app.py

import sklearn
import sqlite3
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
import openai
import os

openai.api_key = os.environ["Secret"] 


def find_closest_neighbors(vector1, dictionary_of_vectors):
    """
    Takes a vector and a dictionary of vectors and returns the three closest neighbors
    """

    # Convert the input string to a vector
    vector = openai.Embedding.create(
        input=vector1,
        engine="text-embedding-ada-002"
    )['data'][0]['embedding']
    
    vector = np.array(vector)

    # Finds cosine similarities between the vector and values in the dictionary and Creates a dictionary of cosine similarities with its text key
    cosine_similarities = {}
    for key, value in dictionary_of_vectors.items():
        cosine_similarities[key] = cosine_similarity(vector.reshape(1, -1), value.reshape(1, -1))[0][0]

    # Sorts the dictionary by value and returns the three highest values
    sorted_cosine_similarities = sorted(cosine_similarities.items(), key=lambda x: x[1], reverse=True)
    match_list = sorted_cosine_similarities[0:4]
    web = str(sorted_cosine_similarities[0][0])
    return match_list

# Connect to the database
conn = sqlite3.connect('QRIdatabase.db')

# Create a cursor
cursor = conn.cursor()

# Select the text and embedding from the chunks table
cursor.execute('''SELECT text, embedding FROM chunks''')

# Fetch the rows
rows = cursor.fetchall()

# Create a dictionary to store the text and embedding for each row
dictionary_of_vectors = {}

# Iterate through the rows and add them to the dictionary
for row in rows:
    text = row[0]
    embedding_str = row[1]
    # Convert the embedding string to a NumPy array
    embedding = np.fromstring(embedding_str, sep=' ')
    dictionary_of_vectors[text] = embedding

# Close the connection
conn.close()

def context_gpt_response(question):
    """
    Takes a question and returns an answer
    """

    # Find the closest neighbors
    match_list = find_closest_neighbors(question, dictionary_of_vectors)

    # Create a string of the text from the closest neighbors
    context = ''
    for match in match_list:
        context += str(match[0])  
    prep = f"This is an OpenAI model tuned to answer questions specific to the Qualia Research institute, a research institute that focuses on consciousness. Here is some question-specific context, and then the Question to answer, related to consciousness, the human experience, and phenomenology: {context}. Here is a question specific to QRI and consciousness in general Q:  {question}  A: "
    # Generate an answer
    response = openai.Completion.create(
        engine="text-davinci-003",
        prompt=prep,
        temperature=0.7,
        max_tokens=200,
    )


    # Return the answer
    return response['choices'][0]['text']

import gradio as gr

iface = gr.Interface(fn=context_gpt_response, inputs="text", outputs="text",title="Qualia Research Institute GPTbot", description="Ask any question and get QRI specific answers!", examples=[["What is QRI?"], ["What is the Symmetry Theory of Valence?"], ["Explain Logarithmic scales of pain and pleasure"]])
iface.launch()