Update watcher.py

watcher.py

@@ -6,7 +6,7 @@ import requests
 from datetime import datetime, timezone
 from bs4 import BeautifulSoup
 
-# ===== Channel.io 設定 =====
+# ===== Channel.io =====
 GET_URL = "https://desk-api.channel.io/desk/channels/200605/groups/519217/messages"
 POST_URL = GET_URL
 
@@ -34,13 +34,14 @@ HEADERS_POST = {
 }
 
 # ===== ssyoutube =====
-SSYOUTUBE_URL = "https://ssyoutube.online/yt-video-detail/"
+DETAIL_URL = "https://ssyoutube.online/yt-video-detail/"
+MERGE_API = "https://ssyoutube.online/wp-admin/admin-ajax.php"
 
 # ===== Utils =====
 def parse_updated_at(value):
     if isinstance(value, (int, float)):
         return datetime.fromtimestamp(value / 1000, tz=timezone.utc)
-    elif isinstance(value, str):
+    if isinstance(value, str):
         return datetime.fromisoformat(value.replace("Z", "+00:00"))
     return None
 
@@ -55,16 +56,31 @@ def extract_youtube_id(text):
             return m.group(1)
     return None
 
-# ===== ssyoutube HTML 解析 =====
+# ===== ssyoutube =====
+def fetch_nonce():
+    res = requests.get(DETAIL_URL, headers={"User-Agent": "Mozilla/5.0"}, timeout=30)
+    res.raise_for_status()
+    soup = BeautifulSoup(res.text, "lxml")
+
+    nonce = soup.select_one('input[name="nonce"]')
+    if nonce:
+        return nonce["value"]
+
+    m = re.search(r'"nonce":"([a-f0-9]+)"', res.text)
+    if m:
+        return m.group(1)
+
+    raise RuntimeError("nonce が取得できません")
+
 def fetch_download_links(youtube_url):
     res = requests.post(
-        SSYOUTUBE_URL,
+        DETAIL_URL,
         data={"videoURL": youtube_url},
-        timeout=30,
         headers={
             "User-Agent": "Mozilla/5.0",
-            "Referer": "https://ssyoutube.online/",
-        }
+            "Referer": DETAIL_URL,
+        },
+        timeout=30,
     )
     res.raise_for_status()
 
@@ -74,59 +90,91 @@ def fetch_download_links(youtube_url):
     results = []
     for btn in buttons:
         url = btn.get("data-url")
-        quality = btn.get("data-quality")  # 例: 1080p / 720p / None
-        has_audio = btn.get("data-has-audio")  # "true" / "false" / None
+        quality = btn.get("data-quality")
+        has_audio = btn.get("data-has-audio")
+        if url:
+            results.append({
+                "url": url,
+                "quality": quality or "audio",
+                "has_audio": has_audio,
+            })
+    return results
 
-        if not url:
-            continue
+def pick_best_streams(items):
+    videos = [i for i in items if i["has_audio"] == "false"]
+    audios = [i for i in items if i["has_audio"] != "false"]
 
-        results.append({
-            "url": url,
-            "quality": quality or "audio",
-            "has_audio": has_audio,
-        })
+    def vkey(x):
+        m = re.search(r"(\d+)p", x["quality"])
+        return int(m.group(1)) if m else 0
 
-    return results
+    best_video = sorted(videos, key=vkey, reverse=True)[0]
+    best_audio = audios[0]
 
-def build_links(items):
-    lines = []
-    for item in items:
-        url = item["url"]
-        quality = item["quality"]
-        has_audio = item["has_audio"]
+    return best_video, best_audio
+
+def start_merge(video, audio, youtube_id, quality, nonce):
+    payload = {
+        "id": f"{youtube_id}_{quality}",
+        "ttl": 3600000,
+        "inputs": [
+            {"url": audio["url"], "ext": "mp4"},
+            {"url": video["url"], "ext": "mp4"},
+        ],
+        "output": {
+            "ext": "mp4",
+            "downloadName": f"{youtube_id}_{quality}.mp4",
+        },
+        "operation": {"type": "replace_audio_in_video"},
+    }
+
+    files = {
+        "action": (None, "process_video_merge"),
+        "nonce": (None, nonce),
+        "request_data": (None, json.dumps(payload)),
+    }
+
+    res = requests.post(
+        MERGE_API,
+        files=files,
+        headers={"User-Agent": "Mozilla/5.0"},
+        timeout=30,
+    )
+    res.raise_for_status()
+    return res.json()
+
+def wait_merge_done(monitor_url):
+    while True:
+        res = requests.get(monitor_url, timeout=15)
+        res.raise_for_status()
+        data = res.json()
 
-        audio_label = ""
-        if has_audio == "false":
-            audio_label = "（映像のみ）"
-        elif has_audio == "true":
-            audio_label = "（音声付き）"
+        if data["result"]["status"] == "done":
+            return data["result"]["output"]["url"]
 
-        line = f'<link type="url" value="{url}"> {quality} {audio_label}</link>'
-        lines.append(line)
+        time.sleep(5)
 
+def build_links(items, merged_url, quality):
+    lines = []
+    for i in items:
+        lines.append(
+            f'<link type="url" value="{i["url"]}"> {i["quality"]}</link>'
+        )
+    lines.append(
+        f'\n<link type="url" value="{merged_url}"> 🔊音声付き {quality}</link>'
+    )
     return "\n".join(lines)
 
 def send_to_channel(text):
     payload = {
-        "requestId": f"desk-web-{int(time.time() * 1000)}",
-        "blocks": [
-            {
-                "type": "text",
-                "value": text
-            }
-        ],
-        "buttons": None,
-        "form": None,
-        "webPage": None,
-        "files": None,
-        "customPayload": None
+        "requestId": f"desk-{int(time.time()*1000)}",
+        "blocks": [{"type": "text", "value": text}],
     }
-
     res = requests.post(
         POST_URL,
         headers=HEADERS_POST,
         data=json.dumps(payload),
-        timeout=30
+        timeout=30,
     )
     res.raise_for_status()
 
@@ -134,39 +182,24 @@ def send_to_channel(text):
 def main():
     while True:
         try:
-            res = requests.get(
-                GET_URL,
-                headers=HEADERS_GET,
-                params=PARAMS,
-                timeout=30,
-            )
+            res = requests.get(GET_URL, headers=HEADERS_GET, params=PARAMS, timeout=30)
             res.raise_for_status()
-
             messages = res.json().get("messages", [])
-            latest_msg = None
-            latest_time = None
-
-            for msg in messages:
-                plain_text = msg.get("plainText")
-                updated_at = msg.get("updatedAt")
-
-                if not plain_text or updated_at is None:
-                    continue
-
-                t = parse_updated_at(updated_at)
-                if not t:
-                    continue
 
-                if latest_time is None or t > latest_time:
-                    latest_time = t
-                    latest_msg = msg
+            latest = max(
+                (
+                    m for m in messages
+                    if m.get("plainText") and m.get("updatedAt")
+                ),
+                key=lambda m: parse_updated_at(m["updatedAt"]),
+                default=None,
+            )
 
-            if not latest_msg:
+            if not latest:
                 time.sleep(10)
                 continue
 
-            text = latest_msg["plainText"]
-            youtube_id = extract_youtube_id(text)
+            youtube_id = extract_youtube_id(latest["plainText"])
             if not youtube_id:
                 time.sleep(10)
                 continue
@@ -174,13 +207,16 @@ def main():
             youtube_url = f"https://www.youtube.com/watch?v={youtube_id}"
 
             items = fetch_download_links(youtube_url)
-            if not items:
-                print("ダウンロードリンクが取得できませんでした")
-                time.sleep(10)
-                continue
+            best_video, best_audio = pick_best_streams(items)
+
+            nonce = fetch_nonce()
+            merge = start_merge(best_video, best_audio, youtube_id, best_video["quality"], nonce)
+
+            monitor = merge["data"]["result"]["monitor"]["http"]
+            merged_url = wait_merge_done(monitor)
 
-            message_text = build_links(items)
-            send_to_channel(message_text)
+            message = build_links(items, merged_url, best_video["quality"])
+            send_to_channel(message)
             print("送信完了")
 
         except Exception as e: