Data_Processing_Pipeline.md

graph LR
    Preprocessing_Orchestrator["Preprocessing Orchestrator"]
    Processor_Interface["Processor Interface"]
    Concrete_Processors["Concrete Processors"]
    Preprocessing_Orchestrator -- "Uses" --> Concrete_Processors
    Concrete_Processors -- "Inherits From" --> Processor_Interface

Loading

Details

One paragraph explaining the functionality which is represented by this graph. What the main flow is and what is its purpose.

Preprocessing Orchestrator

The primary entry point for the data preparation workflow. It parses configuration (e.g., dataset name, paths), selects and instantiates the appropriate Concrete Processor, and drives the overall preprocessing execution.

Related Classes/Methods:

• tensorflow_tts/bin/preprocess.py

Processor Interface

An abstract base class (BaseProcessor) that defines the standard contract for all data processors. It establishes the sequence of operations (the "template method") for converting raw data into features, such as loading text, cleaning text, and saving outputs.

Related Classes/Methods:

• tensorflow_tts/processor/base_processor.py

Concrete Processors

A collection of dataset-specific classes that implement the Processor Interface. Each class provides the concrete logic for handling a specific dataset's format, such as LJSpeech, Baker, or KSS. They function as plugins to the main orchestrator.

Related Classes/Methods:

• tensorflow_tts/processor/ljspeech.py
• tensorflow_tts/processor/baker.py
• tensorflow_tts/processor/kss.py

FAQ