Update app.py

app.py

@@ -555,7 +555,7 @@ def ginigen_app():
 
     # Set default session state
     if "ai_model" not in st.session_state:
-        st.session_state.ai_model = "gpt-4.1-mini"  # 기본 모델로 설정
+        st.session_state.ai_model = "gpt-4.1-mini"  # 고정 모델 설정
     if "messages" not in st.session_state:
         st.session_state.messages = []
     if "auto_save" not in st.session_state:
@@ -575,9 +575,7 @@ def ginigen_app():
     sb = st.sidebar
     sb.title("Blog Settings")
     
-    # 모델 선택 추가
-    available_models = ["gpt-4.1-mini", "gpt-4o", "gpt-4-turbo"]
-    sb.selectbox("AI Model", available_models, key="ai_model")
+    # 모델 선택 제거 (고정 모델 사용)
     
     sb.subheader("Blog Style Settings")
     sb.selectbox(
@@ -595,6 +593,7 @@ def ginigen_app():
     )
     
     sb.slider("Blog Length (word count)", 800, 3000, key="word_count")
+
     
     # Example topics
     sb.subheader("Example Topics")
@@ -734,7 +733,8 @@ def process_input(prompt: str, uploaded_files):
     
     with st.chat_message("assistant"):
         placeholder = st.empty()
-        answer = ""
+        message_placeholder = st.empty()
+        full_response = ""
 
         use_web_search = st.session_state.web_search_enabled
         has_uploaded_files = bool(uploaded_files) and len(uploaded_files) > 0
@@ -742,7 +742,7 @@ def process_input(prompt: str, uploaded_files):
         try:
             # 상태 표시를 위한 상태 컴포넌트
             status = st.status("Preparing to generate blog...")
-            status.update(label="Initializing OpenAI client...")
+            status.update(label="Initializing client...")
             
             client = get_openai_client()
             
@@ -764,7 +764,7 @@ def process_input(prompt: str, uploaded_files):
                     file_content = process_uploaded_files(uploaded_files)
             
             # Build system prompt
-            status.update(label="Preparing system prompt...")
+            status.update(label="Preparing blog draft...")
             sys_prompt = get_system_prompt(
                 template=st.session_state.blog_template,
                 tone=st.session_state.blog_tone,
@@ -774,7 +774,7 @@ def process_input(prompt: str, uploaded_files):
             )
 
             # OpenAI API 호출 준비
-            status.update(label="Generating blog content...")
+            status.update(label="Writing blog content...")
             
             # 메시지 구성
             api_messages = [
@@ -794,98 +794,47 @@ def process_input(prompt: str, uploaded_files):
             # 사용자 메시지 추가
             api_messages.append({"role": "user", "content": user_content})
             
-            # OpenAI API 호출 - 재시도 로직 추가
-            max_retries = 3
-            retry_delay = 2  # 초 단위
-            answer = None
-            used_model = st.session_state.ai_model
-            
-            for retry in range(max_retries):
-                try:
-                    status.update(label=f"Generating blog with {used_model} (attempt {retry+1}/{max_retries})...")
-                    
-                    # 타임아웃 인자 없이 API 호출
-                    response = client.chat.completions.create(
-                        model=used_model,
-                        messages=api_messages,
-                        temperature=1,
-                        max_tokens=MAX_TOKENS,
-                        top_p=1
-                    )
-                    
-                    # 응답 추출
-                    answer = response.choices[0].message.content
-                    status.update(label=f"Blog generated successfully with {used_model}!", state="complete")
-                    break  # 성공하면 반복문 종료
-                    
-                except Exception as api_error:
-                    error_message = str(api_error)
-                    logging.error(f"OpenAI API error (attempt {retry+1}/{max_retries}): {error_message}")
-                    
-                    if retry < max_retries - 1:
-                        wait_time = retry_delay * (retry + 1)  # 점점 길어지는 대기 시간
-                        status.update(label=f"API Error: {error_message}. Retrying in {wait_time}s... ({retry+1}/{max_retries})")
-                        time.sleep(wait_time)
-                    else:
-                        # 마지막 시도에서도 실패하면 대체 모델 시도
-                        fallback_models = ["gpt-3.5-turbo", "gpt-3.5-turbo-16k"]
-                        if used_model in fallback_models:
-                            fallback_models.remove(used_model)
-                        
-                        # 대체 모델 시도
-                        if fallback_models:
-                            status.update(label="Primary model failed. Trying fallback models...")
-                            for fb_model in fallback_models:
-                                try:
-                                    status.update(label=f"Trying with fallback model: {fb_model}...")
-                                    response = client.chat.completions.create(
-                                        model=fb_model,
-                                        messages=api_messages,
-                                        temperature=1,
-                                        max_tokens=min(MAX_TOKENS, 4096 if "16k" not in fb_model else 16000),
-                                        top_p=1
-                                    )
-                                    answer = response.choices[0].message.content
-                                    used_model = fb_model
-                                    status.update(label=f"Blog generated with fallback model: {used_model}", state="complete")
-                                    break
-                                except Exception as fb_error:
-                                    logging.error(f"Fallback model {fb_model} also failed: {str(fb_error)}")
-                            
-                            # 대체 모델 중 하나가 성공했는지 확인
-                            if answer:
-                                break
-                        
-                        # 모든 모델이 실패한 경우
-                        if not answer:
-                            status.update(label="All models failed to generate content", state="error")
-                            user_message = "OpenAI API 연결 문제가 발생했습니다. 인터넷 연결을 확인하거나 나중에 다시 시도해 주세요."
-                            if "rate limit" in error_message.lower():
-                                user_message = "OpenAI API 속도 제한에 도달했습니다. 잠시 후 다시 시도해 주세요."
-                            elif "invalid api key" in error_message.lower():
-                                user_message = "유효하지 않은 API 키입니다. API 키를 확인해 주세요."
-                            
-                            raise Exception(user_message)
-            
-            # 결과가 생성되지 않은 경우 오류 발생
-            if not answer:
-                raise Exception("블로그 콘텐츠를 생성하지 못했습니다.")
+            # OpenAI API 스트리밍 호출 - 고정 모델 "gpt-4.1-mini" 사용
+            try:
+                # 스트리밍 방식으로 API 호출
+                stream = client.chat.completions.create(
+                    model="gpt-4.1-mini",  # 고정 모델 사용
+                    messages=api_messages,
+                    temperature=1,
+                    max_tokens=MAX_TOKENS,
+                    top_p=1,
+                    stream=True  # 스트리밍 활성화
+                )
+                
+                # 스트리밍 응답 처리
+                for chunk in stream:
+                    if chunk.choices and len(chunk.choices) > 0 and chunk.choices[0].delta.content is not None:
+                        content_delta = chunk.choices[0].delta.content
+                        full_response += content_delta
+                        message_placeholder.markdown(full_response + "▌")
+                
+                # 최종 응답 표시 (커서 제거)
+                message_placeholder.markdown(full_response)
+                status.update(label="Blog completed!", state="complete")
+                
+            except Exception as api_error:
+                error_message = str(api_error)
+                logging.error(f"API error: {error_message}")
+                status.update(label=f"Error: {error_message}", state="error")
+                raise Exception(f"Blog generation error: {error_message}")
             
-            # 결과 표시
-            placeholder.markdown(answer)
-
             # 이미지 생성
             answer_entry_saved = False
-            if st.session_state.generate_image and answer:
+            if st.session_state.generate_image and full_response:
                 with st.spinner("Generating image..."):
                     try:
-                        ip = extract_image_prompt(answer, prompt)
+                        ip = extract_image_prompt(full_response, prompt)
                         img, cap = generate_image(ip)
                         if img:
                             st.image(img, caption=cap)
                             st.session_state.messages.append({
                                 "role": "assistant",
-                                "content": answer,
+                                "content": full_response,
                                 "image": img,
                                 "image_caption": cap
                             })
@@ -895,22 +844,22 @@ def process_input(prompt: str, uploaded_files):
                         st.warning("이미지 생성에 실패했습니다. 블로그 콘텐츠만 저장됩니다.")
 
             # Save the answer if not saved above
-            if not answer_entry_saved and answer:
-                st.session_state.messages.append({"role": "assistant", "content": answer})
+            if not answer_entry_saved and full_response:
+                st.session_state.messages.append({"role": "assistant", "content": full_response})
 
             # Download buttons
-            if answer:
+            if full_response:
                 st.subheader("Download This Blog")
                 c1, c2 = st.columns(2)
                 c1.download_button(
                     "Markdown", 
-                    data=answer, 
+                    data=full_response, 
                     file_name=f"{prompt[:30]}.md",
                     mime="text/markdown"
                 )
                 c2.download_button(
                     "HTML",
-                    data=md_to_html(answer, prompt[:30]),
+                    data=md_to_html(full_response, prompt[:30]),
                     file_name=f"{prompt[:30]}.html",
                     mime="text/html"
                 )
@@ -932,7 +881,6 @@ def process_input(prompt: str, uploaded_files):
             st.session_state.messages.append({"role": "assistant", "content": ans})
 
 
-
 # ──────────────────────────────── main ────────────────────────────────────
 def main():
     ginigen_app()