src/config.py

"""Configuration for the Ryugaku Gradio app."""
import os

# ---------------------------------------------------------------------------
# Model paths
# ---------------------------------------------------------------------------
# HF Hub repo id for the fine-tuned Japanese -> English translation model.
HF_MODEL_ID = os.environ.get(
    "RYUGAKU_MODEL_ID",
    "build-small-hackathon/ryugaku-jaen-translator-4b",
)

# Local fallback path for development / offline testing.
LEGACY_MODEL_PATH = os.environ.get(
    "RYUGAKU_LOCAL_MODEL",
    "/mnt/github/Ryugaku/legacy_backup/LLM/outputs_qwen35_jaen/merged_16bit",
)

# Device / dtype overrides.
FORCE_CPU = os.environ.get("RYUGAKU_FORCE_CPU", "0") == "1"
DTYPE = os.environ.get("RYUGAKU_DTYPE", "bfloat16")

# ---------------------------------------------------------------------------
# Generation defaults
# ---------------------------------------------------------------------------
DEFAULT_MAX_NEW_TOKENS = int(os.environ.get("RYUGAKU_MAX_NEW_TOKENS", "512"))
DEFAULT_TEMPERATURE = float(os.environ.get("RYUGAKU_TEMPERATURE", "0.1"))

# ---------------------------------------------------------------------------
# System prompts
# ---------------------------------------------------------------------------
TRANSLATION_SYSTEM_PROMPT = (
    "You are a precise Japanese-to-English translator. "
    "Translate the user's Japanese text into natural English only. "
    "Do not add explanations, notes, or romanizations."
)

SUMMARY_SYSTEM_PROMPT = (
    "You are an assistant for international students in Japan. "
    "Summarize the following Japanese text in clear English. "
    "Use bullet points when appropriate. Keep it concise."
)

EASY_ENGLISH_SYSTEM_PROMPT = (
    "You are an English language assistant for international students. "
    "Rewrite the following English text into simple, easy-to-understand English. "
    "Use short sentences and common words. Keep the meaning the same."
)

KEYWORD_SYSTEM_PROMPT = (
    "You are a Japanese language learning assistant. "
    "Extract 5 to 10 important Japanese vocabulary words from the text. "
    "List each word on its own line in this exact format:\n"
    "word (reading) — meaning\n"
    "Do not add introductions, explanations, or numbering."
)

EXAM_SYSTEM_PROMPT = (
    "You are an academic assistant for international students in Japan. "
    "Based on the lecture text, predict 3 to 5 likely exam questions. "
    "List each prediction on its own block in this exact format:\n"
    "Topic: <topic>\nQuestion: <question>\nAnswer: <sample answer>\n\n"
    "Do not add introductions or explanations."
)

OCR_SYSTEM_PROMPT = (
    "You are a translation helper for international students. "
    "The user has uploaded an image of Japanese text. Translate it into English. "
    "If the text is unclear, say so briefly."
)