Create app.py

app.py

@@ -0,0 +1,156 @@
+import gradio as gr
+from sentence_transformers import SentenceTransformer
+import faiss
+import os
+
+# ============================
+# 1) 대화 데이터 로드
+# ============================
+DATA_FILE = "data/conversation.txt"
+
+if not os.path.exists(DATA_FILE):
+    os.makedirs("data", exist_ok=True)
+    with open(DATA_FILE, "w", encoding="utf-8") as f:
+        f.write("여기에 대화 내용을 넣으세요.\n")
+
+with open(DATA_FILE, "r", encoding="utf-8") as f:
+    text = f.read()
+
+# ============================
+# 2) 청크 처리
+# ============================
+def chunk_text(text, chunk_size=500):
+    lines = text.split("\n")
+    chunks, cur, length = [], [], 0
+
+    for line in lines:
+        line = line.strip()
+        if not line:
+            continue
+        if length + len(line) > chunk_size:
+            chunks.append("\n".join(cur))
+            cur = []
+            length = 0
+        cur.append(line)
+        length += len(line)
+
+    if cur:
+        chunks.append("\n".join(cur))
+
+    return chunks
+
+chunks = chunk_text(text)
+
+# ============================
+# 3) 임베딩 + 인덱스
+# ============================
+model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
+embeddings = model.encode(chunks)
+
+dimension = embeddings.shape[1]
+index = faiss.IndexFlatL2(dimension)
+index.add(embeddings)
+
+def retrieve(query, top_k=3):
+    qvec = model.encode([query])
+    dist, idxs = index.search(qvec, top_k)
+    return "\n\n".join(chunks[i] for i in idxs[0] if 0 <= i < len(chunks))
+
+# ============================
+# 4) LLM API 준비
+# ============================
+import openai
+import google.generativeai as genai
+import requests
+import json
+
+# OPENAI (GPT)
+if os.getenv("OPENAI_API_KEY"):
+    openai.api_key = os.getenv("OPENAI_API_KEY")
+
+# GEMINI
+if os.getenv("GEMINI_API_KEY"):
+    genai.configure(api_key=os.getenv("GEMINI_API_KEY"))
+
+# GROK (xAI)
+GROK_KEY = os.getenv("GROK_API_KEY")
+GROK_URL = "https://api.x.ai/v1/chat/completions"
+
+# ============================
+# 5) 모델별 호출 함수
+# ============================
+def call_gpt(prompt):
+    response = openai.ChatCompletion.create(
+        model="gpt-4o-mini",
+        messages=[{"role": "user", "content": prompt}]
+    )
+    return response["choices"][0]["message"]["content"]
+
+def call_gemini(prompt):
+    model = genai.GenerativeModel("gemini-1.5-flash")
+    response = model.generate_content(prompt)
+    return response.text
+
+def call_grok(prompt):
+    headers = {
+        "Authorization": f"Bearer {GROK_KEY}",
+        "Content-Type": "application/json"
+    }
+    data = {
+        "model": "grok-1.5",
+        "messages": [{"role": "user", "content": prompt}]
+    }
+    r = requests.post(GROK_URL, headers=headers, data=json.dumps(data))
+    out = r.json()
+    return out["choices"][0]["message"]["content"]
+
+# ============================
+# 6) 메인 챗 함수
+# ============================
+def chat(query, history, model_name):
+    related = retrieve(query)
+
+    prompt = f"""
+당신은 업스케일/ONNX/TensorRT 전문가입니다.
+아래는 프로젝트 핵심 문맥입니다:
+
+{related}
+
+위 문맥을 참고해 다음 질문에 답하세요:
+
+질문: {query}
+"""
+
+    if model_name == "GPT-4o-mini":
+        answer = call_gpt(prompt)
+    elif model_name == "Gemini":
+        answer = call_gemini(prompt)
+    elif model_name == "Grok":
+        answer = call_grok(prompt)
+    else:
+        answer = "지원되지 않는 LLM입니다."
+
+    history.append((query, answer))
+    return history, history
+
+# ============================
+# 7) UI
+# ============================
+with gr.Blocks(title="Upscale Frog 🐸 RAG") as app:
+    gr.Markdown("## 🐸 Upscale Frog — Multi-LLM RAG Engine")
+    
+    with gr.Row():
+        llm_selector = gr.Dropdown(
+            ["GPT-4o-mini", "Gemini", "Grok"],
+            value="GPT-4o-mini",
+            label="LLM 엔진 선택"
+        )
+    
+    chatbot = gr.Chatbot()
+    msg = gr.Textbox(label="메시지 입력")
+    clear = gr.ClearButton()
+
+    msg.submit(chat, [msg, chatbot, llm_selector], [chatbot, chatbot])
+    msg.submit(lambda: "", None, msg)
+
+app.launch()