app.py

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


# Your OpenAI API Key
openai.api_key = os.environ["Secret"] 

# Connect to the SQLite database
db_path = "text_chunks_with_embeddings.db"  # Update with the path to your database
conn = sqlite3.connect(db_path)
cursor = conn.cursor()

# Fetch the rows from the database
cursor.execute("SELECT text, embedding FROM chunks")
rows = cursor.fetchall()

# Create a dictionary to store the text and embedding for each row
dictionary_of_vectors = {}
for row in rows:
    text = row[0]
    embedding_str = row[1]
    embedding = np.fromstring(embedding_str, sep=' ')
    dictionary_of_vectors[text] = embedding

# Close the connection
conn.close()

def find_closest_neighbors(vector):
    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]

    sorted_cosine_similarities = sorted(cosine_similarities.items(), key=lambda x: x[1], reverse=True)
    return sorted_cosine_similarities[0:4]

def generate_embedding(text):
    response = openai.Embedding.create(
        input=text,
        engine="text-embedding-ada-002"
    )
    embedding = np.array(response['data'][0]['embedding'])
    return embedding

def context_gpt_response(question):
    vector = generate_embedding(question)
    match_list = find_closest_neighbors(vector)

    context = ''
    for match in match_list:
        context += str(match[0])
    
    context = context[:1500]  # Limit context to the last 1500 characters

    prep = f"This is an OpenAI model designed to answer questions specific to grant-making applications for an aquarium. Here is some question-specific context: {context}. Q: {question} A: "
    response = openai.Completion.create(
        engine="gpt-4",
        prompt=prep,
        temperature=0.7,
        max_tokens=220,
    )

    return response['choices'][0]['text']

iface = gr.Interface(fn=context_gpt_response, inputs="text", outputs="text", title="Aquarium Grant Application Chatbot", description="Context-specific chatbot for grant writing", examples=[["What types of projects are eligible for funding?"], ["Tell me more about the application process."], ["What will be the most impactful grant opportunities?"]])
iface.launch()