event_type == 'PushEvent':
-                                            # Count commits in push event
-                                            commits = event.get('payload', {}).get('commits', [])
-                                            contributions += len(commits) if commits else 1
-                                        elif event_type in ['PullRequestEvent', 'IssuesEvent', 'CreateEvent']:
-                                            contributions += 1
-
-                            # Total contributions is based on last 30 days activity
-                            total_contributions = contributions
+                            # Get accurate contribution count using GraphQL
+                            contributions = get_contribution_count(user['login'], github_token, days)
 
                             all_users.append({
                                 'login': user_data.get('login', ''),
@@ -85,7 +114,7 @@ def fetch_github_users(github_token=None, max_users=200, min_followers=10):
                                 'avatar': user_data.get('avatar_url', ''),
                                 'followers': user_data.get('followers', 0),
                                 'public_repos': user_data.get('public_repos', 0),
-                                'contributions': total_contributions,
+                                'contributions': contributions,
                                 'location': user_data.get('location', ''),
                                 'bio': user_data.get('bio', ''),
                                 'company': user_data.get('company', ''),
@@ -100,7 +129,7 @@ def fetch_github_users(github_token=None, max_users=200, min_followers=10):
                 time.sleep(1)  # Rate limiting between searches
 
             elif response.status_code == 403:
-                status_updates.append(f"⚠️ Rate limit reached. Please add a GitHub token.")
+                status_updates.append(f"⚠️ Rate limit reached.")
                 break
             else:
                 status_updates.append(f"❌ Error searching {location}: {response.status_code}")
@@ -126,7 +155,7 @@ def fetch_github_users(github_token=None, max_users=200, min_followers=10):
             display_df.columns = ['Rank', 'Name', 'Username', 'Contributions', 'Followers', 'Public Repos', 'Location']
             display_df['GitHub Profile'] = df['login'].apply(lambda x: f"https://github.com/{x}")
 
-            status_message = f"✅ Successfully fetched top {len(df)} contributors (sorted by last 30 days contributions)\n" + "\n".join(status_updates[-5:])
+            status_message = f"✅ Successfully fetched top {len(df)} contributors (last {days} days)\n" + "\n".join(status_updates[-5:])
             return display_df, status_message
         else:
             return pd.DataFrame(), "⚠️ No users found\n" + "\n".join(status_updates)
@@ -153,17 +182,17 @@ def search_users(df, search_term):
 with gr.Blocks(theme=gr.themes.Soft(), title="GitHub UAE Top Contributors") as app:
 
     gr.Markdown("""
-    # 🏆 Top 200 GitHub Contributors in UAE (Last 30 Days)
-    ### Ranked by Contributions in the Last 30 Days
+    # 🏆 Top 200 GitHub Contributors in UAE
+    ### Ranked by Accurate Contribution Counts
 
-    **Note:** The app will automatically use `GITHUB_TOKEN` environment variable if set.
-    Without a token, you're limited to 60 requests/hour. With a token: 5000 requests/hour.
+    **IMPORTANT:** This app requires a GitHub Personal Access Token for accurate contribution counts.
+    The token needs the `read:user` scope for GraphQL API access.
     """)
 
     with gr.Row():
         token_input = gr.Textbox(
-            label="GitHub Personal Access Token (Optional - overrides env variable)",
-            placeholder="ghp_xxxxxxxxxxxx or leave empty to use GITHUB_TOKEN env var",
+            label="GitHub Personal Access Token (Required)",
+            placeholder="ghp_xxxxxxxxxxxx or set GITHUB_TOKEN env var",
             type="password",
             scale=3
         )
@@ -184,6 +213,13 @@ with gr.Blocks(theme=gr.themes.Soft(), title="GitHub UAE Top Contributors") as a
             value=10,
             step=1
         )
+        days_input = gr.Slider(
+            label="Days to Count (30 days = last 30 days, not calendar month)",
+            minimum=7,
+            maximum=90,
+            value=30,
+            step=1
+        )
 
     with gr.Row():
         fetch_btn = gr.Button("🚀 Fetch Top Contributors from GitHub", variant="primary", size="lg")
@@ -212,31 +248,31 @@ with gr.Blocks(theme=gr.themes.Soft(), title="GitHub UAE Top Contributors") as a
     gr.Markdown("""
     ---
     **Ranking Criteria:**
-    - **Primary Sort:** Contributions in Last 30 Days (Commits, Pull Requests, Issues, Creates)
+    - **Primary Sort:** Total contributions in selected time period (accurate count from GitHub)
     - **Secondary Sort:** Followers count
-    - Only users from UAE locations are included
-    - Contributions counted: Commits (from PushEvents), Pull Requests, Issues, and Create Events
-
-    **How to set up GitHub Token:**
-
-    **Option 1: Environment Variable (Recommended)**
-    ```bash
-    export GITHUB_TOKEN="ghp_your_token_here"
-    python app.py
-    ```
-
-    **Option 2: Manual Input**
-    1. Go to [GitHub Settings > Developer settings > Personal access tokens](https://github.com/settings/tokens)
-    2. Generate new token (classic)
-    3. No special scopes needed for public data
-    4. Copy and paste the token in the field above
-
-    **Data fetched directly from:** GitHub API
+    - Uses GitHub's GraphQL API for accurate contribution counts
+    - Counts ALL contribution types (commits, PRs, issues, reviews, etc.)
+
+    **How to set up GitHub Token with correct permissions:**
+
+    1. Go to [GitHub Settings > Developer settings > Personal access tokens > Tokens (classic)](https://github.com/settings/tokens)
+    2. Click "Generate new token (classic)"
+    3. Give it a name like "UAE Contributors Tracker"
+    4. Select scopes: **read:user** (this is required for GraphQL API)
+    5. Generate and copy the token
+    6. Either:
+       - Paste it in the field above, OR
+       - Set environment variable: `export GITHUB_TOKEN="ghp_your_token_here"`
+
+    **Why the change?**
+    - The Events API only returns ~100 events and may miss many contributions
+    - GraphQL API gives accurate counts of ALL contribution types
+    - This matches what you see on your GitHub profile
     """)
 
     # Event handlers
-    def fetch_and_display(token, max_users, min_followers):
-        df, msg = fetch_github_users(token if token else None, int(max_users), int(min_followers))
+    def fetch_and_display(token, max_users, min_followers, days):
+        df, msg = fetch_github_users(token if token else None, int(max_users), int(min_followers), int(days))
         return df, df, msg
 
     def filter_data(df, search):
@@ -249,7 +285,7 @@ with gr.Blocks(theme=gr.themes.Soft(), title="GitHub UAE Top Contributors") as a
 
     fetch_btn.click(
         fn=fetch_and_display,
-        inputs=[token_input, max_users_input, min_followers_input],
+        inputs=[token_input, max_users_input, min_followers_input, days_input],
         outputs=[full_data, data_display, status_msg]
     )