Convert a string to a LanguageCode instance. Matches on ISO codes, English name, or native name.

Get environment variable with backwards compatibility fallback. Args: new_name: The new standardized environment variable name

Determines if subtitle generation should be skipped for a file. Args: file_path: Path to the media file. target_language: Th

Only schedules a cleanup timer if the system is actually idle. This prevents unnecessary timer resets when a large batch is being processed.

Handles the possibility of a media file having multiple audio tracks. Returns bytes directly instead of BytesIO to prevent memory leak.

Checks if a task_id is currently queued or processing.

Extracts information about the audio tracks in a file. Returns: List of dictionaries with information about each audio track.

Gets the full path to a file from the Plex server. Args: itemid: The ID of the item in the Plex library. server_ip: The IP address

Extract language codes from each subtitle stream in the video file using pyav. Forced subtitle tracks are excluded when ignore_forced_subtitl

Convert a language name (either English or native) to LanguageCode enum.

Generate a deterministic hash from audio content and optional parameters. Same audio + same task + same language = always same hash.

Checks if the given folder has a subtitle file with the given language. Args: video_file (str): The path of the video file. target

Refreshes the metadata of a Jellyfin library item. Args: itemid: The ID of the item in the Jellyfin library whose metadata needs

Format seconds as [H:]MM:SS without milliseconds.

Get the next episode's ratingKey based on the current episode in Plex. Args: current_episode_rating_key (str): The ratingKey of the c

Wait for file stability before processing.

Checks whether a video container has an embedded subtitle track in the given language. Forced subtitle tracks are excluded when ignore_forced

Name the subtitle file to be written, based on the source file and the language of the subtitle. Args: file_path: The path to

Return True if any ancestor directory of file_path contains a .subgen_skip marker.

Shift all segment and word timestamps forward by the given offset. This compensates for audio start_time offsets in containers where the

Worker function that processes ASR tasks from the queue. Called by transcription_worker when task type is 'asr'.

Determines the language to be used for transcription based on the provided file path and language preferences. Args: file_path:

Determines the naming format for a subtitle language based on the given type. Args: language (LanguageCode): The language code obje

Worker function that processes detect-language tasks from uploaded audio. Sets the result in the result_container when complete.

Worker function that detects language for a local file. Returns the task data to be queued for transcription.

Extract a segment of audio from in-memory content using FFmpeg. Args: audio_content: Raw audio bytes start_time: Start t

Extract a segment of audio from input_file, starting at start_time for duration seconds. :param input_file: UploadFile object or path to

Extract a specific audio track from a video file to memory using FFmpeg. Args: input_video_path (str): The path to the video file.

Finds the language of the default audio track in the given list of audio tracks. Args: audio_tracks (list): A list of dictionaries c

Checks if an audio track with any of the given languages is present in the list of audio tracks. Returns the first language from `find_langua

Generates subtitles for a video file. Args: file_path: str - The path to the video file. transcription_type: str - The type of tr

Extract a chunk of audio from a file, starting at the given offset and of the given length. :param audio_file: The audio file (UploadFil

Extract language codes from each audio stream in the video file. :param video_path: Path to the video file :return: List of language cod

Use ffprobe to detect the audio stream start_time offset from a video file. Some containers (especially Amazon WEB-DL) have audio stream

Returns the first audio track with the given language. Args: audio_tracks (list): A list of dictionaries containing information

Gets the full path to a file from the Jellyfin server. Args: jellyfin_url: The URL of the Jellyfin server. jellyfin_token: The Jel

Returns True if the file size is stable for a given number of checks.

Checks if any part of the subtitle name matches the target language.

Refreshes the metadata of a Plex library item. Args: itemid: The ID of the item in the Plex library whose metadata needs to be r

Schedule model cleanup with a delay to allow concurrent requests. Properly joins cancelled timers to prevent thread accumulation.

Sends a JSON POST request to a configured webhook URL upon task completion.

Determines if a subtitle file with the target language is available for a specified video file. This function checks both within the video f

Compare the LanguageCode instance to another object. Explicitly handle comparison to None.

ASR endpoint that uses audio content hash for deduplication. BLOCKS until processing is complete, then returns the result. If ident

OpenAI-compatible transcription endpoint (/v1/audio/transcriptions).

OpenAI-compatible translation endpoint (/v1/audio/translations). Always translates to English.

Actually perform the model cleanup.