feat: leaderboard

app.py

@@ -9,6 +9,7 @@ import time
 
 import os
 import sys
+from collections import defaultdict
 import pandas as pd
 import json
 import shutil
@@ -262,6 +263,70 @@ def certification(hf_username, first_name, last_name):
     return message, pdf, certificate, output_row.update(visible=visible)
 
 
+def make_clickable_repo(name, repo_type):
+    if repo_type == "space":
+        link = "https://huggingface.co/" + "spaces/" + name
+    elif repo_type == "model":
+        link = "https://huggingface.co/"  + name
+    elif repo_type == "dataset":
+        link = "https://huggingface.co/" + "datasets/" + name
+    return f'<a target="_blank" href="{link}">{name.split("/")[-1]}</a>'
+
+
+def make_clickable_user(user_id):
+    link = "https://huggingface.co/" + user_id
+    return f'<a  target="_blank" href="{link}">{user_id}</a>'
+
+
+def leaderboard():
+    """
+    Get the leaderboard of the hackathon.
+    
+    The leaderboard is a Pandas DataFrame with the following columns:
+    - Rank: the rank of the user in the leaderboard
+    - User: the Hugging Face username of the user
+    - Contributions: the list of contributions of the user (models, datasets, spaces)
+    - Likes: the total number of likes of the user's contributions
+    """
+    repo_list = {
+        'model': api.list_models,
+        'dataset': api.list_datasets,
+        'space': api.list_spaces
+    }
+
+    # Repos that should not be included in the leaderboard
+    not_included = [
+        # 'README',
+        # '2023-Hackathon-Certification',
+        # 'huggingface-krew-hackathon2023'
+    ]
+
+    contributions = defaultdict(list)
+
+    for repo_type in repo_list:
+        for repo in repo_list[repo_type](author=ORGANIZATION):
+            if repo.id.split('/')[-1] in not_included:
+                continue
+            commits = api.list_repo_commits(repo.id, repo_type=repo_type)
+            for author in set(author for commit in commits for author in commit.authors):
+                contributions[author].append((repo_type, repo.id, repo.likes))
+    
+    leaderboard = []
+    for user, repo_likes in contributions.items():
+        repos = []
+        user_likes = 0
+        for repo_type, repo, likes in repo_likes:
+            repos.append(make_clickable_repo(repo, repo_type))
+            user_likes += likes
+        leaderboard.append([make_clickable_user(user), '- ' + '\n- '.join(repos), user_likes])
+    
+    df = pd.DataFrame(data=leaderboard, columns=["User", "Contributions", "Likes"])
+    df.sort_values(by=["Likes"], ascending=False, inplace=True)
+    df.insert(0, "Rank", list(range(1, len(df) + 1)))
+    
+    return df
+
+
 with gr.Blocks() as demo:
     gr.Markdown(
         '<img style="display: block; margin-left: auto; margin-right: auto; height: 10em;"'
@@ -313,5 +378,16 @@ with gr.Blocks() as demo:
                 Don't hesitate to share your contributions on Twitter (tag me [@wonhseo](https://twitter.com/wonhseo) and [@huggingface](https://twitter.com/huggingface)) and on LinkedIn.
                 """
             )
+            with gr.Row():
+                repos_data = gr.components.Dataframe(
+                    type="pandas", datatype=["number", "markdown", "markdown", "number"]
+                )
+            with gr.Row():
+                data_run = gr.Button("Refresh")
+                data_run.click(
+                    leaderboard, outputs=repos_data
+                )
+    
+    demo.load(leaderboard, outputs=repos_data)
 
 demo.launch(debug=True)