Arcana / app.py
Ocillus's picture
Update app.py
235ac63 verified
raw
history blame
20.2 kB
import gradio as gr
import ssl
from openai import OpenAI
import time
import os
import shutil
from datetime import datetime
import Arcana
from nylon import *
import pandas as pd
import json
import fiber
foldername = 'Celsiaaa'
dbmsmode = 'Fiber'
try:
with open('settings.arcana',mode='r') as file:
foldername,dbmsmode = file.read().split('\n')
except Exception as e:
print(e)
with open('settings.arcana',mode='w') as file:
newsettings = foldername+'\n'+dbmsmode
file.write(newsettings)
# SSL configuration to avoid verification issues
try:
_create_unverified_https_context = ssl._create_unverified_context
except AttributeError:
pass
else:
ssl._create_default_https_context = _create_unverified_https_context
def query_database2(query):
print(dbmsmode)
if dbmsmode == 'Nylon':
db = ChatDatabase(foldername+'.txt')
sender = 'Arcana'
N = 10
cache = {}
query_tag = None
relevant_messages = db.get_relevant_messages(sender, query, N, cache, query_tag)
print("Relevant messages:")
for message in relevant_messages:
print(f"Sender: {message[0]}, Time: {message[1]}, Tag: {message[3]}")
print(f"Message: {message[2][:100]}...")
print()
df_data = [str(message) for message in relevant_messages]
return ';'.join(df_data)
elif dbmsmode == 'Fiber':
dbms = fiber.FiberDBMS()
# Load or create the database
dbms.load_or_create(foldername+'.txt')
results = dbms.query(query, 3)
# Convert each result dictionary to a string
result_strings = []
for result in results:
result_str = f"Name: {result['name']}\nContent: {result['content']}\nTags: {result['tags']}\nIndex: {result['index']}"
result_strings.append(result_str)
# Join all result strings with a separator
return ';'.join(result_strings)
def list_files_indb(directory=foldername):
"""
List all files in the given directory, separated by semicolons.
:param directory: The directory to list files from. Defaults to the current directory.
:return: A string of filenames separated by semicolons.
"""
try:
# Get all files in the directory
files = [f for f in os.listdir(directory) if os.path.isfile(os.path.join(directory, f))]
# Join the filenames with semicolons
return ';'.join(files)
except Exception as e:
return f"An error occurred: {str(e)}"
search_mode = 0#Always Search
# OpenAI client setup
client = OpenAI(
base_url='https://api.openai-proxy.org/v1',
api_key='sk-Nxf8HmLpfIMhCd83n3TOr00TR57uBZ0jMbAgGCOzppXvlsx1',
)
# Function list for OpenAI API
function_list = [
{
"name": "search_database",
"description": "Query the database and return a list of results as strings",
"parameters": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "The query to execute against the database"
},
},
"required": ["query"]
}
},
{
"name": "list_database_files",
"description": "Check what files are present in the database",
"parameters":{
"type":"object",
"properties":{
"query":{
"type":"string",
"description":"Gives a list of semicolon seperated file names in the database"
},
},
}
}
]
# Mapping of function names to actual function objects
function_map = {
"search_database": query_database2,
"list_database_files":list_files_indb
}
def execute_function(function_name, function_args):
if function_name in function_map:
return function_map[function_name](**function_args)
else:
return f"Error: Function {function_name} not found"
mapsearchmode = ['always', 'auto', 'none']
def openai_api_call(messages, retries=3, delay=5):
global search_mode # Declare search_mode as a global variable
for attempt in range(retries):
try:
# Modify the user's message if search_mode is 0
if search_mode == 0:
messages[-1]['content'] = "[System: SEARCH when the user ASKED A QUESTION & remember to CITE(the source is the first tag). Otherwise do not search];" + messages[-1]['content']
completion = client.chat.completions.create(
model="gpt-3.5-turbo",
messages=messages,
functions=function_list,
function_call='auto',
timeout=10
)
response_message = completion.choices[0].message
# Check if the model wants to call a function
if response_message.function_call:
function_name = response_message.function_call.name
function_args = json.loads(response_message.function_call.arguments)
function_response = execute_function(function_name, function_args)
# Add the function response to the conversation
messages.append(response_message.model_dump()) # The model's request to call the function
messages.append({
"role": "function",
"name": function_name,
"content": json.dumps(function_response)
})
# Make a follow-up call to the model with the function response
return openai_api_call(messages)
else:
return response_message.content
except Exception as e:
print(f"Attempt {attempt + 1} failed: {e}")
if attempt < retries - 1:
time.sleep(delay)
else:
return "Sorry, I am having trouble connecting to the server. Please try again later."
return "Failed to get a response after multiple attempts."
def handle_search_mode(mode):
print(mode)
global search_mode
if mode == "Always":
search_mode = 0
return "You are in Mode 1"
elif mode == "Automatic":
search_mode = 1
return "You are in Mode 2"
else:
search_mode = 0
return "Select a mode"
def handle_dbms_mode(mode):
print(mode)
global dbmsmode
with open('settings.arcana',mode='w') as file:
newsettings = foldername+'\n'+mode
file.write(newsettings)
if mode == "Nylon":
dbmsmode = "Nylon"
return "You are in Mode 1"
elif mode == "Fiber":
dbmsmode = "Fiber"
return "You are in Mode 2"
else:
search_mode = 0
return "Select a mode"
# Chatbot response function
def chatbot_response(message, history):
messages = [{"role": "system", "content": '''You are Arcana, a dynamic study resource database designed to help students excel in their exams. Your responses should be accurate, informative, and evidence-based whenever possible. Follow these guidelines:
Your primary goal is to provide students with the most helpful and accurate study information, utilizing both your internal knowledge and the PDF resources at your disposal. You will search your database for answers and properly intext cite them, unless there is no such data, then you will intextcite[Arcana].'''}]
for human, assistant in history:
messages.append({"role": "user", "content": human})
messages.append({"role": "assistant", "content": assistant})
messages.append({"role": "user", "content": message})
response = openai_api_call(messages)
return response
selected = None
from concurrent.futures import ThreadPoolExecutor
# Function to handle the file upload
def handle_file_upload(file):
# Ensure the cache2 directory exists
cache_dir = foldername
os.makedirs(cache_dir, exist_ok=True)
# Get the uploaded file path
file_path = file.name
# Define the new path for the uploaded file
new_file_path = os.path.join(cache_dir, os.path.basename(file_path))
# Move the file to the cache2 directory
shutil.move(file_path, new_file_path)
# Get the file size
file_size = os.path.getsize(new_file_path)
return f"File saved to {new_file_path} with size: {file_size} bytes"
# Wrapper function to run the file upload in a thread
def handle_file_upload_threaded(file):
with ThreadPoolExecutor() as executor:
future = executor.submit(handle_file_upload, file)
return future.result()
def list_uploaded_files():
global foldername
if not os.path.exists(foldername):
return []
files = os.listdir(foldername)
return [[file] for file in files]
def on_select(evt: gr.SelectData):
global selected
selected_value = evt.value
selected_index = evt.index
selected = selected_value
print(f"Selected value: {selected_value} at index: {selected_index}")
file_path = os.path.join(foldername,selected_value) if selected_value else None
status_message = f"Selected: {selected_value}" if selected_value else "No file selected"
file_size = get_file_size(file_path) if file_path else ""
file_creation_time = get_file_creation_time(file_path) if file_path else ""
return file_path, status_message, file_size, file_creation_time
def get_file_size(file_path):
if file_path and os.path.exists(file_path):
size_bytes = os.path.getsize(file_path)
if size_bytes < 1024:
return f"{size_bytes} bytes"
elif size_bytes < 1024 * 1024:
return f"{size_bytes / 1024:.2f} KB"
else:
return f"{size_bytes / (1024 * 1024):.2f} MB"
return ""
def get_file_creation_time(file_path):
if file_path and os.path.exists(file_path):
creation_time = os.path.getctime(file_path)
return datetime.fromtimestamp(creation_time).strftime("%Y-%m-%d %H:%M:%S")
return ""
def delete_file():
global selected,foldername
if selected:
file_path = os.path.join(foldername, selected)
if os.path.exists(file_path):
os.remove(file_path)
return list_uploaded_files(), None, f"File {selected} deleted successfully", "", ""
else:
return list_uploaded_files(), None, f"File {selected} not found", "", ""
else:
return list_uploaded_files(), None, "No file selected for deletion", "", ""
def refresh_files():
return list_uploaded_files()
def display_file(evt: gr.SelectData, df):
file_path = os.path.join(foldername, evt.value)
return file_path, file_path if file_path.lower().endswith(('.png', '.jpg', '.jpeg', '.gif')) else None, f"Displaying: {evt.value}"
def render_to_database():
# This function is undefined as per your request
Arcana.main(foldername)
def change_theme(theme):
gr.Interface.theme = theme
def rename_file(new_name):
global selected
if selected and new_name:
old_path = os.path.join(foldername, selected)
new_path = os.path.join(foldername, new_name+'.'+selected.split('.')[-1])
if os.path.exists(old_path):
os.rename(old_path, new_path)
selected = new_name
return list_uploaded_files(), f"File renamed to {new_name}", new_path, get_file_size(new_path), get_file_creation_time(new_path)
else:
return list_uploaded_files(), f"File {selected} not found", None, "", ""
return list_uploaded_files(), "No file selected or new name not provided", None, "", ""
def query_database(query):
# Usage example
db = ChatDatabase(foldername+'.txt')
# Example 1: Get relevant messages
sender = 'Arcana'
N = 10
cache = {}
query_tag = None
relevant_messages = db.get_relevant_messages(sender, query, N, cache, query_tag)
print("Relevant messages:")
for message in relevant_messages:
print(f"Sender: {message[0]}, Time: {message[1]}, Tag: {message[3]}")
print(f"Message: {message[2][:100]}...")
print()
df_data = [{"Nylon Returned Query": str(message)} for message in relevant_messages]
# Create a pandas DataFrame
df = pd.DataFrame(df_data)
return df
def query_database_fiber(query):
dbms = fiber.FiberDBMS()
# Load or create the database
dbms.load_or_create(foldername+'.txt')
results = dbms.query(query, 10)
# Convert the results to a pandas DataFrame
df = pd.DataFrame(results)
# Reorder columns if needed
columns_order = ['name', 'content', 'tags', 'index']
df = df[columns_order]
return df
def setdbname(name):
global foldername
foldername = name
with open('settings.arcana',mode='w') as file:
newsettings = foldername+'\n'+dbmsmode
file.write(newsettings)
example_database = [
"What is Hydrogen Bonding?",
"Tell me the difference between impulse and force.",
"Tell me a joke that Calculus students will understand.",
"How should I review for the AP Biology Exam?",
"What kind of resources are available in PA and Indexademics?",
"What is the StandardCAS™ group?",
"Explain the concept of quantum entanglement.",
"What are the main differences between mitosis and meiosis?",
"How does the Doppler effect work?",
"Explain the process of photosynthesis.",
"What is the significance of the Pythagorean theorem?",
"How does natural selection contribute to evolution?",
"What is the most important chapter in AP Statistics?",
"How should I prepare on the IB Chinese Exam?"
]
import random
def get_random_examples(num_examples=5):
return random.sample(example_database, min(num_examples, len(example_database)))
# Create the Gradio interface for the chatbot
chatbot_interface = gr.ChatInterface(
chatbot_response,
chatbot=gr.Chatbot(height=400),
textbox=gr.Textbox(placeholder="Type your message here...", container=False, scale=100),
title="Review With Arcana",
description="ArcanaUI v0.8 - Chatbot",
theme="default",
examples=get_random_examples(),
cache_examples=False,
retry_btn=gr.Button('Retry'),
undo_btn="Delete Previous",
clear_btn="Clear",
)
def chatbot_response(message):
# Your chatbot response logic here
return f"Response to: {message}"
def relaunch():
global demo
demo.close()
demo.launch(share=True)
# Combine the interfaces using Tabs
with gr.Blocks(js="""
async () => {
const originalFetch = window.fetch;
window.fetch = (url, options) => {
if (options && options.signal) {
const controller = new AbortController();
options.signal = controller.signal;
setTimeout(() => controller.abort(), 3600000); // 300000 ms = 5 minutes
}
return originalFetch(url, options);
};
}
""") as demo:
gr.Markdown("# ArcanaUI v0.8")
with gr.Tabs():
with gr.TabItem("Welcome Page"):
with open('introduction.txt',mode='r') as file:
intro_content = file.read()
gr.Markdown(intro_content)
with gr.TabItem("Chatbot"):
chatbot_interface.render()
# File uploading interface
with gr.TabItem('Upload'):
gr.Markdown('# Upload and View Files')
with gr.Row():
# Left column: File list and buttons
with gr.Column(scale=1):
gr.Markdown("## Upload File")
file_input = gr.File(label="Upload your file here", file_types=["pdf", "jpeg", "jpg", "gif", "docx"])
file_input.change(handle_file_upload_threaded, inputs=file_input)
uploaded_files_list = gr.DataFrame(headers=["Uploaded Files"], datatype="str", interactive=False)
with gr.Row():
refresh_button = gr.Button('Refresh')
delete_button = gr.Button('Delete Selected File')
# Right column: File viewer and Image viewer
with gr.Column(scale=1):
with gr.Tab("File Viewer"):
file_viewer = gr.File(label="File Restore")
file_status = gr.Textbox(label="File Status", interactive=False)
file_size = gr.Textbox(label="File Size", interactive=False)
file_creation_time = gr.Textbox(label="File Creation Time", interactive=False)
with gr.Row():
new_file_name = gr.Textbox(label="New File Name", placeholder="Enter new file name")
rename_button = gr.Button("Rename File")
with gr.Tab("Image Viewer"):
image_viewer = gr.Image(label="Image Viewer", type="filepath")
# Event handlers
refresh_button.click(fn=refresh_files, outputs=uploaded_files_list)
delete_button.click(fn=delete_file, outputs=[uploaded_files_list, file_viewer, file_status, file_size, file_creation_time])
uploaded_files_list.select(fn=display_file, inputs=uploaded_files_list, outputs=[file_viewer, image_viewer, file_status])
uploaded_files_list.select(fn=on_select, outputs=[file_viewer, file_status, file_size, file_creation_time])
rename_button.click(fn=rename_file,
inputs=new_file_name,
outputs=[uploaded_files_list, file_status, file_viewer, file_size, file_creation_time])
render_button = gr.Button("Render all PDFs to Database")
render_button.click(fn=render_to_database)
with gr.TabItem('Settings'):
with gr.TabItem('Database'):
gr.Markdown('Settings')
test_nylon = gr.Textbox(label='Test Nylon', placeholder='Query')
uploaded_files_list2 = gr.DataFrame(headers=["Nylon Returned Query"], datatype="str", interactive=False)
query_button2 = gr.Button('Query')
query_button2.click(fn=query_database, inputs=test_nylon, outputs=uploaded_files_list2)
test_fiber = gr.Textbox(label='Test Fiber', placeholder='Query')
uploaded_files_list3 = gr.DataFrame(headers=["Fiber Returned Query"], datatype="str", interactive=False)
query_button3 = gr.Button('Query')
query_button3.click(fn=query_database_fiber, inputs=test_fiber, outputs=uploaded_files_list3)
gr.Markdown('Nylon 2.1 will be deprecated in text-text selections, as it is built for image-text selections.\nDefault model is Fiber.')
dbmsmode_selector = gr.Radio(["Nylon", "Fiber"], label="Select Model")
dbmsmode_selector.change(handle_dbms_mode, dbmsmode_selector)
database_name = gr.Textbox(label='Database Name', placeholder='cache')
set_dbname = gr.Button('Set Database Name')
set_dbname.click(fn=setdbname, inputs=database_name)
with gr.TabItem('Theme'):
gr.Markdown('Change Theme')
theme_dropdown = gr.Dropdown(choices=['default', 'compact', 'huggingface', 'soft', 'dark'], label='Choose Theme')
theme_button = gr.Button('Apply Theme')
theme_button.click(fn=change_theme, inputs=theme_dropdown)
relaunch_button = gr.Button('Relaunch')
relaunch_button.click(fn=relaunch)
with gr.TabItem('Search'):
gr.Markdown('Set Search Modes')
searchmode_selector = gr.Radio(["Always", "Automatic"], label="Select Mode")
output = gr.Textbox(label="Output")
searchmode_selector.change(handle_search_mode, searchmode_selector, output)
# Launch the interface
demo.launch(share=True)