Create app.py

app.py

@@ -0,0 +1,68 @@
+from langchain.document_loaders import HuggingFaceDatasetLoader
+from langchain_community.document_loaders.csv_loader import CSVLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter, CharacterTextSplitter
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.vectorstores import FAISS
+from transformers import AutoTokenizer, AutoModelForQuestionAnswering
+from transformers import AutoTokenizer, pipeline
+from langchain import HuggingFacePipeline
+from langchain.chains import RetrievalQA
+from langchain.schema import AIMessage, HumanMessage, SystemMessage
+from langchain.schema.runnable import RunnablePassthrough, RunnableLambda
+from langchain.prompts import ChatPromptTemplate
+from langchain_community.vectorstores import FAISS
+from langchain_core.documents import Document
+import gradio as gr
+import os
+import difflib
+
+
+modelPath = "MSEAJYTHTH/NFPC"
+
+model_kwargs = {'device':'cpu'}
+
+encode_kwargs = {'normalize_embeddings': False}
+
+embeddings = HuggingFaceEmbeddings(
+    model_name=modelPath,     # Provide the pre-trained model's path
+    model_kwargs=model_kwargs, # Pass the model configuration options
+    encode_kwargs=encode_kwargs # Pass the encoding options
+)
+
+
+def find_best_page_content(question, keywords, db):
+
+    db = FAISS.load_local("MSEAJYTHTH/NFPC/faiss_index", embeddings, allow_dangerous_deserialization=True)
+    results = db.similarity_search(question, k=50, fetch_k=50)
+    keywords = ', '.join(keywords)  #
+
+    max_similarity = 0  # 최대 유사도 초기화
+    best_page_content = ""    # 가장 유사한 결과 초기화
+    page_content = ""
+    metadata = ""
+    similarity_score = ""
+
+    for i in range(50) :
+       page_content = results[i].page_content
+       metadata = results[i].metadata['source']
+       similarity_score = difflib.SequenceMatcher(None, keywords, metadata).ratio()
+
+       if similarity_score > max_similarity:
+             max_similarity = similarity_score
+             best_page_content = page_content
+
+    return best_page_content
+
+
+gr.Interface(
+    fn=find_best_page_content,
+    inputs=[
+        gr.Textbox(label="질문을 입력해주세요", placeholder="물분무소화설비 화재안전성능기준에서 헤드 관련 기준을 찾아줘"),
+        gr.Textbox(label="Keywords를 입력해주세요", placeholder="물분무, 헤드"),
+    ],
+    outputs=gr.Textbox(label="검색 결과", type="text"),  # 수정된 부분
+    title="화재안전성능기준 Question & Answering",
+    description="화재안전성능기준를 찾아주는 Chatbot",
+    theme="soft",
+    examples=[["물분무소화설비 헤드에 대한 것을 찾아줘", "물분무, 헤드"]],
+).launch()