Spaces:
Sleeping
Sleeping
| import os | |
| import re | |
| import gradio as gr | |
| import qdrant_client | |
| from langchain_community.embeddings import HuggingFaceEmbeddings | |
| from langchain_qdrant import Qdrant | |
| from langchain_openai import ChatOpenAI | |
| class HadithChatApp: | |
| def __init__(self): | |
| self.QDRANT_URL = os.getenv('QDRANT_URL') | |
| self.QDRANT_API_KEY = os.getenv('QDRANT_API_KEY') | |
| self.GOOGLE_API_KEY = os.getenv('GOOGLE_API_KEY') | |
| self.collection_name = "Cluster0" | |
| self.client = qdrant_client.QdrantClient( | |
| url=self.QDRANT_URL, | |
| api_key=self.QDRANT_API_KEY | |
| ) | |
| self.collection_config = qdrant_client.http.models.VectorParams( | |
| size=384, | |
| distance=qdrant_client.http.models.Distance.COSINE | |
| ) | |
| self.embeddings = HuggingFaceEmbeddings( | |
| model_name="intfloat/multilingual-e5-small" | |
| ) | |
| self.vectorStore = Qdrant( | |
| client=self.client, | |
| collection_name=self.collection_name, | |
| embeddings=self.embeddings | |
| ) | |
| self.chat = ChatGoogleGenerativeAI(model="gemini-1.5-flash-latest",google_api_key=self.GOOGLE_API_KEY) | |
| def clean_text(self, text): | |
| text = re.sub(r'<[^>]*>', '', text) | |
| text = re.sub(r'[^\w\s]', '', text) | |
| text = re.sub(r'\s+', ' ', text) | |
| return text.lower().strip() | |
| def get_relevant_docs(self, question, k): | |
| relevant_docs = self.vectorStore.similarity_search_with_score(query=question, k=k) | |
| return relevant_docs | |
| def extract_contexts(self, relevant_docs): | |
| contexts = [] | |
| for doc in relevant_docs: | |
| contexts.append(doc[0].page_content) | |
| return contexts | |
| def create_template(self, question, k): | |
| relevant_docs = self.get_relevant_docs(question, k) | |
| contexts = self.extract_contexts(relevant_docs) | |
| template = f""" | |
| Engage in a conversation with the user, responding to their question: | |
| {question} | |
| within this contexts of Hadiths: | |
| {contexts} | |
| Encourage the model to provide informative and culturally sensitive answers, reflecting Islamic teachings. Maintain a conversational tone and aim for clarity in responses and make sure they are restricted extracted from the provided contexts and i want you to answer me in arabic.""" | |
| return template | |
| def generate_answer(self, question): | |
| cleaned_question = self.clean_text(question) | |
| query = self.create_template(cleaned_question, 10) | |
| response = self.clean_text(self.chat.invoke(query).content) | |
| return response | |
| def greet(self, question): | |
| answer = self.generate_answer(question) | |
| return answer | |
| if __name__ == "__main__": | |
| hadith_chat_app = HadithChatApp() | |
| iface = gr.Interface(fn=hadith_chat_app.greet, inputs="text", outputs="text") | |
| iface.launch(inline=False) |