scare_engine #

A pure Dart package for psychological vulnerability assessment through a guided, choice-based dialogue.

Features #

• Psychological Profiling: Categorizes user fears into Paranoia, Isolation, Loss of Control, and Existential.
• Dynamic Question Generation: Selects questions based on the user's dominant fear for targeted assessment.
• AI Integration Ready: Designed to work seamlessly with generative AI APIs for personalized interactions.
• Redux Friendly: Integrates well with Redux for state management, promoting immutability.
• Local Fallback: Provides a local ScareEngine implementation for continuity if AI integration fails.
• Multilingual Support: Questions provided in English and Arabic.

Getting Started #

To use this package, add scare_engine to your pubspec.yaml file:

dependencies:
  scare_engine: ^0.0.1

Then, run dart pub get to install the package.

Usage #

The ScareEngine orchestrates the psychological assessment process.

Basic Usage #

First, import the necessary components:

import 'package:scare_engine/scare_engine.dart';
import 'package:scare_engine/src/models/psych_profile.dart';
import 'package:scare_engine/src/models/scare_question.dart';
import 'package:scare_engine/src/data/question_loader.dart';

Initialize the engine and get the first question:

Future<void> main() async {
  // Load questions from a file path.
  // In a real app, this path would point to your questions JSON file.
  // For example, in a Flutter app, you might first copy it from assets.
  await QuestionLoader.loadQuestions('assets/questions.json');
  final List<ScareQuestion> questions = QuestionLoader.questions;

  // Initialize with an optional existing profile
  final scareEngine = ScareEngine(
    questionPool: questions,
    doctorName: "Dr. Enigma", // Optional: Customize the doctor's name
  );

  // You can also provide custom doctor notes:
  final customNotes = {
    FearType.paranoia: ["A custom note for paranoia."]
  };
  final customScareEngine = ScareEngine(
    questionPool: questions,
    customDoctorNotes: customNotes, // Provide your custom notes here
  );

  print("--- ${scareEngine.doctorName} Begins ---");

  // Get the first question for the 'en' (English) locale
  ScareQuestion currentQuestion = scareEngine.getNextQuestion(locale: 'en');
  print("Question: ${currentQuestion.getLocalizedQuestionText('en')}");
  print("Option A: ${currentQuestion.optionA.getLocalizedText('en')} (Impact: ${currentQuestion.optionA.impact})");
  print("Option B: ${currentQuestion.optionB.getLocalizedText('en')} (Impact: ${currentQuestion.optionB.impact})");

  // Simulate a user choice (e.g., choosing Option A)
  ScareOption userChoice = currentQuestion.optionA;
  String doctorNote = scareEngine.processUserChoice(userChoice);

  print("\nUser chose: ${userChoice.getLocalizedText('en')}");
  print("Updated PsychProfile scores: ${scareEngine.profile.scores}");
  print("Dominant Fear: ${scareEngine.profile.dominantFear}");
  print("${scareEngine.doctorName}'s Observation: $doctorNote");

  // Get the next question, which might be targeted based on the new dominant fear
  currentQuestion = scareEngine.getNextQuestion(locale: 'en');
  print("\nNext Question: ${currentQuestion.getLocalizedQuestionText('en')}");
}

AI Integration #

This package is designed to integrate with a generative AI API (e.g., Google Generative AI). The AI should act as "Dr. Abdelmoati" and adhere to the "ScareEngine System Prompt" provided in the docs/prompt.txt file.

The AI is expected to return questions in a specific JSON format:

{
  "question": "العربية (The Question Text)",
  "options": [
    {"text": "Option A", "impact": "fear_category"},
    {"text": "Option B", "impact": "fear_category"}
  ],
  "doctor_note": "A creepy observation about why they chose the previous answer."
}

Your app would send context (like user name/age) to the AI and parse its JSON response into a ScareQuestion object.

Redux Integration #

The PsychProfile class is designed with immutability in mind, making it suitable for Redux state management. When the AI returns an impact field or a local ScareOption is chosen, your Redux middleware should update the PsychProfile in your application's state.

The PsychProfile.copyWithAddedScore(FearType type) method returns a new PsychProfile instance with updated scores, facilitating immutable updates.

// Example of how Redux middleware might update the profile
// Assuming 'state.psychProfile' is the current profile in your Redux store
PsychProfile newProfile = state.psychProfile.copyWithAddedScore(chosenFearType);
// Dispatch an action to update the state with newProfile

Fear Categories #

The package categorizes psychological fears into:

• Paranoia: Fear of being watched or followed.
• Isolation: Fear of being alone or forgotten.
• Loss of Control: Fear of an external entity (like an app or AI) taking over.
• Existential: Fear of the unknown or death.

Contributing #

Contributions are welcome! Please feel free to open an issue or submit a pull request.

License #

This package is distributed under the MIT License. See the LICENSE file for more information.