create daily/unlimited mode, CDN audio file for daily mode

1 files changed, 91 insertions, 0 deletions

diff --git a/playlist_generator/generate_daily.py b/playlist_generator/generate_daily.py
new file mode 100644
index 0000000..7bd4683
--- /dev/null
+++ b/playlist_generator/generate_daily.py
@@ -0,0 +1,91 @@
+import os
+import boto3
+import json
+import requests
+import random
+from dotenv import load_dotenv
+from datetime import datetime, timezone
+import yt_dlp
+
+load_dotenv()
+
+ACCOUNT_ID = os.getenv("R2_ACCOUNT_ID")
+ACCESS_KEY = os.getenv("R2_ACCESS_KEY")
+SECRET_KEY = os.getenv("R2_SECRET_KEY")
+BUCKET = os.getenv("R2_BUCKET")
+API_URL = os.getenv("API_URL")
+OBFUSCATION_KEY = os.getenv("OBFUSCATION_KEY")
+
+
+def xor_buffer(data: bytes, key: bytes) -> bytes:
+    return bytes(b ^ key[i % len(key)] for i, b in enumerate(data))
+
+def get_obfuscation_key() -> bytes:
+    date = datetime.now(timezone.utc).strftime("%Y-%m-%d")
+    return (OBFUSCATION_KEY + date).encode("utf-8")
+
+
+def decode_data(hex_data: str):
+    encrypted = bytes.fromhex(hex_data)
+    key = get_obfuscation_key()
+    decrypted = xor_buffer(encrypted, key)
+    return json.loads(decrypted.decode("utf-8"))
+
+
+def fetch_daily() -> dict:
+    url = f"{API_URL}/today"
+    response = requests.get(url)
+    response.raise_for_status()
+    return response.json()
+
+
+def download_random_segment_mp3(youtube_id: str, output_file="today.mp3") -> str:
+    url = f"https://www.youtube.com/watch?v={youtube_id}"
+    with yt_dlp.YoutubeDL({"quiet": True}) as ydl:
+        info = ydl.extract_info(url, download=False)
+        duration = info.get("duration", 60)
+    start = 0 if duration <= 17 else random.randint(0, duration - 17)
+    ydl_opts = {
+        "format": "bestaudio/best",
+        "outtmpl": "today.%(ext)s",
+        "quiet": True,
+        "download_ranges": lambda info, _: [
+            {"start_time": start, "end_time": start + 17}
+        ],
+        "force_keyframes_at_cuts": True,
+        "postprocessors": [
+            {
+                "key": "FFmpegExtractAudio",
+                "preferredcodec": "mp3",
+                "preferredquality": "192",
+            }
+        ],
+    }
+
+    with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+        ydl.download([url])
+    return output_file
+
+
+def upload_to_r2(file_path: str, object_key: str):
+    s3 = boto3.client(
+        "s3",
+        endpoint_url=f"https://{ACCOUNT_ID}.r2.cloudflarestorage.com",
+        aws_access_key_id=ACCESS_KEY,
+        aws_secret_access_key=SECRET_KEY,
+        region_name="auto",
+    )
+    s3.upload_file(file_path, BUCKET, object_key)
+
+
+def main():
+    daily_data = fetch_daily()
+    data = decode_data(daily_data["data"])
+    print(data)
+    youtube_id = data["youtubeId"]
+    clip_path = download_random_segment_mp3(youtube_id)
+    date = datetime.now(timezone.utc).strftime("%Y-%m-%d")
+    upload_to_r2(clip_path, f"kheardle/{date}.mp3")
+
+if __name__ == "__main__":
+    main()