Update app.py

app.py

@@ -6,6 +6,7 @@ import tempfile
 from collections.abc import Iterator
 from threading import Thread
 
+import requests  # <-- For SERPHouse web search
 import cv2
 import gradio as gr
 import spaces
@@ -20,6 +21,54 @@ import pandas as pd
 # PDF 텍스트 추출
 import PyPDF2
 
+##############################################################################
+# SERPHouse API key for web search
+##############################################################################
+SERPHOUSE_API_KEY = "V38CNn4HXpLtynJQyOeoUensTEYoFy8PBUxKpDqAW1pawT1vfJ2BWtPQ98h6"
+
+##############################################################################
+# Simple function to call the SERPHouse Live endpoint
+# https://api.serphouse.com/serp/live
+##############################################################################
+def do_web_search(query: str) -> str:
+    """
+    Calls SERPHouse live endpoint with the given query (q).
+    Returns a simple text summary or error message.
+    """
+    try:
+        url = "https://api.serphouse.com/serp/live"
+        params = {
+            "q": query,
+            "domain": "google.com",
+            "lang": "en",
+            "device": "desktop",
+            "serp_type": "web",
+            "api_token": SERPHOUSE_API_KEY,
+        }
+        resp = requests.get(url, params=params, timeout=30)
+        resp.raise_for_status()  # Raise an exception for 4xx/5xx errors
+        data = resp.json()
+
+        # For demonstration, let's extract top 3 organic results:
+        results = data.get("results", {})
+        organic = results.get("results", {}).get("organic", [])
+        if not organic:
+            return "No web search results found."
+
+        summary_lines = []
+        for item in organic[:3]:
+            rank = item.get("position", "-")
+            title = item.get("title", "No Title")
+            link = item.get("link", "No Link")
+            snippet = item.get("snippet", "(No snippet)")
+            summary_lines.append(f"**Rank {rank}:** [{title}]({link})\n\n> {snippet}")
+
+        return "\n\n".join(summary_lines) if summary_lines else "No web search results found."
+    except Exception as e:
+        logger.error(f"Web search failed: {e}")
+        return f"Web search failed: {str(e)}"
+
+
 MAX_CONTENT_CHARS = 4000  # 너무 큰 파일을 막기 위해 최대 표시 4000자
 
 model_id = os.getenv("MODEL_ID", "google/gemma-3-27b-it")
@@ -350,12 +399,31 @@ def process_history(history: list[dict]) -> list[dict]:
 # 메인 추론 함수
 ##################################################
 @spaces.GPU(duration=120)
-def run(message: dict, history: list[dict], system_prompt: str = "", max_new_tokens: int = 512) -> Iterator[str]:
+def run(
+    message: dict,
+    history: list[dict],
+    system_prompt: str = "",
+    max_new_tokens: int = 512,
+    use_web_search: bool = False,
+    web_search_query: str = "",
+) -> Iterator[str]:
+    """
+    The main inference function. Now extended with optional web_search arguments:
+    - use_web_search: bool
+    - web_search_query: str
+    If `use_web_search` is True, calls SERPHouse for the given `web_search_query`.
+    """
+    # Validate media constraints first
     if not validate_media_constraints(message, history):
         yield ""
         return
 
     try:
+        # If user opted for "Web Search", do it here and yield a prefix message
+        if use_web_search and web_search_query.strip():
+            ws_result = do_web_search(web_search_query.strip())
+            yield f"**[Web Search Results for '{web_search_query.strip()}':]**\n\n{ws_result}\n\n---\n"
+
         messages = []
         if system_prompt:
             messages.append({"role": "system", "content": [{"type": "text", "text": system_prompt}]})
@@ -372,7 +440,6 @@ def run(message: dict, history: list[dict], system_prompt: str = "", max_new_tok
         messages.append({"role": "user", "content": user_content})
 
         # 모델 입력 생성 전 최종 확인
-        # 이미지나 비디오가 아닌 파일들은 모델의 "image" 파이프라인으로 전달되지 않도록 필터링
         for msg in messages:
             if msg["role"] != "user":
                 continue
@@ -426,8 +493,6 @@ def run(message: dict, history: list[dict], system_prompt: str = "", max_new_tok
 
 
 
-
-
 ##################################################
 # 예시들 (한글화 버전)
 ##################################################
@@ -536,41 +601,161 @@ examples = [
 ]
 
 
-##################################################
-# Gradio 인터페이스 설정
-##################################################
-demo = gr.ChatInterface(
-    fn=run,
-    type="messages",
-    chatbot=gr.Chatbot(type="messages", scale=1, allow_tags=["image"]),
-    # .webp, .png, .jpg, .jpeg, .gif, .mp4, .csv, .txt, .pdf 모두 허용
-    textbox=gr.MultimodalTextbox(
-        file_types=[
-            ".webp", ".png", ".jpg", ".jpeg", ".gif",
-            ".mp4", ".csv", ".txt", ".pdf"
-        ],
-        file_count="multiple",
-        autofocus=True
-    ),
-    multimodal=True,
-    additional_inputs=[
-        gr.Textbox(
-            label="System Prompt",
-            value=(
-                "You are a deeply thoughtful AI. Consider problems thoroughly and derive "
-                "correct solutions through systematic reasoning. Please answer in korean."
+
+
+##############################################################################
+# Custom CSS similar to second example (colorful background, panel, etc.)
+##############################################################################
+css = """
+body {
+    background: linear-gradient(135deg, #667eea, #764ba2);
+    font-family: 'Helvetica Neue', Helvetica, Arial, sans-serif;
+    color: #333;
+    margin: 0;
+    padding: 0;
+}
+.gradio-container {
+    background: rgba(255, 255, 255, 0.95);
+    border-radius: 15px;
+    padding: 30px 40px;
+    box-shadow: 0 8px 30px rgba(0, 0, 0, 0.3);
+    margin: 40px auto;
+    max-width: 1200px;
+}
+.gradio-container h1 {
+    color: #333;
+    text-shadow: 1px 1px 2px rgba(0, 0, 0, 0.2);
+}
+.fillable { 
+    width: 95% !important; 
+    max-width: unset !important;
+}
+#examples_container {
+    margin: auto;
+    width: 90%;
+}
+#examples_row {
+    justify-content: center;
+}
+.sidebar {
+    background: rgba(255, 255, 255, 0.98);
+    border-radius: 10px;
+    padding: 20px;
+    box-shadow: 0 4px 15px rgba(0, 0, 0, 0.2);
+}
+button, .btn {
+    background: linear-gradient(90deg, #ff8a00, #e52e71);
+    border: none;
+    color: #fff;
+    padding: 12px 24px;
+    text-transform: uppercase;
+    font-weight: bold;
+    letter-spacing: 1px;
+    border-radius: 5px;
+    cursor: pointer;
+    transition: transform 0.2s ease-in-out;
+}
+button:hover, .btn:hover {
+    transform: scale(1.05);
+}
+"""
+
+title_html = """
+<h1 align="center" style="margin-bottom: 0.2em;"> 🤗 Vidraft-Gemma-3-27B </h1>
+<p align="center" style="font-size:1.1em; color:#555;">
+    Multimodal Chat Interface + Optional Web Search
+</p>
+"""
+
+##############################################################################
+# Build a Blocks layout that includes:
+#   - A left sidebar with "Web Search" controls
+#   - The main ChatInterface in the center or right
+##############################################################################
+with gr.Blocks(css=css, title="Vidraft-Gemma-3-27B") as demo:
+    gr.Markdown(title_html)
+
+    with gr.Row():
+        # Left Sidebar
+        with gr.Column(scale=3, variant="panel"):
+            gr.Markdown("### Menu / Options")
+            with gr.Row():
+                web_search_checkbox = gr.Checkbox(
+                    label="Web Search",
+                    value=False,
+                    info="Check to enable a SERPHouse web search before the chat reply"
+                )
+            web_search_text = gr.Textbox(
+                lines=1,
+                label="Web Search Query",
+                placeholder="Enter search keywords..."
+            )
+            # No immediate function needed; the ChatInterface call will read these states
+
+            gr.Markdown("---")
+            gr.Markdown("#### System Prompt")
+            system_prompt_box = gr.Textbox(
+                lines=3,
+                value=(
+                    "You are a deeply thoughtful AI. Consider problems thoroughly and derive "
+                    "correct solutions through systematic reasoning. Please answer in korean."
+                ),
             )
-        ),
-        gr.Slider(label="Max New Tokens", minimum=100, maximum=8000, step=50, value=2000),
-    ],
-    stop_btn=False,
-    title="Vidraft-Gemma-3-27B",
-    examples=examples,
-    run_examples_on_click=False,
-    cache_examples=False,
-    css_paths="style.css",
-    delete_cache=(1800, 1800),
-)
 
+            max_tokens_slider = gr.Slider(
+                label="Max New Tokens",
+                minimum=100,
+                maximum=8000,
+                step=50,
+                value=2000,
+            )
+
+            gr.Markdown("<br><br>")  # spacing
+
+        # Main ChatInterface to the right
+        with gr.Column(scale=7):
+            chat = gr.ChatInterface(
+                fn=run,
+                type="messages",
+                chatbot=gr.Chatbot(type="messages", scale=1, allow_tags=["image"]),
+                # .webp, .png, .jpg, .jpeg, .gif, .mp4, .csv, .txt, .pdf 모두 허용
+                textbox=gr.MultimodalTextbox(
+                    file_types=[
+                        ".webp", ".png", ".jpg", ".jpeg", ".gif",
+                        ".mp4", ".csv", ".txt", ".pdf"
+                    ],
+                    file_count="multiple",
+                    autofocus=True
+                ),
+                multimodal=True,
+                additional_inputs=[
+                    system_prompt_box,
+                    max_tokens_slider,
+                    web_search_checkbox,
+                    web_search_text,
+                ],
+                stop_btn=False,
+                title="Vidraft-Gemma-3-27B",
+                examples=examples,
+                run_examples_on_click=False,
+                cache_examples=False,
+                css_paths=None,  # We'll rely on the custom Blocks css
+                delete_cache=(1800, 1800),
+            )
+
+    with gr.Row(elem_id="examples_row"):
+        with gr.Column(scale=12, elem_id="examples_container"):
+            gr.Markdown("### Example Inputs (click to load)")
+            # We attach the same examples again at the bottom for a visual layout
+            # But note that ChatInterface already has these examples.
+            # Optionally, we could remove them here to avoid duplication:
+            gr.Examples(
+                examples=examples,
+                inputs=chat,
+                cache_examples=False
+            )
+
+# Finally, launch
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()
+