File size: 6,429 Bytes
42e3a78 2f2c2eb 42e3a78 2f2c2eb 10f03bd 2f2c2eb 42e3a78 3e07b40 42e3a78 3e07b40 42e3a78 2f2c2eb 42e3a78 2f2c2eb 42e3a78 2f2c2eb 42e3a78 2f2c2eb 42e3a78 2f2c2eb 42e3a78 |
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 |
import time
import streamlit as st
from streamlit_chat import message
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
import textwrap
from chat import generate_response, generate_tag
@st.cache_data()
def create_database():
import json
from langchain.docstore.document import Document
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.vectorstores import FAISS
# Maybe it is better to save the embeddings than the text dataset
if "db" not in st.session_state:
json_file_path = "./new_dataset.json"
string_chunks = []
with open(json_file_path, "r") as json_file:
for line in json_file:
if line != '\n':
json_string = json.loads(line)
string_chunks.append(json_string)
documents_ = []
for line in string_chunks:
loader = Document(page_content=line)
documents_.append(loader)
embeddings = HuggingFaceEmbeddings()
db = FAISS.from_documents(documents_, embeddings)
print(type(db))
return db
db = create_database()
if "tokenizer" not in st.session_state:
st.session_state["tokenizer"] = AutoTokenizer.from_pretrained(
"MBZUAI/LaMini-Flan-T5-783M"
)
st.session_state["model"] = AutoModelForSeq2SeqLM.from_pretrained(
"MBZUAI/LaMini-Flan-T5-783M"
)
st.title("BGPT : Bibek's Personal Chatbot")
# Storing the chat
if "generated" not in st.session_state:
st.session_state["generated"] = []
if "past" not in st.session_state:
st.session_state["past"] = []
# We will get the user's input by calling the get_text function
def get_text():
input_text = st.text_input("Enter your inquiries here: ", "Hi!!")
return input_text
user_input = get_text()
def wrap_text_preserve_newlines(text, width=110):
# Split the input text into lines based on newline characters
lines = text.split('\n')
# Wrap each line individually
wrapped_lines = [textwrap.fill(line, width=width) for line in lines]
# Join the wrapped lines back together using newline characters
wrapped_text = '\n'.join(wrapped_lines).replace("page_content=", "").replace("metadata={}", "")
return wrapped_text
if user_input:
tag = generate_tag(user_input)
start = time.time()
# print(tag)
if tag in ["greeting"]:
output = "Hello 👋! Thanks for visiting!\n I am BGPT! I am here to assist you in obtaining information about Bibek. Feel free to ask me any questions about Bibek. These are some sample questions:\n (I) Tell me about Bibek.\n (II) What skills does Bibek have?\n (III) What work experience does Bibek have?\n (IV) What is Bibek's educational background?\n (V) What awards has Bibek won?\n (VI) What projects have Bibek completed? &\n (VII) How can I contact Bibek?"
else:
tokenizer = st.session_state["tokenizer"]
model = st.session_state["model"]
docs = db.similarity_search(user_input)
output = wrap_text_preserve_newlines(str(docs[0]))
if tag in ["welcome", "thanks", "exit"]:
input = user_input
elif tag in ["BibekBOT-introduction"]:
input = "I am BGPT, a large language model. I am here to assist you in obtaining information about Bibek. Feel free to ask me any questions about Bibek and I will make every effort to respond to all inquiries. These are some sample questions:\n (I) Tell me about Bibek.\n (II) What skills does Bibek have?\n (III) What work experience does Bibek have?\n (IV) What is Bibek's educational background?\n (V) What awards has Bibek won?\n (VI) What projects have Bibek completed? &\n (VII) How can I contact Bibek?. \n Can you paraphrase the above without changing the tone and contents."
elif tag in ["decline"]:
input = "Okay, if there's anything else I can assist with, please don't hesitate to ask. \n Can you paraphrase the above without changing much content and tone."
else:
# output = generate_response(user_input)
task_description_prompt = "I want you to act like my personal assistant chatbot named 'BGPT'. You are provided with some content and you will get one question. Try to answer the question in details based on the provided content. You may paraphrase the contents to reach your answer too. The below is the content: \n"
prompt_template = "\nBased on the above content, try to answer the following question.\n\n"
end_prompt = "\nPlease make meaningful sentence and try to be descriptive as possible responding with many sentences and ending with proper punctuations. If you think the content doesn't contain good answer to the question, give some polite respones telling them that you do not have specific response to the query and apologize and refer them to contact Bibek directly.\n" # NoQA"
short_response_template = "\nIf your response is very short like 1 or 2 sentence, add a followup sentence like 'Let me know if there's anything else I can help you with. or If there's anything else I can assist with, please don't hesitate to ask. I mean something similar in polite way." # NoQA
input = task_description_prompt + output + prompt_template + user_input + end_prompt
input_ids = tokenizer(
input,
return_tensors="pt",
).input_ids
outputs = model.generate(input_ids, max_length=512, do_sample=True)
output = tokenizer.decode(outputs[0]).strip("<pad></s>").strip()
end = time.time()
# print(input)
print("Time for model inference: ", end - start)
# Checks for memory overflow
if len(st.session_state.past) == 15:
st.session_state.past.pop(0)
st.session_state.generated.pop(0)
# store the output
st.session_state.past.append(user_input)
st.session_state.generated.append(output)
if st.session_state["generated"]:
# print(st.session_state)
for i in range(len(st.session_state["generated"]) - 1, -1, -1):
message(
st.session_state["generated"][i],
avatar_style="bottts",
seed=39,
key=str(i), # NoQA
)
message(
st.session_state["past"][i],
is_user=True,
avatar_style="identicon",
seed=4,
key=str(i) + "_user",
) # NoQA
|