Fix provider HLS downloads

provider_downloader.py

@@ -8,6 +8,7 @@ import os
 import re
 import subprocess
 import sys
+from urllib.parse import urljoin
 import urllib.request
 from pathlib import Path
 
@@ -37,6 +38,19 @@ def parse_show_id(value, default_provider="anikoto"):
     return default_provider, text
 
 
+def episode_key(value):
+    text = str(value or "").strip()
+    if not text:
+        return ""
+    try:
+        number = float(text)
+    except ValueError:
+        return text
+    if number.is_integer():
+        return str(int(number))
+    return str(number).rstrip("0").rstrip(".")
+
+
 def bridge(command, provider, *args):
     cmd = ["node", str(BRIDGE), command, provider, *[str(arg) for arg in args]]
     proc = subprocess.run(cmd, cwd=PROJECT_ROOT, text=True, capture_output=True, check=False)
@@ -96,7 +110,7 @@ def find_provider_show(provider, title, original_provider, original_show_id):
 
 def episode_values(spec, episodes):
     values = []
-    by_number = {str(ep.get("number")).rstrip("0").rstrip("."): ep for ep in episodes}
+    by_number = {episode_key(ep.get("number")): ep for ep in episodes}
     if not str(spec or "").strip():
         return episodes
     for part in re.split(r"[\s,]+", str(spec).strip()):
@@ -115,7 +129,7 @@ def episode_values(spec, episodes):
                 if item and item not in values:
                     values.append(item)
             continue
-        item = by_number.get(str(float(part)).rstrip("0").rstrip(".")) if re.match(r"^\d+(?:\.\d+)?$", part) else None
+        item = by_number.get(episode_key(part)) if re.match(r"^\d+(?:\.\d+)?$", part) else None
         if item and item not in values:
             values.append(item)
     return values
@@ -145,18 +159,18 @@ def requested_episode_numbers(spec, episodes):
 
 
 def wanted_episodes(spec, episodes):
-    by_number = {str(ep.get("number")).rstrip("0").rstrip("."): ep for ep in episodes or []}
+    by_number = {episode_key(ep.get("number")): ep for ep in episodes or []}
     wanted = []
     for number in requested_episode_numbers(spec, episodes):
-        key = str(number).rstrip("0").rstrip(".")
+        key = episode_key(number)
         wanted.append(by_number.get(key) or {"number": number, "id": ""})
     return wanted
 
 
 def episode_by_number(number, episodes):
-    wanted = str(number).rstrip("0").rstrip(".")
+    wanted = episode_key(number)
     for episode in episodes or []:
-        current = str(episode.get("number")).rstrip("0").rstrip(".")
+        current = episode_key(episode.get("number"))
         if current == wanted:
             return episode
     return None
@@ -170,6 +184,62 @@ def ffmpeg_headers(headers):
     return "\r\n".join(pairs) + ("\r\n" if pairs else "")
 
 
+def best_hls_variant_url(master_url, playlist_text):
+    variants = []
+    pending_bandwidth = 0
+    for raw_line in str(playlist_text or "").splitlines():
+        line = raw_line.strip()
+        if not line:
+            continue
+        if line.startswith("#EXT-X-STREAM-INF"):
+            match = re.search(r"BANDWIDTH=(\d+)", line)
+            pending_bandwidth = int(match.group(1)) if match else 0
+            continue
+        if pending_bandwidth and not line.startswith("#"):
+            variants.append((pending_bandwidth, urljoin(master_url, line)))
+            pending_bandwidth = 0
+    if not variants:
+        return master_url
+    variants.sort(key=lambda item: item[0], reverse=True)
+    return variants[0][1]
+
+
+def resolve_hls_input_url(stream):
+    url = stream.get("url")
+    if not stream.get("isM3U8") and ".m3u8" not in str(url):
+        return url
+    headers = {"User-Agent": "Mozilla/5.0", **(stream.get("headers") or {})}
+    request = urllib.request.Request(url, headers=headers)
+    with urllib.request.urlopen(request, timeout=30) as response:
+        content_type = response.headers.get("Content-Type", "")
+        data = response.read(2_000_000).decode("utf-8", errors="replace")
+    if "#EXT-X-STREAM-INF" not in data and "mpegurl" not in content_type.lower():
+        return url
+    return best_hls_variant_url(url, data)
+
+
+def ffmpeg_base_command(stream, input_url):
+    cmd = [
+        "ffmpeg",
+        "-hide_banner",
+        "-loglevel",
+        "info",
+        "-y",
+        "-protocol_whitelist",
+        "file,http,https,tcp,tls,crypto",
+        "-allowed_extensions",
+        "ALL",
+    ]
+    headers = ffmpeg_headers(stream.get("headers") or {})
+    if headers:
+        cmd.extend(["-headers", headers])
+    referer = (stream.get("headers") or {}).get("Referer")
+    if referer:
+        cmd.extend(["-referer", referer])
+    cmd.extend(["-i", input_url])
+    return cmd
+
+
 def download_subtitles(subtitles, output_base):
     saved = []
     english = [
@@ -188,15 +258,60 @@ def download_subtitles(subtitles, output_base):
 
 
 def download_episode(stream, target):
-    cmd = ["ffmpeg", "-hide_banner", "-loglevel", "info", "-y"]
-    headers = ffmpeg_headers(stream.get("headers") or {})
-    if headers:
-        cmd.extend(["-headers", headers])
-    referer = (stream.get("headers") or {}).get("Referer")
-    if referer:
-        cmd.extend(["-referer", referer])
-    cmd.extend(["-i", stream["url"], "-c", "copy", "-bsf:a", "aac_adtstoasc", str(target)])
-    return subprocess.call(cmd)
+    input_url = resolve_hls_input_url(stream)
+    partial = target.with_suffix(target.suffix + ".part")
+    for path in (target, partial):
+        try:
+            path.unlink()
+        except FileNotFoundError:
+            pass
+    attempts = [
+        [
+            *ffmpeg_base_command(stream, input_url),
+            "-map",
+            "0:v:0?",
+            "-map",
+            "0:a:0?",
+            "-c",
+            "copy",
+            "-bsf:a",
+            "aac_adtstoasc",
+            "-movflags",
+            "+faststart",
+            "-f",
+            "mp4",
+            str(partial),
+        ],
+        [
+            *ffmpeg_base_command(stream, input_url),
+            "-map",
+            "0",
+            "-c",
+            "copy",
+            "-bsf:a",
+            "aac_adtstoasc",
+            "-movflags",
+            "+faststart",
+            "-f",
+            "mp4",
+            str(partial),
+        ],
+    ]
+    code = 1
+    for cmd in attempts:
+        try:
+            partial.unlink()
+        except FileNotFoundError:
+            pass
+        code = subprocess.call(cmd)
+        if code == 0 and partial.exists() and partial.stat().st_size > 0:
+            partial.replace(target)
+            return 0
+    try:
+        partial.unlink()
+    except FileNotFoundError:
+        pass
+    return code
 
 
 def provider_episode_candidates(primary_provider, primary_show_id, title, wanted_episode):