api_host.py

from Perceptrix.engine import perceptrix, robotix, identify_objects_from_text, search_keyword
from CircumSpect import answer_question, find_object_description, locate_object
from flask import Flask, request, jsonify
import numpy as np
import threading
import whisper
import cv2
import os

model = whisper.load_model("base")


def transcribe(audio):
    result = model.transcribe(audio)
    transcription = result['text']
    print(transcription)
    return transcription


app = Flask(__name__)


@app.route('/locate_object', methods=['POST'])
def display_image():
    try:
        image_data = request.json['image_data']
        prompt = request.json['prompt']

        image_data = np.array(image_data, dtype=np.uint8)
        image = cv2.imdecode(image_data, cv2.IMREAD_COLOR)

        cv2.imwrite('API.jpg', image)
        answer, annotated_image = locate_object(prompt, "API.jpg")

        return jsonify({'message': answer, 'annotated_image': annotated_image})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/vqa', methods=['POST'])
def display_image():
    try:
        image_data = request.json['image_data']
        prompt = request.json['prompt']

        image_data = np.array(image_data, dtype=np.uint8)
        image = cv2.imdecode(image_data, cv2.IMREAD_COLOR)

        cv2.imwrite('API.jpg', image)
        answer = answer_question(prompt, "API.jpg")

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/object_description', methods=['POST'])
def display_image():
    try:
        image_data = request.json['image_data']

        image_data = np.array(image_data, dtype=np.uint8)
        image = cv2.imdecode(image_data, cv2.IMREAD_COLOR)

        cv2.imwrite('API.jpg', image)
        answer = find_object_description("API.jpg")

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/perceptrix', methods=['POST'])
def display_image():
    try:
        prompt = request.json['prompt']

        answer = perceptrix(prompt)

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/robotix', methods=['POST'])
def display_image():
    try:
        prompt = request.json['prompt']

        answer = robotix(prompt)

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/search_keyword', methods=['POST'])
def display_image():
    try:
        prompt = request.json['prompt']

        answer = search_keyword(prompt)

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/identify_objects_from_text', methods=['POST'])
def display_image():
    try:
        prompt = request.json['prompt']

        answer = identify_objects_from_text(prompt)

        return jsonify({'message': answer})

    except Exception as e:
        return jsonify({'error': str(e)})


@app.route('/transcribe', methods=['POST'])
def upload_audio():
    audio_file = request.files['audio']

    filename = os.path.join("database", audio_file.filename)
    audio_file.save(filename)
    return jsonify({'message': transcribe(filename)})


def run_app():
    app.run(port=7777)

if __name__ == "__main__":
    runner = threading.Thread(target=run_app)
    runner.start()