wiydarrr's picture
Upload folder using huggingface_hub
5f685fd verified

A newer version of the Gradio SDK is available: 5.9.1

Upgrade

Module: engine

This module contains the main engine classes for generating different types of short videos. There are four main engine classes in this module:

  • AbstractContentEngine: This is an abstract base class that provides the basic functionalities and attributes required by all content engines. It implements common methods for initializing the content engine, preparing editing paths, verifying parameters, and rendering the short video.

  • ContentShortEngine: This class extends AbstractContentEngine and is used for generating general content short videos. It implements specific methods for generating a script, generating temporary audio, speeding up the audio, timing captions, generating image search terms, generating image URLs, choosing background music and video, and preparing background and custom assets. It also overrides the __generateScript method to generate the script for the content short video.

  • ContentVideoEngine: This class extends AbstractContentEngine and is used for generating general content videos. It implements specific methods for generating temporary audio, speeding up the audio, timing captions, generating video search terms, generating video URLs, choosing background music, and preparing background and custom assets.

  • FactsShortEngine: This class extends ContentShortEngine and is used for generating facts short videos. It overrides the _generateScript method to generate the script for the facts short video.

  • RedditShortEngine: This class extends ContentShortEngine and is used for generating reddit short videos. It overrides the _generateScript method to generate the script for the reddit short video and adds a custom step for preparing a reddit image.


File: abstract_content_engine.py

This file contains the AbstractContentEngine class, which is an abstract base class for all content engines. It provides the basic functionalities and attributes required by all content engines.

Class: AbstractContentEngine

Attributes:

  • CONTENT_DB: An instance of the ContentDatabase class, which is used to store and retrieve content data.

Methods:

  • __init__(self, short_id: str, content_type:str, language: Language, voiceName: str): Initializes an instance of the AbstractContentEngine class with the given parameters. It sets the dataManager, id, _db_language, voiceModule, assetStore, stepDict, and logger attributes.

  • __getattr__(self, name): Overrides the __getattr__ method to retrieve attributes that start with 'db' from the dataManager.

  • __setattr__(self, name, value): Overrides the __setattr__ method to save attributes that start with 'db' to the dataManager.

  • prepareEditingPaths(self): Creates the directory for storing dynamic assets if it doesn't already exist.

  • verifyParameters(*args, **kwargs): Verifies that all the required parameters are not null. If any parameter is null, it raises an exception.

  • isShortDone(self): Checks if the short video is done rendering by checking the value of the '_db_ready_to_upload' attribute.

  • makeContent(self): Generates the short video by executing the steps defined in the stepDict. It yields the current step number and a message indicating the progress.

  • get_video_output_path(self): Returns the path of the rendered video.

  • get_total_steps(self): Returns the total number of steps in the stepDict.

  • set_logger(self, logger): Sets the logger function for logging the progress of the short video rendering.

  • initializeMagickAndFFMPEG(self): Initializes the paths for FFmpeg, FFProbe, and ImageMagick. If any of these programs are not found, it raises an exception.


File: content_short_engine.py

This file contains the ContentShortEngine class, which is used for generating general content short videos. It extends the AbstractContentEngine class and adds specific methods for generating a script, generating temporary audio, speeding up the audio, timing captions, generating image search terms, generating image URLs, choosing background music and video, and preparing background and custom assets.

Class: ContentShortEngine

Attributes:

  • stepDict: A dictionary that maps step numbers to their corresponding methods for generating the short video.

Methods:

  • __init__(self, short_type: str, background_video_name: str, background_music_name: str, short_id="", num_images=None, watermark=None, language: Language = Language.ENGLISH, voiceName=""): Initializes an instance of the ContentShortEngine class with the given parameters. It sets the stepDict attribute with the specific methods for generating the short video.

  • __generateScript(self): Abstract method that generates the script for the content short video. This method needs to be implemented by the child classes.

  • __prepareCustomAssets(self): Abstract method that prepares the custom assets for the content short video. This method needs to be implemented by the child classes.

  • __editAndRenderShort(self): Abstract method that performs the editing and rendering of the content short video. This method needs to be implemented by the child classes.


File: content_video_engine.py

This file contains the ContentVideoEngine class, which is used for generating general content videos. It extends the AbstractContentEngine class and adds specific methods for generating temporary audio, speeding up the audio, timing captions, generating video search terms, generating video URLs, choosing background music, and preparing background and custom assets.

Class: ContentVideoEngine

Methods:

  • __generateTempAudio(self): Generates the temporary audio for the content video by using the voiceModule to generate a voice from the script.

  • __speedUpAudio(self): Speeds up the temporary audio to match the duration of the background video.

  • __timeCaptions(self): Converts the audio to text and then generates captions with time based on the text.

  • __generateVideoSearchTerms(self): Generates the video search terms by using the timed captions.

  • __generateVideoUrls(self): Generates the video URLs by using the video search terms and the getBestVideo function from the pexels_api.

  • __chooseBackgroundMusic(self): Retrieves the background music URL from the assetStore based on the background music name.

  • __prepareBackgroundAssets(self): Prepares the background assets for the content video by retrieving the voiceover audio duration, trimming the background video, and extracting a random clip from the background video.

  • __prepareCustomAssets(self): Abstract method that prepares the custom assets for the content video. This method needs to be implemented by the child classes.

  • __editAndRenderShort(self): Performs the editing and rendering of the content video by using the videoEditor and the editing steps defined in the stepDict.


File: facts_short_engine.py

This file contains the FactsShortEngine class, which is used for generating facts short videos. It extends the ContentShortEngine class and overrides the _generateScript method to generate the script for the facts short video.

Class: FactsShortEngine

Methods:

  • _generateScript(self): Generates the script for the facts short video by using the generateFacts function from the facts_gpt module.

File: reddit_short_engine.py

This file contains the RedditShortEngine class, which is used for generating reddit short videos. It extends the ContentShortEngine class and overrides the _generateScript method to generate the script for the reddit short video. It also adds a custom step for preparing a reddit image.

Class: RedditShortEngine

Methods:

  • _generateScript(self): Generates the script for the reddit short video by using the getInterestingRedditQuestion function from the reddit_gpt module.

  • _prepareCustomAssets(self): Prepares the custom assets for the reddit short video by using the ingestFlow method from the imageEditingEngine to create a reddit image.

  • _editAndRenderShort(self): Performs the editing and rendering of the reddit short video by using the videoEditor and the editing steps defined in the stepDict.