Create app.py

app.py

@@ -0,0 +1,241 @@
+import gradio as gr
+from openai import OpenAI
+import os
+import tempfile
+from dotenv import load_dotenv
+from markitdown import MarkItDown
+
+load_dotenv()
+api_key = os.getenv("OPENAI_API_KEY")
+api_base = os.getenv("OPENAI_API_BASE")
+# 刪除全域 client，改由 generate_questions 動態初始化
+
+# ✅ 合併多檔案文字
+
+def extract_text_from_files(files):
+    from openai import OpenAI
+    import os
+
+    api_key = os.getenv("OPENAI_API_KEY")
+    api_base = os.getenv("OPENAI_API_BASE")
+    client = OpenAI(api_key=api_key, base_url=api_base)
+
+    image_exts = {".jpg", ".jpeg", ".png", ".bmp", ".gif", ".tiff", ".webp"}
+    merged_text = ""
+    for f in files:
+        ext = os.path.splitext(f.name)[1].lower()
+        if ext in image_exts:
+            md = MarkItDown(llm_client=client, llm_model="gpt-4.1")
+        else:
+            md = MarkItDown()
+        result = md.convert(f.name)
+        merged_text += result.text_content + "\n"
+    return merged_text
+
+# ✅ 產出題目與答案（根據語言與題型）
+
+def generate_questions(files, question_types, num_questions, lang, llm_key, baseurl, model=None):
+    try:
+        text = extract_text_from_files(files)
+        trimmed_text = text[:200000]
+
+        # 優先使用 .env，否則用 UI 傳入值
+        key = os.getenv("OPENAI_API_KEY") or llm_key
+        base = os.getenv("OPENAI_API_BASE") or baseurl
+        model_name = model or "gpt-4.1"
+        if not key or not base:
+            return "⚠️ 請輸入 LLM key 與 baseurl", ""
+        client = OpenAI(api_key=key, base_url=base)
+
+        type_map = {
+            "單選選擇題": {
+                "zh-Hant": "單選選擇題（每題四個選項）",
+                "zh-Hans": "单选选择题（每题四个选项）",
+                "en": "single choice question (4 options)",
+                "ja": "四択問題"
+            },
+            "多選選擇題": {
+                "zh-Hant": "多選選擇題（每題四到五個選項）",
+                "zh-Hans": "多选选择题（每题四到五个选项）",
+                "en": "multiple choice question (4-5 options)",
+                "ja": "複数選択問題"
+            },
+            "問答題": {
+                "zh-Hant": "簡答題",
+                "zh-Hans": "简答题",
+                "en": "short answer",
+                "ja": "短答式問題"
+            },
+            "申論題": {
+                "zh-Hant": "申論題",
+                "zh-Hans": "申论题",
+                "en": "essay question",
+                "ja": "記述式問題"
+            }
+        }
+
+        prompt_map = {
+            "繁體中文": "你是一位專業的出題者，請根據以下內容，設計 {n} 題以下類型的題目：{types}。每題後面請標註【答案】。內容如下：\n{text}",
+            "簡體中文": "你是一位专业的出题者，请根据以下内容，设计 {n} 题以下类型的题目：{types}。每题后面请标注【答案】。内容如下：\n{text}",
+            "English": "You are a professional exam writer. Based on the following content, generate {n} questions of types: {types}. Please mark the answer after each question using [Answer:]. Content:\n{text}",
+            "日本語": "あなたはプロの出題者です。以下の内容に基づいて、{types}を含む{n}問の問題を作成してください。各問題の後に【答え】を付けてください。内容：\n{text}"
+        }
+
+        lang_key_map = {
+            "繁體中文": "zh-Hant",
+            "簡體中文": "zh-Hans",
+            "English": "en",
+            "日本語": "ja"
+        }
+
+        lang_key = lang_key_map[lang]
+        types_str = "、".join([type_map[t][lang_key] for t in question_types])
+        prompt = prompt_map[lang].format(n=num_questions, types=types_str, text=trimmed_text)
+
+        response = client.chat.completions.create(
+            model=model_name,
+            messages=[{"role": "user", "content": prompt}]
+        )
+        content = response.choices[0].message.content
+
+        questions, answers = [], []
+        for line in content.strip().split("\n"):
+            if not line.strip():
+                continue
+            try:
+                if "【答案】" in line:
+                    q, a = line.split("【答案】", 1)
+                elif "[Answer:" in line:
+                    q, a = line.split("[Answer:", 1)
+                    a = a.rstrip("]")
+                elif "【答え】" in line:
+                    q, a = line.split("【答え】", 1)
+                else:
+                    questions.append(line.strip())
+                    answers.append("")
+                    continue
+                questions.append(q.strip())
+                answers.append(a.strip())
+            except Exception:
+                questions.append(line.strip())
+                answers.append("")
+
+        if not questions:
+            return "⚠️ 無法解析 AI 回傳內容，請��查輸入內容或稍後再試。", ""
+
+        return "\n\n".join(questions), "\n\n".join(answers)
+    except Exception as e:
+        return f"⚠️ 發生錯誤：{str(e)}", ""
+
+# ✅ 匯出 Markdown, Quizlet（TSV）
+
+def export_files(questions_text, answers_text):
+    md_path = tempfile.NamedTemporaryFile(delete=False, suffix=".md").name
+    with open(md_path, "w", encoding="utf-8") as f:
+        f.write("# 📘 題目 Questions\n\n" + questions_text + "\n\n# ✅ 解答 Answers\n\n" + answers_text)
+
+    quizlet_path = tempfile.NamedTemporaryFile(delete=False, suffix=".tsv").name
+    with open(quizlet_path, "w", encoding="utf-8") as f:
+        for q, a in zip(questions_text.split("\n\n"), answers_text.split("\n\n")):
+            q_clean = q.replace("\n", " ").replace("\r", " ")
+            a_clean = a.replace("\n", " ").replace("\r", " ")
+            f.write(f"{q_clean}\t{a_clean}\n")
+
+    return md_path, quizlet_path
+
+# ✅ Gradio UI
+
+# --- FastAPI + Gradio 整合 ---
+from fastapi import FastAPI, UploadFile, File, Form
+from fastapi.responses import JSONResponse
+from typing import List, Optional
+import uvicorn
+
+def build_gradio_blocks():
+    with gr.Blocks() as demo:
+        gr.Markdown("# 📄 通用 AI 出題系統（支援多檔、多語、匯出格式）")
+
+        with gr.Row():
+            with gr.Column():
+                file_input = gr.File(
+                    label="上傳文件（可多檔）",
+                    file_types=[
+                        ".pdf", ".ppt", ".pptx", ".doc", ".docx", ".xls", ".xlsx", ".csv",
+                        ".jpg", ".jpeg", ".png", ".bmp", ".gif", ".tiff", ".webp",
+                        ".mp3", ".wav", ".m4a", ".flac", ".ogg", ".aac", ".amr", ".wma", ".opus",
+                        ".html", ".htm", ".json", ".xml", ".txt", ".md", ".rtf", ".log",
+                        ".zip", ".epub"
+                    ],
+                    file_count="multiple"
+                )
+                lang = gr.Dropdown(["繁體中文", "簡體中文", "English", "日本語"], value="繁體中文", label="語言 Language")
+                question_types = gr.CheckboxGroup(["單選選擇題", "多選選擇題", "問答題", "申論題"],
+                                                  label="選擇題型（可複選）",
+                                                  value=["單選選擇題"])
+                num_questions = gr.Slider(1, 20, value=10, step=1, label="題目數量")
+                llm_key = gr.Textbox(label="LLM Key (不會儲存)", type="password", placeholder="請輸入你的 OpenAI API Key")
+                baseurl = gr.Textbox(label="Base URL (如 https://api.openai.com/v1)",value="https://api.openai.com/v1", placeholder="請輸入 API Base URL")
+                model_box = gr.Textbox(label="Model 名稱", value="gpt-4.1", placeholder="如 gpt-4.1, gpt-3.5-turbo, ...")
+                generate_btn = gr.Button("✏️ 開始出題")
+
+            with gr.Column():
+                qbox = gr.Textbox(label="📘 題目 Questions", lines=15)
+                abox = gr.Textbox(label="✅ 解答 Answers", lines=15)
+                export_btn = gr.Button("📤 匯出 Markdown / Quizlet")
+                md_out = gr.File(label="📝 Markdown 檔下載")
+                quizlet_out = gr.File(label="📋 Quizlet (TSV) 檔下載")
+
+
+        generate_btn.click(fn=generate_questions,
+                           inputs=[file_input, question_types, num_questions, lang, llm_key, baseurl, model_box],
+                           outputs=[qbox, abox])
+
+        export_btn.click(fn=export_files,
+                         inputs=[qbox, abox],
+                         outputs=[md_out, quizlet_out])
+    return demo
+
+if __name__ == "__main__":
+    demo = build_gradio_blocks()
+    demo.launch()
+
+# --- FastAPI API 介面 ---
+from fastapi import FastAPI, UploadFile, File, Form
+from fastapi.responses import JSONResponse
+from typing import List, Optional
+import uvicorn
+
+api_app = FastAPI(title="AI 出題系統 API")
+
+@api_app.post("/api/generate")
+async def api_generate(
+    files: List[UploadFile] = File(...),
+    question_types: List[str] = Form(...),
+    num_questions: int = Form(...),
+    lang: str = Form(...),
+    llm_key: Optional[str] = Form(None),
+    baseurl: Optional[str] = Form(None),
+    model: Optional[str] = Form(None)
+):
+    # 將 UploadFile 轉為臨時檔案物件，與 Gradio 行為一致
+    temp_files = []
+    for f in files:
+        temp = tempfile.NamedTemporaryFile(delete=False)
+        temp.write(await f.read())
+        temp.flush()
+        temp_files.append(temp)
+        temp.name = temp.name  # 保持介面一致
+
+    # 呼叫原本的出題邏輯
+    questions, answers = generate_questions(
+        temp_files, question_types, num_questions, lang, llm_key, baseurl, model
+    )
+
+    # 關閉臨時檔案
+    for temp in temp_files:
+        temp.close()
+
+    return JSONResponse({"questions": questions, "answers": answers})
+
+# 若要啟動 API 伺服器，請執行：
+# uvicorn app:api_app --host 0.0.0.0 --port 7861