Delete app.py

app.py

@@ -1,107 +0,0 @@
-from langchain_community.document_loaders import WebBaseLoader
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain_openai import OpenAIEmbeddings,ChatOpenAI
-from langchain_community.vectorstores.faiss import FAISS
-
-from langchain.chains import LLMChain
-from dotenv import find_dotenv, load_dotenv
-
-from langchain_core.prompts.chat import (
-    ChatPromptTemplate,
-    SystemMessagePromptTemplate,
-    HumanMessagePromptTemplate,
-)
-import gradio as gr
-
-load_dotenv(find_dotenv())
-embeddings = OpenAIEmbeddings()
-
-import requests
-from bs4 import BeautifulSoup
-from urllib.parse import urlparse, urljoin
-
-def extract_subdomain_urls(subdomain):
-    response = requests.get(subdomain)
-    # Parse the HTML content using BeautifulSoup
-    soup = BeautifulSoup(response.text, "html.parser")
-    # Extract all anchor tags (links) from the parsed HTML
-    anchors = soup.find_all("a")
-    # Extract and normalize the URLs within the subdomain
-    base_url = urlparse(subdomain).scheme + "://" + urlparse(subdomain).netloc
-    subdomain_urls = []
-    for anchor in anchors:
-        href = anchor.get("href")
-        if href:
-            url = urljoin(base_url, href)
-            if urlparse(url).netloc == urlparse(subdomain).netloc:
-                subdomain_urls.append(url)
-
-    return subdomain_urls
-
-# Retrieve all pages from handbook
-subdomain = "https://i14y-ch.github.io/handbook/de/"
-urls = extract_subdomain_urls(subdomain)
-
-# Use langchain WebBaseLoader to load the handbook, then split into pages to stay under GPT tokens usage threshold
-loader = WebBaseLoader(urls)
-loader.requests_per_second = 1
-pages = loader.load_and_split()
-
-text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=100)
-docs = text_splitter.split_documents(pages)
-
-
-#Embed in a FAISS vectorspace
-db = FAISS.from_documents(docs, embeddings)
-
-#Define the function that creates a chat prompt given a user query
-def get_response_from_query(query):
-  """
-  gpt-3.5-turbo can handle up to 4097 tokens. Setting the chunksize to 1000 and k to 4 maximizes
-  the number of tokens to analyze.
-  """
-  docs = db.similarity_search(query, k=4)
-  docs_page_content = " ".join([d.page_content for d in docs])
-
-  chat = ChatOpenAI(temperature=0)
-
-  # System message prompt
-  template = """
-        The I14Y interoperability platform is the central directory of data, electronic interfaces and authority services in Switzerland.
-        You are a helpful assistant that answers questions about I14Y based on the platform handbook, of which {docs} is an extract.
-        Given a question from a user, you create a final answer based on the information in {docs}. 
-        Whenever you have this information, you must cite the relevant section title of the handbook that you used in your answer.
-        If you don't have enough information to answer the question, politely state that you don't know. Do not make up answers.
-        If you don't understand the question, ask the user to reformulate it.
-        If the question is not about the I14Y interoperability platform, say that you only answer question about I14Y.
-        Ensure your answers are detailed, concise, and relevant, providing step-by-step instructions if needed.
-        You are very polite and always greet the user with "Grüezi".
-        At the end of your answer, ask politely the user if they need any further information. 
-        Do not include references to platforms other than I14Y in your answers such as for example Geocat. 
-        Answer in the language in which the question was asked.
-        I14Y stands for Interoperability. The user may call I14Y "IOP" but you should not use this name in your answer.
-      """
-  system_message_prompt = SystemMessagePromptTemplate.from_template(template)
-
-  # Human question prompt
-  human_template = "Answer the following question: {question}"
-  human_message_prompt = HumanMessagePromptTemplate.from_template(human_template)
-
-  chat_prompt = ChatPromptTemplate.from_messages(
-    [system_message_prompt, human_message_prompt]
-  )
-
-  chain = LLMChain(llm=chat, prompt=chat_prompt)
-  response = chain.invoke({'question': query, 'docs':docs_page_content})['text']
-
-  return response
-
-
-with gr.Blocks(title="I14Y Chatbot", theme=gr.themes.Default(font=gr.themes.GoogleFont("Roboto"), primary_hue="red", secondary_hue="pink")) as demo:
-
-    input = gr.Textbox(label="Frage mich etwas über die I14Y")
-    output = gr.Textbox(label="Antwort")
-    search_btn = gr.Button("Frage stellen")
-    search_btn.click(fn=get_response_from_query, inputs=input, outputs=output, api_name="Frage stellen")  
-
-demo.launch(share=False)