backend/src/config.py

"""Configuration settings for EUMORA lyrical analysis."""

import os
from dataclasses import dataclass, field
from pathlib import Path
from typing import Dict, List, Optional


@dataclass
class Config:
    """Training and model configuration."""

    # Paths
    # EUMORA_ROOT can be set to the repo root when running from outside backend/
    project_root: Path = field(
        default_factory=lambda: Path(
            os.environ.get("EUMORA_ROOT", str(Path(__file__).parent.parent))
        )
    )

    @property
    def data_dir(self) -> Path:
        path = self.project_root / "data"
        path.mkdir(exist_ok=True)
        return path

    @property
    def model_dir(self) -> Path:
        path = self.project_root / "models"
        path.mkdir(exist_ok=True)
        return path

    # Model settings - using DeBERTa-v3-Base for superior emotion understanding
    model_name: str = "microsoft/deberta-v3-base"
    max_length: int = 256  # Increased for longer lyrics/text

    # Training settings
    batch_size: int = 16
    learning_rate: float = 2e-5
    num_epochs: int = 4
    warmup_ratio: float = 0.1
    weight_decay: float = 0.01

    # Emotion labels (8 emotions including neutral and sarcasm)
    label_names: tuple = ("sadness", "joy", "love", "anger", "fear", "surprise", "neutral", "sarcasm")
    num_labels: int = 8

    # Sarcasm calibration defaults for lyric-domain inference
    target_sarcasm_prior: float = 0.15
    assumed_train_sarcasm_prior: float = 0.5
    sarcasm_threshold: Optional[float] = None

    # Dataset options
    hf_dataset_name: str = "dair-ai/emotion"
    goemotions_name: str = "google-research-datasets/go_emotions"

    # Emotion to music context mapping
    @property
    def emotion_to_music_mood(self) -> Dict:
        return {
            "sadness": {"mood": "melancholic", "energy": "low", "valence": "negative"},
            "joy": {"mood": "happy", "energy": "high", "valence": "positive"},
            "love": {"mood": "romantic", "energy": "medium", "valence": "positive"},
            "anger": {"mood": "intense", "energy": "high", "valence": "negative"},
            "fear": {"mood": "anxious", "energy": "medium", "valence": "negative"},
            "surprise": {"mood": "excited", "energy": "high", "valence": "mixed"},
            "neutral": {"mood": "calm", "energy": "low", "valence": "neutral"},
            "sarcasm": {"mood": "ironic", "energy": "medium", "valence": "mixed"},
        }


# Global config instance
config = Config()