Malin/MeDBia
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: add libonig-dev for mbstring; update scanner for data discs

Browse Source 
Dockerfile: add libonig-dev (oniguruma) — required by mbstring extension

scanner/scan_disc.py:
- Treat all discs as data discs (mp4/mkv/etc), no VIDEO_TS/BDMV logic
- List video files by extension (.mp4 .mkv .avi .mov .m4v .ts .m2ts …)
- Infer media type from drutil type string, fall back to used_gb capacity
  (>8 GB → Blu-ray, >0.68 GB → DVD, smaller → CD)
- Store video file names in comment field, count in custom2

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Malin
2026-05-11 10:12:34 +02:00
parent 6002fc6e58
commit 451afc0440
2 changed files with 140 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files

261
scanner/scan_disc.py
View File

@@ -2,8 +2,11 @@
""" """
MeDBia Disc Scanner — macOS client MeDBia Disc Scanner — macOS client
=================================== ===================================
Polls the optical drive, reads disc metadata, posts to the remote Polls the optical drive, reads the disc content (video/data files),
videoDB API, then ejects the disc. posts to the remote videoDB API, then ejects the disc.
All discs are treated as data discs containing media files (mp4, mkv, etc.).
Media type is inferred from disc capacity reported by drutil.
Setup: Setup:
pip3 install requests pip3 install requests
@@ -24,174 +27,176 @@ try:
except ImportError: except ImportError:
sys.exit("Install requests first: pip3 install requests") sys.exit("Install requests first: pip3 install requests")
# ── Config (override with environment variables) ─────────────────────────────── # ── Config ─────────────────────────────────────────────────────────────────────
API_URL = os.environ.get("VIDEODB_URL", "http://your-server:6761").rstrip("/") + "/api_ingest.php" API_URL = os.environ.get("VIDEODB_URL", "http://your-server:6761").rstrip("/") + "/api_ingest.php"
API_TOKEN = os.environ.get("VIDEODB_TOKEN", "change_this_secret_token") API_TOKEN = os.environ.get("VIDEODB_TOKEN", "change_this_secret_token")
POLL_SEC = int(os.environ.get("POLL_INTERVAL", "5")) POLL_SEC = int(os.environ.get("POLL_INTERVAL", "5"))
# videoDB mediatype IDs (must match install.sql) # videoDB mediatype IDs (from install.sql)
MEDIATYPE = { MT_DVD = 1
"dvd": 1, MT_BLURAY = 16
"bluray": 16, MT_CD = 18
"cd": 18,
"data_cd": 18, # Video file extensions to list in the index
VIDEO_EXT = {
".mp4", ".mkv", ".avi", ".mov", ".m4v", ".ts", ".m2ts",
".wmv", ".flv", ".webm", ".vob", ".mpg", ".mpeg", ".iso",
} }
# Known system volumes to ignore when scanning /Volumes/ # System volumes to ignore when scanning /Volumes/
IGNORE_VOLUMES = {"Macintosh HD", "Preboot", "Recovery", "VM", "Data", "Update"} IGNORE_VOLUMES = {"Macintosh HD", "Preboot", "Recovery", "VM", "Data", "Update"}
# ── Shell helpers ────────────────────────────────────────────────────────────── # ── Shell helper ───────────────────────────────────────────────────────────────
def run(cmd: list) -> str: def run(cmd: list) -> str:
"""Run a command, return stdout (empty string on error)."""
try: try:
r = subprocess.run(cmd, capture_output=True, text=True, timeout=15) r = subprocess.run(cmd, capture_output=True, text=True, timeout=20)
return r.stdout return r.stdout
except Exception: except Exception:
return "" return ""
# ── Disc detection ───────────────────────────────────────────────────────────── # ── Disc presence ──────────────────────────────────────────────────────────────
def disc_status() -> dict | None: def disc_status() -> dict | None:
""" """Returns disc info dict, or None when no disc is present."""
Returns a dict with keys: drutil_type, device, tracks
Returns None when no disc is present.
"""
out = run(["drutil", "status"]) out = run(["drutil", "status"])
if not out or "No Media" in out: if not out or "No Media" in out:
return None return None
info: dict = {} info: dict = {}
# Type: DVD-ROM / Audio CD / CD-ROM / BD-ROM …
m = re.search(r"Type:\s+(.+?)(?:\s{3,}|$)", out, re.MULTILINE) m = re.search(r"Type:\s+(.+?)(?:\s{3,}|$)", out, re.MULTILINE)
if m: if m:
info["drutil_type"] = m.group(1).strip() info["drutil_type"] = m.group(1).strip()
# Device node: /dev/disk2
m = re.search(r"Name:\s+(/dev/\S+)", out) m = re.search(r"Name:\s+(/dev/\S+)", out)
if m: if m:
info["device"] = m.group(1) info["device"] = m.group(1)
# Track count # Space Used in GB (e.g. "Space Used: 7.88 GB")
m = re.search(r"Tracks:\s+(\d+)", out) m = re.search(r"Space Used:\s+([\d.]+)\s*GB", out)
if m: if m:
info["tracks"] = int(m.group(1)) info["used_gb"] = float(m.group(1))
return info return info
# ── Mount ──────────────────────────────────────────────────────────────────────
def find_mount(device: str) -> str | None: def find_mount(device: str) -> str | None:
"""Find where the optical disc is mounted.""" """Return the mount point for the optical disc."""
# Try 'mount' output first
for line in run(["mount"]).splitlines(): for line in run(["mount"]).splitlines():
if device in line: if device in line:
# "... on /Volumes/FOO (…)" m = re.search(r" on (/Volumes/[^\s(]+)", line)
m = re.search(r" on (/Volumes/\S+)", line)
if m: if m:
return m.group(1).rstrip("()") return m.group(1)
# Fallback: new entry in /Volumes/ that isn't a system volume # Fallback: first non-system entry in /Volumes/
volumes = set(os.listdir("/Volumes/")) - IGNORE_VOLUMES try:
if volumes: volumes = set(os.listdir("/Volumes/")) - IGNORE_VOLUMES
# Return the first one alphabetically if volumes:
return f"/Volumes/{sorted(volumes)[0]}" return f"/Volumes/{sorted(volumes)[0]}"
except Exception:
pass
return None return None
# ── Disc classification ─────────────────────────────────────────────────────── # ── Media type from disc capacity ──────────────────────────────────────────────
def classify(mount: str | None, drutil_type: str) -> tuple[int, str]: def mediatype_from_size(used_gb: float) -> tuple[int, str]:
""" """
Returns (mediatype_id, label) for videoDB. Infer videoDB mediatype from used disc capacity.
Checks filesystem structure first, falls back to drutil type string. Blu-ray discs hold 25/50 GB; DVDs hold ~4.7/8.5 GB; CDs ~0.7 GB.
""" """
if mount: if used_gb > 8.0:
p = Path(mount) return MT_BLURAY, "Blu-ray"
if (p / "BDMV").exists(): if used_gb > 0.68:
return MEDIATYPE["bluray"], "Blu-ray" return MT_DVD, "DVD"
if (p / "VIDEO_TS").exists() or (p / "VIDEO_TS.IFO").exists(): return MT_CD, "CD"
return MEDIATYPE["dvd"], "DVD"
def mediatype_from_drutil(drutil_type: str) -> tuple[int, str] | None:
"""Parse drutil type string if available."""
t = drutil_type.upper() t = drutil_type.upper()
if "AUDIO" in t:
return MEDIATYPE["cd"], "Audio CD"
if "BD" in t: if "BD" in t:
return MEDIATYPE["bluray"], "Blu-ray" return MT_BLURAY, "Blu-ray"
if "DVD" in t: if "DVD" in t:
return MEDIATYPE["dvd"], "DVD" return MT_DVD, "DVD"
if "CD" in t:
return MEDIATYPE["cd"], "CD/Data" return MT_CD, "CD"
return None
def volume_label(mount: str | None, device: str) -> str: # ── File listing ───────────────────────────────────────────────────────────────
"""Get the disc's volume label."""
if mount:
label = os.path.basename(mount)
if label:
return label
if device: def list_video_files(mount: str) -> list[str]:
for line in run(["diskutil", "info", device]).splitlines(): """Return relative paths of all video files on the disc."""
if "Volume Name" in line: found = []
return line.split(":", 1)[-1].strip() try:
for root, _dirs, files in os.walk(mount):
return "Unknown Disc" for f in files:
if Path(f).suffix.lower() in VIDEO_EXT:
rel = os.path.relpath(os.path.join(root, f), mount)
found.append(rel)
except PermissionError:
pass
return sorted(found)
def disc_size_bytes(mount: str | None) -> int: def all_files_count(mount: str) -> int:
"""Total used space on the disc in bytes.""" """Count every file on the disc (for discs with no video files)."""
if not mount: count = 0
return 0 try:
for _root, _dirs, files in os.walk(mount):
count += len(files)
except PermissionError:
pass
return count
def disc_size_bytes(mount: str) -> int:
"""Used space in bytes via df."""
out = run(["df", "-k", mount]) out = run(["df", "-k", mount])
for line in out.splitlines()[1:]: for line in out.splitlines()[1:]:
parts = line.split() parts = line.split()
if len(parts) >= 3: if len(parts) >= 3:
try: try:
return int(parts[2]) * 1024 # 'Used' column, KB→B return int(parts[2]) * 1024
except ValueError: except ValueError:
pass pass
return 0 return 0
def sample_files(mount: str | None, limit: int = 30) -> list[str]: # ── Volume label ───────────────────────────────────────────────────────────────
"""Return a sample of file paths on the disc."""
if not mount or not os.path.exists(mount): def volume_label(mount: str | None, device: str) -> str:
return [] if mount:
found = [] label = os.path.basename(mount)
try: if label:
for root, _dirs, files in os.walk(mount): return label
for f in files: if device:
rel = os.path.relpath(os.path.join(root, f), mount) for line in run(["diskutil", "info", device]).splitlines():
found.append(rel) if "Volume Name" in line:
if len(found) >= limit: return line.split(":", 1)[-1].strip()
return found return "Unknown Disc"
except PermissionError:
pass
return found
# ── Eject ───────────────────────────────────────────────────────────────────── # ── Eject ──────────────────────────────────────────────────────────────────────
def eject(mount: str | None): def eject(mount: str | None):
"""Eject the disc. Try diskutil first, fall back to drutil."""
if mount: if mount:
result = subprocess.run( r = subprocess.run(["diskutil", "eject", mount], capture_output=True)
["diskutil", "eject", mount], if r.returncode == 0:
capture_output=True, text=True
)
if result.returncode == 0:
return return
run(["drutil", "eject"]) run(["drutil", "eject"])
# ── API submission ──────────────────────────────────────────────────────────── # ── API submission ─────────────────────────────────────────────────────────────
def submit(payload: dict) -> bool: def submit(payload: dict) -> bool:
"""POST disc data to the videoDB API. Returns True on success."""
try: try:
resp = requests.post( resp = requests.post(
API_URL, API_URL,
@@ -201,20 +206,21 @@ def submit(payload: dict) -> bool:
) )
if resp.status_code == 200: if resp.status_code == 200:
data = resp.json() data = resp.json()
print(f" [OK] Indexed as entry #{data.get('id', '?')}: {payload['title']}") print(f" [OK] Entry #{data.get('id', '?')}: {payload['title']}")
return True return True
else: else:
print(f" [ERR] API {resp.status_code}: {resp.text[:200]}") print(f" [ERR] API {resp.status_code}: {resp.text[:300]}")
return False return False
except requests.ConnectionError: except requests.ConnectionError:
print(f" [ERR] Cannot reach {API_URL} — check VIDEODB_URL and network") print(f" [ERR] Cannot reach {API_URL}")
print(f" Check VIDEODB_URL and that the server is up.")
return False return False
except Exception as e: except Exception as e:
print(f" [ERR] {e}") print(f" [ERR] {e}")
return False return False
# ── Main scan ───────────────────────────────────────────────────────────────── # ── Main scan routine ──────────────────────────────────────────────────────────
def scan_and_submit(): def scan_and_submit():
status = disc_status() status = disc_status()
@@ -223,58 +229,70 @@ def scan_and_submit():
drutil_type = status.get("drutil_type", "") drutil_type = status.get("drutil_type", "")
device = status.get("device", "") device = status.get("device", "")
tracks = status.get("tracks", 0) used_gb = status.get("used_gb", 0.0)
print(f" drutil type : {drutil_type}") print(f" drutil type : {drutil_type}")
print(f" device : {device}") print(f" device : {device}")
print(f" tracks : {tracks}") print(f" used : {used_gb:.2f} GB")
# Give macOS a moment to finish mounting # Give macOS a moment to finish mounting the filesystem
time.sleep(3) time.sleep(4)
mount = find_mount(device) mount = find_mount(device)
print(f" mount point : {mount or '(not mounted)'}") print(f" mount : {mount or '(not mounted)'}")
mediatype_id, mediatype_name = classify(mount, drutil_type) # Determine media type — prefer drutil string, fall back to size
title = volume_label(mount, device) mt = mediatype_from_drutil(drutil_type)
size = disc_size_bytes(mount) if mt:
files = sample_files(mount) mediatype_id, mediatype_name = mt
# Build a short content summary for the 'comment' field
if tracks and mediatype_id == MEDIATYPE["cd"]:
summary = f"{tracks} audio tracks"
elif files:
summary = f"{len(files)} files"
if len(files) <= 10:
summary += ": " + ", ".join(Path(f).name for f in files[:10])
else: else:
summary = "" mediatype_id, mediatype_name = mediatype_from_size(used_gb)
title = volume_label(mount, device)
size_b = disc_size_bytes(mount) if mount else int(used_gb * 1024**3)
# Build file listing
video_files = list_video_files(mount) if mount else []
total_files = all_files_count(mount) if mount else 0
if video_files:
# List video file names (not full paths) for the comment field
names = [Path(f).name for f in video_files]
summary = f"{len(video_files)} video files: " + ", ".join(names[:10])
if len(names) > 10:
summary += f" … +{len(names) - 10} more"
# Store full file list in custom2 (255 char limit — truncate gracefully)
file_detail = "\n".join(video_files)
else:
summary = f"{total_files} files (no video files detected)"
file_detail = ""
payload = { payload = {
"title": title, "title": title,
"mediatype": mediatype_id, "mediatype": mediatype_id,
"comment": summary[:255], "comment": summary[:255],
"filesize": size, "filesize": size_b,
"disklabel": title[:32], "disklabel": title[:32],
"custom1": drutil_type[:255], # raw disc type string "custom1": drutil_type[:255],
"custom2": str(tracks) if tracks else str(len(files)), # track/file count "custom2": str(len(video_files) or total_files),
} }
print(f"\n Submitting [{mediatype_name}] \"{title}\" ({size // (1024*1024)} MB)") print(f"\n [{mediatype_name}] \"{title}\"")
print(f" {summary[:120]}")
ok = submit(payload) ok = submit(payload)
print(" Ejecting...") print(" Ejecting disc...")
eject(mount) eject(mount)
return ok return ok
# ── Entry point ─────────────────────────────────────────────────────────────── # ── Entry point ────────────────────────────────────────────────────────────────
def main(): def main():
print("=" * 55) print("=" * 55)
print(" MeDBia Disc Scanner") print(" MeDBia Disc Scanner")
print(f" API : {API_URL}") print(f" API : {API_URL}")
print(f" Poll: every {POLL_SEC}s") print(f" Poll : every {POLL_SEC}s")
print("=" * 55) print("=" * 55)
print("Insert a disc to index it. Ctrl-C to stop.\n") print("Insert a disc to index it. Ctrl-C to stop.\n")
@@ -287,11 +305,10 @@ def main():
was_present = True was_present = True
print("Disc detected!") print("Disc detected!")
scan_and_submit() scan_and_submit()
was_present = False # reset — disc was ejected was_present = False
print("\nReady — insert next disc.\n") print("\nReady — insert next disc.\n")
elif not status and was_present: elif not status and was_present:
# Disc manually removed before scan completed
was_present = False was_present = False
time.sleep(POLL_SEC) time.sleep(POLL_SEC)

1
videodb/Dockerfile
View File

@@ -5,6 +5,7 @@ RUN apt-get update && apt-get install -y \
libpng-dev \ libpng-dev \
libjpeg-dev \ libjpeg-dev \
libfreetype6-dev \ libfreetype6-dev \
libonig-dev \
default-mysql-client \ default-mysql-client \
&& rm -rf /var/lib/apt/lists/* && rm -rf /var/lib/apt/lists/*