Update app.py

app.py

@@ -1,91 +1,26 @@
 import gradio as gr
-from gpt4all import GPT4All
-from huggingface_hub import hf_hub_download
-import faiss
-#from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain_huggingface import HuggingFaceEmbeddings
-import numpy as np
-from pypdf import PdfReader
 from gradio_pdf import PDF
-from pdf2image import convert_from_path
-from transformers import pipeline
-from pathlib import Path
-from langchain_chroma import Chroma
-
-title = "Mistral-7B-Instruct-GGUF Run On CPU-Basic Free Hardware"
-
-description = """
-🔎 [Mistral AI's Mistral 7B Instruct v0.1](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) [GGUF format model](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF) , 4-bit quantization balanced quality gguf version, running on CPU. English Only (Also support other languages but the quality's not good). Using [GitHub - llama.cpp](https://github.com/ggerganov/llama.cpp) [GitHub - gpt4all](https://github.com/nomic-ai/gpt4all). 
-🔨 Running on CPU-Basic free hardware. Suggest duplicating this space to run without a queue. 
-Mistral does not support system prompt symbol (such as ```<<SYS>>```) now, input your system prompt in the first message if you need. Learn more: [Guardrailing Mistral 7B](https://docs.mistral.ai/usage/guardrailing). 
-"""
-
-"""
-[Model From TheBloke/Mistral-7B-Instruct-v0.1-GGUF](https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF)
-[Mistral-instruct-v0.1 System prompt](https://docs.mistral.ai/usage/guardrailing)
-"""
-
-model_path = "models"
-model_name = "mistral-7b-instruct-v0.1.Q4_K_M.gguf"
-
-hf_hub_download(repo_id="TheBloke/Mistral-7B-Instruct-v0.1-GGUF", filename=model_name, local_dir=model_path, local_dir_use_symlinks=False)
-
-print("Start the model init process")
-model = model = GPT4All(model_name, model_path, allow_download = False, device="cpu")
-
-
-model.config["promptTemplate"] = "[INST] {0} [/INST]"
-model.config["systemPrompt"] = "Tu es un assitant et tu dois répondre en français"
-model._is_chat_session_activated = False
-
-max_new_tokens = 2048
-
-model_kwargs = {'device': 'cpu'}
-encode_kwargs = {'normalize_embeddings': False}
-embeddings = HuggingFaceEmbeddings(
-    
-    model_kwargs=model_kwargs,
-    encode_kwargs=encode_kwargs
-)
-
-chunk_size = 2048
-
-# creating a pdf reader object
-
-vectordb = Chroma(
-    persist_directory="./resource/chroma/",
-    embedding_function=embeddings
-)
-
-print("Finish the model init process")
+from pypdf import PdfReader
 
-def qa(question: str) -> str:
+def upload_file(files):
+    file_paths = [file.name for file in files]
+    return file_paths
     
+def summarise_file(file):
+    text = PdfReader(file)
     
-    docs = vectordb.max_marginal_relevance_search(question,k=3)
+    return text
 
+with gr.Blocks() as demo:
+    file_output = gr.File()
+    upload_button = gr.UploadButton("Click to Upload a File", file_types=["pdf", "doc"], file_count="multiple")
+    upload_button.upload(upload_file, upload_button, file_output)
     
-
-   # prompt = f"""<s>[INST]
-    #        Les informations contextuelles sont ci-dessous.
-     #       ---------------------
-      #      {docs[0].page_content}
-       #     ---------------------
-        #    [/INST]
-         #   Compte tenu des informations contextuelles et non des connaissances préalables, répondez à la requête.  </s>
-          #  [INST] Requête: {question} [/INST]
-           # Réponse: 
-            #    """
-
-    #outputs = model.generate(prompt=prompt, temp=0.5, top_k = 40, top_p = 1, max_tokens = max_new_tokens)
-    return vectordb._collection.count() #"".join(outputs)
+    file_summary = gr.Textbox()			
+    sum_button = gr.Button("Click to summarise file")
+    
+    sum_button.click(summarise_file, file_output, file_summary)
 
 
-demo = gr.Interface(
-    qa,
-    [gr.Textbox(label="Question")#, PDF(label="Document")
-    ],
-    gr.Textbox()
-)
 if __name__ == "__main__":
-    demo.queue(max_size=3).launch()
+demo.queue(max_size=3).launch()