Add files via upload

GenerativeAI.ipynb

@@ -0,0 +1,191 @@
+{
+  "nbformat": 4,
+  "nbformat_minor": 0,
+  "metadata": {
+    "colab": {
+      "provenance": []
+    },
+    "kernelspec": {
+      "name": "python3",
+      "display_name": "Python 3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "cells": [
+    {
+      "cell_type": "markdown",
+      "source": [
+        "\n",
+        "\n",
+        "\n",
+        "# Initialize generative AI: Gemini and pyttsx3 text-to-speech\n",
+        "\n",
+        "\n"
+      ],
+      "metadata": {
+        "id": "xajjvRn4ZQj_"
+      }
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "uDmI30KVpRmP"
+      },
+      "outputs": [],
+      "source": [
+        "import google.generativeai as genai\n",
+        "import pyttsx3\n",
+        "import cv2\n",
+        "import numpy as np\n",
+        "import time\n",
+        "from PIL import Image, ImageGrab\n",
+        "import customtkinter as tk\n",
+        "import pyautogui\n",
+        "import secret\n",
+        "\n",
+        "# Initialize Gemini Pro Vision\n",
+        "API_KEY = secret.API_KEY\n",
+        "genai.configure(api_key=API_KEY)\n",
+        "model = genai.GenerativeModel('gemini-pro-vision')\n",
+        "\n",
+        "# Initialize pyttsx3\n",
+        "speaker = pyttsx3.init()\n",
+        "speaker.setProperty('rate', 120)\n",
+        "voices = speaker.getProperty('voices')\n",
+        "speaker.setProperty('voice', voices[1].id)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "# Menu button function to start distraction hunting"
+      ],
+      "metadata": {
+        "id": "8r8RiXC0gn9M"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "def activate():\n",
+        "    if user_input.get() == \"\":\n",
+        "        label3 = tk.CTkLabel(app, text = \"You can't plan to do nothing!\", font=my_font, fg_color='indigo') # Empty input not allowed\n",
+        "        canvas.create_window(400,425, window=label3)\n",
+        "        return\n",
+        "    keyword = user_input.get()\n",
+        "    canvas.delete(\"all\") # Replace with new screen to signal to user distraction hunting has started\n",
+        "    button1 = tk.CTkButton(app, text=\"I'm done!\", command = app.destroy, font=my_font)\n",
+        "    canvas.create_window(400,400, window=button1)\n",
+        "    label5 = tk.CTkLabel(app, text = f\"Task: {keyword}\", font=my_font, fg_color='indigo') # Displays inputted task\n",
+        "    canvas.create_window(400,300, window=label5)\n",
+        "    focus(keyword, streak) # actual distraction hunting function"
+      ],
+      "metadata": {
+        "id": "psloRkHxgqeg"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "# Distraction hunting loop (AI backend)"
+      ],
+      "metadata": {
+        "id": "iHADdEm_hDOo"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# main distraction hunting manager\n",
+        "def focus(keyword, streak):\n",
+        "    screenshot() # take screenshot of user's screen\n",
+        "    picture = Image.open('s1.png')\n",
+        "    unrelated = judge(keyword, picture) # judge from the image whether the user is distarcted from their desired task\n",
+        "    if unrelated:\n",
+        "        streak += 1 # streak counter increases by 1 if screenshot implies that the user is distracted from their task\n",
+        "        if streak >= 2:\n",
+        "            taunt(keyword, picture) # streak of 2 or more shows that user is confirmed to be distracted for some time, so an audio reminder will be played\n",
+        "    else:\n",
+        "        streak = 0 # streak resets to 0 when user is deemed focused on their task again\n",
+        "    print(streak)\n",
+        "    app.after(1000, focus, keyword, streak) # constantly loops every 1 second while in distraction hunting phase\n",
+        "\n",
+        "# record screen and take screenshots\n",
+        "def screenshot():\n",
+        "    ss = pyautogui.screenshot()\n",
+        "    ss.save(\"s1.png\")\n",
+        "\n",
+        "# send a text prompt based on the task keyword and make judgement based on the screenshot\n",
+        "def judge(keyword, image):\n",
+        "    qn = f'Answer YES or NO in caps, nothing else. Is this screenshot directly or indirectly related to \"{keyword}\"?' # first prompt\n",
+        "    print(qn)\n",
+        "    response = model.generate_content(contents=[qn, image])\n",
+        "    print(response.text)\n",
+        "    if response.text == \"NO\" or \"NO\" in response.text:\n",
+        "        return True\n",
+        "    qn = f'Answer YES or NO in caps, nothing else. Is the user doing something else unrelated to the task \"{keyword}\"?' # second prompt\n",
+        "    print(qn)\n",
+        "    response = model.generate_content(contents=[qn, image])\n",
+        "    print(response.text)\n",
+        "    if response.text == \"YES\" or \"YES\" in response.text:\n",
+        "        return True\n",
+        "    return False # user is deemed not distracted when both prompts infer that the user is doing the inputted task\n",
+        "\n",
+        "# taunt the user with text-to-speech audio if it's clear they are distracted\n",
+        "def taunt(keyword, image):\n",
+        "    qn = f'This user is distracted right now. Reply below in ONE SENTENCE ONLY, NOTHING ELSE in a cheeky manner\\\n",
+        "    to remind the user based on what they are currently doing and they are distracted from the task \"{keyword}\"'\n",
+        "    print(qn)\n",
+        "    response = model.generate_content(contents=[qn, image])\n",
+        "    print(response.text)\n",
+        "    speaker.say(response.text)\n",
+        "    speaker.runAndWait()"
+      ],
+      "metadata": {
+        "id": "6P3b_tgThHtk"
+      },
+      "execution_count": null,
+      "outputs": []
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "# Main app + menu interface"
+      ],
+      "metadata": {
+        "id": "Wa2wT7RyhSlG"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "app = tk.CTk()\n",
+        "app.title(\"Anti-Distraction Assistant\")\n",
+        "my_font = tk.CTkFont(family=\"Courier\", size=35, weight=\"bold\")\n",
+        "user_input = tk.StringVar()\n",
+        "times = tk.StringVar()\n",
+        "keyword = \"\"\n",
+        "streak = 0\n",
+        "canvas = tk.CTkCanvas(app, bg='indigo', width=800, height=600)\n",
+        "canvas.pack()\n",
+        "label1 = tk.CTkLabel(app, text = \"What do you plan to do today?\", font=my_font, fg_color='indigo')\n",
+        "canvas.create_window(400,150, window=label1)\n",
+        "entry1 = tk.CTkEntry(app, textvariable = user_input, font=my_font, width=600)\n",
+        "canvas.create_window(400,300, window=entry1)\n",
+        "button = tk.CTkButton(app, text=\"Lock In\", command = activate, font=my_font) # calls the activate() function\n",
+        "canvas.create_window(400,500, window=button)\n",
+        "app.mainloop()"
+      ],
+      "metadata": {
+        "id": "cOHC2IwPhWKa"
+      },
+      "execution_count": null,
+      "outputs": []
+    }
+  ]
+}