update app

app.py

@@ -1,15 +1,18 @@
+import huggingface_hub
 import gradio as gr
 import pandas as pd
 import random
 import openai
 import os
 import requests
-from datetime import datetime, timedelta, timezone
-from enum import Enum
 import sqlite3
 import string
 import hashlib
 import dotenv
+import shutil
+from apscheduler.schedulers.background import BackgroundScheduler
+from datetime import datetime, timedelta, timezone
+from enum import Enum
 
 dotenv.load_dotenv()
 openai_api_key = os.getenv("OPENAI_API_KEY")
@@ -18,19 +21,43 @@ discord_webhook_url_easy = os.getenv("DISCORD_WEBHOOK_URL_EASY")
 
 secret_key = os.getenv("CTF_SECRET_KEY", "ctf_secret_key")
 hard_challenge_secret = os.getenv("HARD_CHALLENGE_SECRET", "hard_challenge_secret")
-
-DB_FILE = "./data.db"
-db = sqlite3.connect(DB_FILE)
-
+hf_ctf_sync_token = os.getenv("HF_CTF_SYNC_TOKEN")
 
 class Env(str, Enum):
     PLAYGROUND = "playground"
-    CHALLENGE_EASY = "challenge_easy"
-    CHALLENGE_HARD = "challenge_hard"
+    CHALLENGE_EASY = "ctf_easy"
+    CHALLENGE_HARD = "ctf_hard"
 
+DB_FILE = "./reviews.db"
+
+repo = huggingface_hub.Repository(
+    local_dir="hf_data",
+    repo_type="dataset",
+    clone_from="https://huggingface.co/datasets/mislavb/test-ctf",
+    use_auth_token=hf_ctf_sync_token,
+)
+repo.git_pull()
+shutil.copyfile("./hf_data/reviews.db", DB_FILE)
+
+def backup_db():
+    db = sqlite3.connect(DB_FILE)
+    cur = db.cursor()
+    shutil.copyfile(DB_FILE, "./hf_data/reviews.db")
+    print("here")
+    for level in [Env.PLAYGROUND, Env.CHALLENGE_EASY, Env.CHALLENGE_HARD]:
+        reviews = cur.execute(f"SELECT * FROM {level.value}").fetchall()
+        pd_data = pd.DataFrame(reviews, columns=["id", "timestamp", "name", "feedback", "summary"])
+        pd_data.to_csv(f"./hf_data/data/reviews_{level.value}-00000-of-00001.csv", index=False)
+        repo.push_to_hub(blocking=False, commit_message=f"Updating data at {datetime.now()}")
+    db.close()
+
+scheduler = BackgroundScheduler()
+scheduler.add_job(func=backup_db, trigger="interval", seconds=60)
+scheduler.start()
 
 # Create table if it doesn't already exist
 def create_tables():
+    db = sqlite3.connect(DB_FILE)
     for level in [Env.PLAYGROUND, Env.CHALLENGE_EASY, Env.CHALLENGE_HARD]:
         try:
             db.execute(f"SELECT * FROM {level.value}").fetchall()
@@ -48,8 +75,10 @@ def create_tables():
 
 # Add review to the db
 def add_review_db(level: str, name: str, feedback: str, summary: str):
+    print("add_review_db: ", level, name, feedback, summary)
     db = sqlite3.connect(DB_FILE)
-    db.execute(
+    cur = db.cursor()
+    cur.execute(
         f"INSERT INTO {level} (name, feedback, summary) VALUES (?, ?, ?)",
         (name, feedback, summary),
     )
@@ -129,11 +158,14 @@ def summarize_feedback(rows: list[list[str]], env: Env, username: str) -> str:
 
     prompt = f"Read all received user feedback and summarize three most positive comments:\n{text}"
 
-    client = openai.Client(api_key=openai_api_key)
-    completion = client.chat.completions.create(
-        model="gpt-4o-mini", messages=[{"role": "user", "content": prompt}]
-    )
-    summary = completion.choices[0].message.content
+    if env == Env.CHALLENGE_HARD:
+        summary = "-"
+    else:
+        client = openai.Client(api_key=openai_api_key)
+        completion = client.chat.completions.create(
+            model="gpt-4o-mini", messages=[{"role": "user", "content": prompt}]
+        )
+        summary = completion.choices[0].message.content
 
     add_review_db(env.value, user, feedback, summary)