# Dataset Configuration - Base
name: myanmar_ghost_base
version: 1.0.0

# Data paths
data:
  raw_dir: data/raw
  processed_dir: data/processed
  augmented_dir: data/augmented
  annotations_dir: data/annotations

# Processing settings
processing:
  audio:
    sample_rate: 16000
    n_fft: 512
    hop_length: 160
    n_mels: 80
  
  text:
    max_length: 512
    min_length: 1
    remove_non_myanmar: false

# Split ratios
splits:
  train: 0.8
  val: 0.1
  test: 0.1
  stratify: true

# Label scheme
labels:
  sentiment:
    - negative
    - neutral
    - positive
    - sarcastic

# Data augmentation
augmentation:
  enabled: true
  synonym_replacement:
    enabled: true
    probability: 0.3
  perturbation:
    enabled: true
    probability: 0.1
    types:
      - char_swap
      - word_swap
      - keyboard_typo