Added requirements.txt

.gitignore

@@ -12,3 +12,6 @@ build/
 .idea/
 *.swp
 *~
+tmp
+Music
+cookies.txt

happy_ripper.py

@@ -1,65 +1,340 @@
-from flask import Flask, render_template, request, jsonify
-from threading import Thread
-import time
-from yt_dlp import YoutubeDL
-from yt_dlp.utils import DownloadError
+from dataclasses import dataclass
 import pathlib
 import os
+import re
 import shutil
+import urllib.request
+from threading import Thread
+import time
+from collections.abc import Callable
+
+from yt_dlp import YoutubeDL
+from yt_dlp.postprocessor import PostProcessor
+from yt_dlp.utils import DownloadError, ExtractorError
+import mutagen
+from mutagen.easyid3 import EasyID3
+from mutagen.mp3 import MP3
+from mutagen.id3 import ID3, APIC
+from flask import Flask, render_template, request, jsonify
+
+MUSIC_FOLDER = "Music"
 
 
-def rip_worker(rip_queue: list[str]):
-    project_root_path = pathlib.Path(__file__).parent.resolve()
-    tmp_path = f"{project_root_path}/tmp"
+@dataclass
+class ReleaseInfo:
+    artist: str
+    title: str
+    art_url: str
+    is_album: bool
+    is_topic: bool
 
-    while True:
-        queue_is_empty = len(rip_queue) == 0
-        if queue_is_empty:
-            continue
 
-        # Creates tmp directory, if it already exists, it clears it, and makes a new one.
-        if not os.path.isdir(tmp_path):
-            os.mkdir(tmp_path)
+@dataclass
+class ReleaseStatus:
+    info: ReleaseInfo | None
+    status: str
+
+
+class RipError(Exception):
+    def __init__(self, msg):
+        super().__init__(msg)
+        self.msg = msg
+
+
+class MetadataPostProcessor(PostProcessor):
+    def __init__(self, info: ReleaseInfo, art_path: str, on_finish: Callable):
+        super().__init__()
+        self.info = info
+        self.art_path = art_path
+        self.on_finish = on_finish
+
+    def run(self, information):
+        file = pathlib.Path(information["filepath"])
+
+        audio = EasyID3(file)
+        audio.delete()
+        audio["title"] = information["title"]
+        audio["artist"] = information["uploader"]
+
+        audio["album"] = self.info.title
+
+        index = information.get("playlist_index")
+        if index is None:
+            index = 1
+        audio["tracknumber"] = str(index)
+
+        audio.save()
+
+        id3audio = MP3(file, ID3=ID3)
+        id3audio_tags = id3audio.tags
+
+        if id3audio_tags is None:
+            raise RipError("ID3 tags missing")
+
+        id3audio_tags.add(
+            APIC(
+                mime="image/jpeg",
+                type=3,
+                desc="Cover",
+                data=open(self.art_path, "rb").read(),
+            )
+        )
+        id3audio.save()
+
+        self.on_finish()
+        return [], information
+
+
+def create_or_wipe_folder(path: str) -> None:
+    """Creates tmp directory, if it already exists, it clears it, and makes a new one."""
+
+    if not os.path.isdir(path):
+        os.mkdir(path)
     else:
-            shutil.rmtree(tmp_path)
-            os.mkdir(tmp_path)
+        shutil.rmtree(path)
+        os.mkdir(path)
 
-        current_url = rip_queue[0]
 
+def get_content_json(url: str):
+    with YoutubeDL({"playlist_items": "1"}) as ydl:
+        return ydl.extract_info(url, download=False)
+
+
+def download_content(path: str, url: str, post_processor: PostProcessor) -> None:
     try:
         with YoutubeDL(
             {
-                    "format": "m4a/bestaudio/best",
+                "format": "bestaudio/best",
+                "ignoreerrors": True,
+                "restrictfilenames": True,
                 "postprocessors": [
                     {
                         "key": "FFmpegExtractAudio",
-                            "preferredcodec": "m4a",
+                        "preferredcodec": "mp3",
                     }
                 ],
-                    "outtmpl": f"{tmp_path}/%(title)s.%(ext)s",
+                "outtmpl": f"{path}/%(title)s.%(ext)s",
             }
         ) as ydl:
-                error_code = ydl.download([current_url])
+            ydl.add_post_processor(post_processor)
+            error_code = ydl.download([url])
+            if error_code > 1:
+                raise RipError(f"Content download failed ({error_code})")
 
     except DownloadError:
-            print("Download Error")
+        raise RipError("Content Download Failed")
 
-        rip_queue.pop(0)
+    except ExtractorError:
+        raise RipError("Content Extraction Failed")
+
+
+def find_release_info(url: str) -> ReleaseInfo:
+    try:
+        info = get_content_json(url)
+
+    except DownloadError:
+        raise RipError("Info Download Failed")
+
+    except ExtractorError:
+        raise RipError("Info Extraction Failed")
+
+    entries = info.get("entries")
+    if entries is None:
+        first_entry = info
+    else:
+        first_entry = next(iter(entries), None)
+        if not first_entry:
+            raise RipError("Entries is empty")
+
+    artist = first_entry.get("uploader")
+    if artist is None:
+        raise RipError("Uploader tag is none")
+
+    is_topic = False
+    if " - Topic" in artist:
+        is_topic = True
+        artist = artist.replace(" - Topic", "")
+
+    title = info.get("title")
+    if title is None:
+        raise RipError("Title tag is none")
+
+    is_album = False
+    if "Album - " in title:
+        is_album = True
+        title = title.replace("Album - ", "")
+
+    thumbnails = info.get("thumbnails")
+    if thumbnails is None:
+        raise RipError("Thumbnails is none")
+
+    art_url = thumbnails[1]["url"]
+
+    return ReleaseInfo(artist, title, art_url, is_album, is_topic)
+
+
+def find_tmp_path() -> str:
+    project_root_path = pathlib.Path(__file__).parent.resolve()
+    return f"{project_root_path}/tmp"
+
+
+def sort_into_music(path: str, info: ReleaseInfo) -> None:
+    try:
+        for file_name in os.listdir(path):
+            if file_name == "cover.jpg":
+                continue
+
+            audio = EasyID3(os.path.join(path, file_name))
+            artist_list = audio.get("artist")
+            if artist_list is None:
+                raise RipError("Artist tag was None")
+
+            artist = artist_list[0]
+            re.sub(r"[^\w_. -]", "_", artist)
+            title = info.title
+            re.sub(r"[^\w_. -]", "_", title)
+
+            artist_path = os.path.join(MUSIC_FOLDER, artist)
+            release_path = os.path.join(MUSIC_FOLDER, artist, title)
+
+            if not os.path.exists(artist_path):
+                os.makedirs(artist_path)
+
+            if not os.path.exists(release_path):
+                os.mkdir(release_path)
+
+            dst_file_path = os.path.join(release_path, file_name)
+            dst_file_path_cover = os.path.join(release_path, "cover.jpg")
+
+            if os.path.exists(dst_file_path):
+                os.remove(dst_file_path)
+
+            if os.path.exists(dst_file_path_cover):
+                os.remove(dst_file_path_cover)
+
+            source_file_path = os.path.join(path, file_name)
+            shutil.move(source_file_path, release_path)
+
+            source_file_path_cover = os.path.join(path, "cover.jpg")
+            shutil.copy(source_file_path_cover, release_path)
+
+    except Exception:
+        raise RipError("Filesystem error")
+
+
+def rip_worker(rip_queue: list[str], status_callback: Callable[[ReleaseInfo | str, bool],]):
+    tmp_path = find_tmp_path()
+
+    while True:
+        if len(rip_queue) == 0:
             time.sleep(1)
+            continue
+
+        status_callback("Gathering info", False)
+
+        url = rip_queue.pop(0)
+        create_or_wipe_folder(tmp_path)
+
+        try:
+            info = find_release_info(url)
+            status_callback(info, False)
+
+            status_callback("Downloading art", False)
+            art_path = os.path.join(tmp_path, "cover.jpg")
+            urllib.request.urlretrieve(info.art_url, art_path)
+            if not os.path.exists(art_path):
+                raise RipError("Art download failed")
+
+            status_callback("Downloading content (0)", False)
+
+            on_index = 0
+
+            def on_download_finish() -> None:
+                nonlocal on_index
+                on_index += 1
+                status_callback(f"Downloading content ({on_index})", False)
+
+            download_content(tmp_path, url, MetadataPostProcessor(info, art_path, on_download_finish))
+            if len(os.listdir(tmp_path)) == 0:
+                raise RipError("No content was downloaded")
+
+            status_callback("Sorting content", False)
+            sort_into_music(tmp_path, info)
+
+        except RipError as e:
+            status_callback(e.msg, True)
+
+        except Exception:
+            status_callback("Unknown Error", True)
+
+        else:
+            status_callback("Finished sucessfully", True)
 
 
+if __name__ == "__main__":
     app = Flask(__name__)
-rip_queue = []
-ripping_thread = Thread(target=rip_worker, args=(rip_queue,), daemon=True).start()
+    rip_queue: list[str] = []
+    finished_queue: list[ReleaseStatus] = []
 
+    current_release_info: ReleaseInfo | None = None
+    current_release_status_message: str = ""
+
+    def get_current_release_status(x: ReleaseInfo | str, finished: bool):
+        global current_release_info
+        global current_release_status_message
+        global finished_queue
+
+        if type(x) is ReleaseInfo:
+            current_release_info = x
+
+        elif type(x) is str:
+            current_release_status_message = x
+
+        if finished:
+            if current_release_info is None:
+                current_release_info = ReleaseInfo("Unknown", "Unknown", "url", False, False)
+
+            finished_queue.append(ReleaseStatus(current_release_info, current_release_status_message))
+            current_release_info = None
+            current_release_status_message = "Waiting"
+
+            if len(finished_queue) > 30:
+                finished_queue.pop(0)
+
+    ripping_thread = Thread(target=rip_worker, args=(rip_queue, get_current_release_status), daemon=True).start()
 
     @app.route("/")
     def index():
         return render_template("index.html")
 
-
     @app.route("/rip", methods=["POST"])
     def rip():
         url = request.form.get("url")
+        if url is None:
+            return jsonify({"success": False, "message": "Url is None"})
+
         rip_queue.append(url)
         return jsonify({"success": True, "message": f"Added, {url}!"})
+
+    @app.route("/status", methods=["GET"])
+    def status():
+        data = {"success": True, "queue": rip_queue, "finishedQueue": [], "status": current_release_status_message}
+
+        for q in finished_queue:
+            if q.info is None:
+                continue
+
+            data["finishedQueue"].append(
+                {"artist": q.info.artist, "title": q.info.title, "status": q.status, "art_url": q.info.art_url}
+            )
+
+        if current_release_info is not None:
+            data["currentReleaseInfo"] = {
+                "artist": current_release_info.artist,
+                "title": current_release_info.title,
+                "art_url": current_release_info.art_url,
+            }
+
+        return jsonify(data)
+
+    app.run()

requirements.txt

@@ -0,0 +1,3 @@
+Flask==3.1.3
+mutagen==1.47.0
+yt_dlp==2026.8.19

templates/index.html

@@ -30,7 +30,57 @@
       </form>
     </div>
 
+    </br>
+
+    <div class="container">
+      <h2>Current</h2>
+      <div id="status-message"></div>
+    </div>
+
+    </br>
+
+    <div class="container">
+      <h2>Queue</h2>
+      <table class="table" id="queue-table">
+        <thead>
+          <tr>
+            <th scope="col">#</th>
+            <th scope="col">URL</th>
+          </tr>
+        </thead>
+        <tbody id="queue-table-body"></tbody>
+      </table>
+    </div>
+
+    </br>
+
+    <div class="container">
+      <h2>Finished</h2>
+      <div id=finish-table-body></div>
+    </div>
+
     <script>
+      function createCard(artist, title, status, art) {
+        return `
+          <div class="card mb-3" style="max-width: 540px;">
+            <div class="row g-0">
+              <div class="col-md-4">
+                <img src="${art}" class="img-fluid rounded-start">
+              </div>
+              <div class="col-md-8">
+                <div class="card-body">
+                  <h5 class="card-title">${title}</h5>
+                  <p class="card-text">${artist}</p>
+                  <p class="card-text"><small class="text-body-secondary">${status}</small></p>
+                </div>
+              </div>
+            </div>
+          </div>
+        `;
+      }
+
+      const ripFormInput = document.getElementById("urlInput");
+
       document
         .getElementById("rip-form")
         .addEventListener("submit", async function (e) {
@@ -44,8 +94,43 @@
           });
 
           const data = await response.json();
+          ripFormInput.value = "";
         });
+
+      const queueTableBody = document.getElementById("queue-table-body");
+      const finishTableBody = document.getElementById("finish-table-body");
+      const statusMessage = document.getElementById("status-message");
+      const statusMessageImage = document.getElementById("status-message-img");
+
+      async function status() {
+        const response = await fetch("/status", { method: "GET" });
+        const statusJson = await response.json();
+
+        if ("currentReleaseInfo" in statusJson) {
+          const currentInfo = statusJson.currentReleaseInfo;
+          statusMessage.innerHTML = createCard(currentInfo.artist, currentInfo.title, statusJson.status, currentInfo.art_url);
+        }
+        else {
+          statusMessage.innerHTML = `<p>${statusJson.status}</p>`;
+        }
+
+        queueTableBody.innerHTML = "";
+
+        var i = 1;
+        for (const url of statusJson.queue) {
+          queueTableBody.innerHTML += `<tr><td>${i}</td><td>${url}</td></tr>`;
+          i ++;
+        }
+
+        finishTableBody.innerHTML = "";
+        for (const q of statusJson.finishedQueue) {
+          finishTableBody.innerHTML += createCard(q.artist, q.title, q.status, q.art_url);
+        }
+      }
+
+      setInterval(status, 100);
     </script>
+
     <script
       src="https://cdn.jsdelivr.net/npm/bootstrap@5.3.8/dist/js/bootstrap.bundle.min.js"
       integrity="sha384-FKyoEForCGlyvwx9Hj09JcYn3nv7wiPVlz7YYwJrWVcXK/BmnVDxM+D2scQbITxI"