Spaces:
Build error
Build error
#the below import has been replaced by the later mentioned import, recently by langchain as a per of their improvement strategy :) | |
#from langchain.chat_models import ChatOpenAI | |
#from langchain_openai import ChatOpenAI | |
from langchain_community.llms import HuggingFaceEndpoint | |
from langchain.schema import HumanMessage, SystemMessage | |
from io import StringIO | |
import streamlit as st | |
from dotenv import load_dotenv | |
import time | |
import base64 | |
#This function is typically used in Python to load environment variables from a .env file into the application's environment. | |
load_dotenv() | |
st.title("Let's do code review for your python code") | |
st.header("Please upload your .py file here:") | |
# Function to download text content as a file using Streamlit | |
def text_downloader(raw_text): | |
# Generate a timestamp for the filename to ensure uniqueness | |
timestr = time.strftime("%Y%m%d-%H%M%S") | |
# Encode the raw text in base64 format for file download | |
b64 = base64.b64encode(raw_text.encode()).decode() | |
# Create a new filename with a timestamp | |
new_filename = "code_review_analysis_file_{}_.txt".format(timestr) | |
st.markdown("#### Download File ✅###") | |
# Create an HTML link with the encoded content and filename for download | |
href = f'<a href="data:file/txt;base64,{b64}" download="{new_filename}">Click Here!!</a>' | |
# Display the HTML link using Streamlit markdown | |
st.markdown(href, unsafe_allow_html=True) | |
# Capture the .py file data | |
data = st.file_uploader("Upload python file",type=".py") | |
if data: | |
# Create a StringIO object and initialize it with the decoded content of 'data' | |
stringio = StringIO(data.getvalue().decode('utf-8')) | |
# Read the content of the StringIO object and store it in the variable 'read_data' | |
fetched_data = stringio.read() | |
# Optionally, uncomment the following line to write the read data to the streamlit app | |
st.write(fetched_data) | |
# Initialize a ChatOpenAI instance with the specified model name "gpt-3.5-turbo" and a temperature of 0.9. | |
#chat = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0.9) | |
chat = HuggingFaceEndpoint(temperature=0.9,repo_id="mistralai/Mixtral-8x7B-Instruct-v0.1") #"mistralai/Mistral-7B-Instruct-v0.2" # 'text-davinci-003' model is depreciated now, so we are using the openai's recommended model | |
# Create a SystemMessage instance with the specified content, providing information about the assistant's role. | |
systemMessage = SystemMessage(content="You are a code review assistant. Provide detailed suggestions to improve the given Python code along by mentioning the existing code line by line with proper indent") | |
# Create a HumanMessage instance with content read from some data source. | |
humanMessage = HumanMessage(content=fetched_data) | |
# Call the chat method of the ChatOpenAI instance, passing a list of messages containing the system and human messages. | |
# Recently langchain has recommended to use invoke function for the below please :) | |
finalResponse = chat.invoke([systemMessage, humanMessage]) | |
#Display review comments | |
st.markdown(finalResponse) | |
text_downloader(finalResponse) | |