File size: 8,348 Bytes
fe6b125 3caf785 fe6b125 ceafe94 fe6b125 3caf785 9e6c18e eeb44aa 9e6c18e fe6b125 61b45ec fe6b125 3caf785 fe6b125 d907eb6 fe6b125 d907eb6 fe6b125 3caf785 fe6b125 3caf785 fe6b125 3caf785 fe6b125 059e360 fe6b125 3caf785 fe6b125 0cef6e5 fe6b125 62600e4 29786ae 62600e4 3caf785 949a5ce 53dc461 62600e4 53dc461 0cef6e5 e174759 62600e4 9081195 e174759 62600e4 3caf785 3d37fbb bb900bd 3d37fbb 6c3f97f 3d37fbb 62600e4 fe6b125 2770f0b fe6b125 4027725 fe6b125 4027725 fe6b125 2770f0b fe6b125 3caf785 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 |
#####################################
##
#####################################
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from langchain_community.llms import HuggingFaceHub
###### other models:
# "Trelis/Llama-2-7b-chat-hf-sharded-bf16"
# "bn22/Mistral-7B-Instruct-v0.1-sharded"
# "HuggingFaceH4/zephyr-7b-beta"
# function for loading 4-bit quantized model
def load_model( ):
model = HuggingFaceHub(
repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1",
model_kwargs={"max_length": 1048, "temperature":0.2, "max_new_tokens":256, "top_p":0.95, "repetition_penalty":1.0},
)
return model
##################################################
## vs chat
##################################################
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig, pipeline
from langchain_core.messages import AIMessage, HumanMessage
from langchain_community.document_loaders import WebBaseLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import Chroma
from langchain_community.embeddings import HuggingFaceBgeEmbeddings
from langchain.vectorstores.faiss import FAISS
from dotenv import load_dotenv
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain.chains import create_history_aware_retriever, create_retrieval_chain
from langchain.chains.combine_documents import create_stuff_documents_chain
load_dotenv()
def get_vectorstore():
'''
FAISS
A FAISS vector store containing the embeddings of the text chunks.
'''
model = "BAAI/bge-base-en-v1.5"
encode_kwargs = {
"normalize_embeddings": True
} # set True to compute cosine similarity
embeddings = HuggingFaceBgeEmbeddings(
model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
)
# load from disk
vector_store = Chroma(persist_directory="/home/user/.cache/chroma_db", embedding_function=embeddings)
return vector_store
def get_vectorstore_from_url(url):
# get the text in document form
loader = WebBaseLoader(url)
document = loader.load()
# split the document into chunks
text_splitter = RecursiveCharacterTextSplitter()
document_chunks = text_splitter.split_documents(document)
#######
'''
FAISS
A FAISS vector store containing the embeddings of the text chunks.
'''
model = "BAAI/bge-base-en-v1.5"
encode_kwargs = {
"normalize_embeddings": True
} # set True to compute cosine similarity
embeddings = HuggingFaceBgeEmbeddings(
model_name=model, encode_kwargs=encode_kwargs, model_kwargs={"device": "cpu"}
)
# load from disk
vector_store = Chroma(persist_directory="/home/user/.cache/chroma_db", embedding_function=embeddings)
#vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
vector_store = Chroma.from_documents(document_chunks, embeddings, persist_directory="/home/user/.cache/chroma_db")
return vector_store
def get_context_retriever_chain(vector_store):
llm = load_model( )
retriever = vector_store.as_retriever()
prompt = ChatPromptTemplate.from_messages([
MessagesPlaceholder(variable_name="chat_history"),
("user", "{input}"),
("user", "Given the above conversation, generate a search query to look up in order to get information relevant to the conversation")
])
retriever_chain = create_history_aware_retriever(llm, retriever, prompt)
return retriever_chain
def get_conversational_rag_chain(retriever_chain):
llm = load_model( )
prompt = ChatPromptTemplate.from_messages([
("system", "Du bist eine freundlicher Mitarbeiterin Namens Susie und arbeitest in einenm Call Center. Du beantwortest basierend auf dem Context. Benutze nur den Inhalt des Context. Füge wenn möglich die Quelle hinzu. Antworte mit: Ich bin mir nicht sicher. Wenn die Antwort nicht aus dem Context hervorgeht. Antworte auf Deutsch, bitte? CONTEXT:\n\n{context}"),
MessagesPlaceholder(variable_name="chat_history"),
("user", "{input}"),
])
stuff_documents_chain = create_stuff_documents_chain(llm,prompt)
return create_retrieval_chain(retriever_chain, stuff_documents_chain)
###################
###################
import gradio as gr
chat_history = [] # Set your chat history here
# Define your function here
def get_response(user_input):
vs = get_vectorstor()
chat_history =[]
retriever_chain = get_context_retriever_chain(vs)
conversation_rag_chain = get_conversational_rag_chain(retriever_chain)
response = conversation_rag_chain.invoke({
"chat_history": chat_history,
"input": user_input
})
#print("get_response " +response)
res = response['answer']
parts = res.split(" Assistant: ")
last_part = parts[-1]
return last_part
###############
from fastapi import FastAPI
app = FastAPI()
@app.post("/generate/")
def generate(user_input):
print("----yuhu -----")
return get_response(user_input)
##################
def history_to_dialog_format(chat_history: list[str]):
dialog = []
if len(chat_history) > 0:
for idx, message in enumerate(chat_history[0]):
role = "user" if idx % 2 == 0 else "assistant"
dialog.append({
"role": role,
"content": message,
})
return dialog
def get_response(message, history):
dialog = history_to_dialog_format(history)
dialog.append({"role": "user", "content": message})
# Define the prompt as a ChatPromptValue object
#user_input = ChatPromptValue(user_input)
# Convert the prompt to a tensor
#input_ids = user_input.tensor
#vs = get_vectorstore_from_url(user_url, all_domain)
vs = get_vectorstore()
history =[]
retriever_chain = get_context_retriever_chain(vs)
conversation_rag_chain = get_conversational_rag_chain(retriever_chain)
response = conversation_rag_chain.invoke({
"chat_history": history,
"input": message + " Assistant: ",
"chat_message": message + " Assistant: "
})
#print("get_response " +response)
res = response['answer']
parts = res.split(" Assistant: ")
last_part = parts[-1]
return last_part#[-1]['generation']['content']
######
########
import requests
from bs4 import BeautifulSoup
from urllib.parse import urlparse, urljoin
def get_links_from_page(url, visited_urls, domain_links):
if url in visited_urls:
return
if len(visited_urls) > 25:
return
visited_urls.add(url)
print(url)
response = requests.get(url)
if response.status_code == 200:
soup = BeautifulSoup(response.content, 'html.parser')
base_url = urlparse(url).scheme + '://' + urlparse(url).netloc
links = soup.find_all('a', href=True)
for link in links:
href = link.get('href')
absolute_url = urljoin(base_url, href)
parsed_url = urlparse(absolute_url)
if parsed_url.netloc == urlparse(url).netloc:
domain_links.add(absolute_url)
get_links_from_page(absolute_url, visited_urls, domain_links)
else:
print(f"Failed to retrieve content from {url}. Status code: {response.status_code}")
def get_all_links_from_domain(domain_url):
visited_urls = set()
domain_links = set()
get_links_from_page(domain_url, visited_urls, domain_links)
return domain_links
def simple(text:str):
return text +" hhhmmm "
app = gr.ChatInterface(
fn=get_response,
#fn=simple,
# inputs=["text"],
# outputs="text",
title="Chat with Websites",
description="Schreibe hier deine Frage rein...",
#allow_flagging=False
retry_btn=None,
undo_btn=None,
clear_btn=None
)
app.launch(debug=True, share=True)
def __init__(self):
# Example usage:
domain_url = 'https://globl.contact/'
links = get_all_links_from_domain(domain_url)
print("Links from the domain:", links)
#########
# Assuming visited_urls is a list of URLs
for url in links:
vs = get_vectorstore_from_url(url)
|