workbook_server.py

import base64
import uvicorn
from contextlib import asynccontextmanager
import sqlite3
from fastapi import FastAPI, HTTPException
from fastapi.responses import JSONResponse
from pydantic import BaseModel
from openai import OpenAI
import os

# OpenAI APIクライアントを初期化
client = OpenAI()
client.api_key = os.getenv('OPENAI_API_KEY')

@asynccontextmanager
async def lifespan(app: FastAPI):
    print("startup event")
    yield
    print("shutdown event")

app = FastAPI(lifespan=lifespan)

DATABASE = 'workbook.db'

def get_db():
    db = getattr(app.state, '_database', None)
    if db is None:
        db = app.state._database = sqlite3.connect(DATABASE)
    return db

@app.get('/api/chapters')
def get_chapters():
    db = get_db()
    cursor = db.cursor()
    cursor.execute("""

        SELECT c.chapter, c.chapter_name, COUNT(q.question) as n_questions

        FROM chapter c

        LEFT JOIN question q ON c.chapter = q.chapter

        GROUP BY c.chapter, c.chapter_name

    """)
    chapters = cursor.fetchall()
    chapter_list = [
        {
            "chapter_number": row[0],
            "chapter_name": row[1],
            "n_questions": row[2]
        } for row in chapters
    ]
    return JSONResponse(content=chapter_list)

class ChapterRequest(BaseModel):
    chapter_number: int

@app.post('/api/questions')
def get_questions_by_chapter(request: ChapterRequest):
    db = get_db()
    cursor = db.cursor()
    cursor.execute("SELECT question, options, answer, explain, image FROM question WHERE chapter = ? AND kind = '正誤問題'", (request.chapter_number,))
    questions = cursor.fetchall()
    if not questions:
        raise HTTPException(status_code=404, detail="Questions not found for the given chapter number")
    question_list = []
    for row in questions:
        if row[4] is None:
            encoded_string = ''
        else:
            image_path = os.path.join('./image', row[4])
            with open(image_path, "rb") as image_file:
                encoded_string = base64.b64encode(image_file.read()).decode('utf-8')
        question_list.append({
            "question_text": row[0],
            "options": row[1],
            "answer": row[2],
            "explanation": row[3],
            "image": encoded_string,
        })
    return JSONResponse(content=question_list)

class HintRequest(BaseModel):
    question_text: str
    options: str
    answer: str

@app.post('/api/hint')
def generate_hint(request: HintRequest):
    prompt = f"設問: {request.question_text}\n選択肢: {request.options}\n正解: {request.answer}\nこの設問に対するヒントを生成してください。なおレスポンスはヒント文章のみとせよ。冒頭に「ヒント：」などの見出しをつけるな。"
    print(prompt)
    try:
        response = client.chat.completions.create(
            model="gpt-4o-mini",
            messages=[
                {
                    "role": "user",
                    "content": [{'type': 'text', 'text': prompt}],
                },
            ],
        )
        response_content = response.choices[0].message.content
        return JSONResponse(content={"hint": response_content})
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

if __name__ == '__main__':
    uvicorn.run(app, host="127.0.0.1", port=8000, log_level="debug")