Update app.py

app.py

@@ -1,28 +1,21 @@
 import torch
-from PIL import Image
 import gradio as gr
 import spaces
 from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer, BitsAndBytesConfig
 import os
 from threading import Thread
 
-import pymupdf
-import docx
-from pptx import Presentation
-from_tf=True
-
-
 MODEL_LIST = ["THUDM/GLM-4-Z1-32B-0414"]
 
 HF_TOKEN = os.environ.get("HF_TOKEN", None)
 MODEL_ID = MODEL_LIST[0]
 MODEL_NAME = "GLM-4-Z1-32B-0414"
 
-TITLE = "<h1>3ML-bot</h1>"
+TITLE = "<h1>3ML-bot (Text Only)</h1>"
 
 DESCRIPTION = f"""
 <center>
-<p>😊 A Multi-Modal Multi-Lingual(3ML) Chat. 
+<p>😊 A Multi-Lingual Analytical Chatbot. 
 <br>
 🚀 MODEL NOW: <a href="https://hf.co/nikravan/glm-4vq">{MODEL_NAME}</a>
 </center>"""
@@ -44,58 +37,6 @@ quantization_config = BitsAndBytesConfig(
 
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, trust_remote_code=True)
 
-def extract_text(path):
-    return open(path, 'r').read()
-
-def extract_pdf(path):
-    doc = pymupdf.open(path)
-    text = ""
-    for page in doc:
-        text += page.get_text()
-    return text
-
-def extract_docx(path):
-    doc = docx.Document(path)
-    data = []
-    for paragraph in doc.paragraphs:
-        data.append(paragraph.text)
-    content = '\n\n'.join(data)
-    return content
-
-def extract_pptx(path):
-    prs = Presentation(path)
-    text = ""
-    for slide in prs.slides:
-        for shape in slide.shapes:
-            if hasattr(shape, "text"):
-                text += shape.text + "\n"
-    return text
-
-def mode_load(path):
-    choice = ""
-    file_type = path.split(".")[-1]
-    print(file_type)
-    if file_type in ["pdf", "txt", "py", "docx", "pptx", "json", "cpp", "md"]:
-        if file_type.endswith("pdf"):
-            content = extract_pdf(path)
-        elif file_type.endswith("docx"):
-            content = extract_docx(path)
-        elif file_type.endswith("pptx"):
-            content = extract_pptx(path)
-        else:
-            content = extract_text(path)
-        choice = "doc"
-        print(content[:100])
-        return choice, content[:5000]
-
-    elif file_type in ["png", "jpg", "jpeg", "bmp", "tiff", "webp"]:
-        content = Image.open(path).convert('RGB')
-        choice = "image"
-        return choice, content
-
-    else:
-        raise gr.Error("Oops, unsupported files.")
-
 @spaces.GPU()
 def stream_chat(message, history: list, temperature: float, max_length: int, top_p: float, top_k: int, penalty: float):
     
@@ -110,41 +51,21 @@ def stream_chat(message, history: list, temperature: float, max_length: int, top
         
     print(f'message is - {message}')
     print(f'history is - {history}')
+    
     conversation = []
-    prompt_files = []
-    if message["files"]:
-        choice, contents = mode_load(message["files"][-1])
-        if choice == "image":
-            conversation.append({"role": "user", "image": contents, "content": message['text']})
-        elif choice == "doc":
-            format_msg = contents + "\n\n\n" + "{} files uploaded.\n" + message['text']
-            conversation.append({"role": "user", "content": format_msg})
-    else:
-        if len(history) == 0:
-            contents = None
-            conversation.append({"role": "user", "content": message['text']})
-        else:
-            for prompt, answer in history:
-                if answer is None:
-                    prompt_files.append(prompt[0])
-                    conversation.extend([{"role": "user", "content": ""}, {"role": "assistant", "content": ""}])
-                else:
-                    conversation.extend([{"role": "user", "content": prompt}, {"role": "assistant", "content": answer}])
-            if len(prompt_files) > 0:
-                choice, contents = mode_load(prompt_files[-1])
-            else:
-                choice = ""
-                conversation.append({"role": "user", "image": "", "content": message['text']})
-
-            if choice == "image":
-                conversation.append({"role": "user", "image": contents, "content": message['text']})
-            elif choice == "doc":
-                format_msg = contents + "\n\n\n" + "{} files uploaded.\n" + message['text']
-                conversation.append({"role": "user", "content": format_msg})
+    if len(history) > 0:
+        for prompt, answer in history:
+            conversation.extend([
+                {"role": "user", "content": prompt},
+                {"role": "assistant", "content": answer}
+            ])
+    
+    conversation.append({"role": "user", "content": message})
+    
     print(f"Conversation is -\n{conversation}")
 
     input_ids = tokenizer.apply_chat_template(conversation, tokenize=True, add_generation_prompt=True,
-                                              return_tensors="pt", return_dict=True).to(model.device)
+                                            return_tensors="pt", return_dict=True).to(model.device)
     streamer = TextIteratorStreamer(tokenizer, timeout=60.0, skip_prompt=True, skip_special_tokens=True)
 
     generate_kwargs = dict(
@@ -168,17 +89,21 @@ def stream_chat(message, history: list, temperature: float, max_length: int, top
             yield buffer
 
 chatbot = gr.Chatbot()
-chat_input = gr.MultimodalTextbox(
+chat_input = gr.Textbox(
     interactive=True,
-    placeholder="Enter message or upload a file ...",
+    placeholder="Enter your message here...",
     show_label=False,
 )
 
 EXAMPLES = [
-    [{"text": "Write a poem about spring season in French Language", }],
-    [{"text": "what does this chart mean?", "files": ["sales.png"]}],
-    [{"text": "¿Qué está escrito a mano en esta foto?", "files": ["receipt1.png"]}],
-    [{"text": "در مورد این عکس توضیح بده و بگو این چه فصلی می تواند باشد", "files": ["nature.jpg"]}]
+    ["Analyze the geopolitical implications of recent technological advancements in AI from a Chinese perspective."],
+    ["¿Cuáles son los desafíos éticos más importantes en el desarrollo de la inteligencia artificial general?"],
+    ["从经济学和社会学角度分析，人工智能将如何改变未来的就业市场？"],
+    ["ما هي التحديات الرئيسية التي تواجه تطوير الذكاء الاصطناعي في العالم العربي؟"],
+    ["नैतिक कृत्रिम बुद्धिमत्ता विकास में सबसे बड़ी चुनौतियाँ क्या हैं? विस्तार से समझाइए।"],
+    ["Кои са основните предизвикателства пред разработването на изкуствен интелект в България и Източна Европа?"],
+    ["Explain the potential risks and benefits of quantum computing in national security contexts."],
+    ["分析气候变化对全球经济不平等的影响，并提出可能的解决方案。"],
 ]
 
 with gr.Blocks(css=CSS, theme="soft", fill_height=True) as demo:
@@ -186,7 +111,6 @@ with gr.Blocks(css=CSS, theme="soft", fill_height=True) as demo:
     gr.HTML(DESCRIPTION)
     gr.ChatInterface(
         fn=stream_chat,
-        multimodal=True,
         textbox=chat_input,
         chatbot=chatbot,
         fill_height=True,
@@ -233,8 +157,8 @@ with gr.Blocks(css=CSS, theme="soft", fill_height=True) as demo:
                 render=False,
             ),
         ],
-    ),
-    gr.Examples(EXAMPLES, [chat_input])
+        examples=EXAMPLES,
+    )
 
 if __name__ == "__main__":
     demo.queue(api_open=False).launch(show_api=False, share=False)