Skip to content

SPECIFICATIONS.md

Latest commit

 

History

History
118 lines (97 loc) · 4.65 KB

SPECIFICATIONS.md

File metadata and controls

118 lines (97 loc) · 4.65 KB

Extension Specifications v1.1 – 21st June 2024

Context and Objective

  1. Context:

    • Google Chrome Extension
    • Semi-automates Prompt creation
    • User selects options in predefined selectors
    • Creates a prompt based on template filled by user’s selections
    • Copy-pastes prompt in AI chatbot (e.g., ChatGPT, Perplexity.ai, Google Gemini)

  2. Objective:

    • Present two user stories (US_1 and US_2) and related coding notes
    • Example of application of US_1 and US_2
    • Technical considerations for coding the extension

User Stories and Notes for Coding

  1. User Story 1 (US_1): User Profile Screen

    • US_1-1: User discovers the interface
      • USER_LANGUAGE Button: User selects language option (Français or English)
        • System action: Changes interface language and displays introduction text
      • Introduction Text: Display text based on selected language
    • US_1-2: User completes their profile
      • ROLE Selector: User selects one option (Student or Teacher)
      • COUNTRY Selector: User selects one option (Country options)
      • CLASS Selector: User selects option(s) based on role and country
      • SUBJECTS Selector: User selects multiple options (Subjects list)
      • TOWN Entry: User enters town
      • BIRTHDATE Entry: User enters birth date (DD/MM/YYYY)
      • GENDER Selector: User selects gender (Woman or Man)
      • EMAIL Entry: User enters email
    • US_1-3: User is redirected to ‘Prompt for AI’ screen
      • VALIDATE Button: Submits profile, saves to local storage, adds row in Google Sheet, redirects to ‘Prompt for AI’

  2. User Story 2 (US_2): Prompt for AI Screen

    • US_2-1: User selects options
      • CLASS Selector: User selects class
      • SUBJECT Selector: User selects subject
      • OBJECTIVE Selector: User selects objective (Narrowed by role)
      • NEED Selector: User selects need (Narrowed by objective)
      • RESPONSE_LANGUAGE Selector: User selects response language (Narrowed by subject and user language)
      • DETAILS Entry: User enters specific details of request
    • US_2-2: Customized editable prompt created
      • PROMPT_CREATION Button: Validates to get the customized prompt
      • Prompt Template filled: Based on user language and selections
    • US_2-3: Prompt pasted in external chatbot
      • LAUNCH_AI Button: Validates prompt, copies it, opens chatbot URL, pastes prompt, increments Google Sheet

Selectors’ Dependencies

  • US_2-1_E (Response Language): Narrowed by US_1-1_A (User Language) and US_2-1_B (Subject)
  • US_1-2_C (Class): Narrowed by US_1-2_B (Country)
  • US_2-1_A (Class): Narrowed by US_1-2_C (Class)
  • US_2-1_B (Subject): Narrowed by US_1-2_D (Subject)
  • US_2-1_C (Objective): Narrowed by US_1-2_A (Role)
  • US_2-1_D (Need): Narrowed by US_2-1_C (Objective)
  • US_2-3_A (Short Prompt): Narrowed by US_2-1_D (Need)
  • US_2-1_D (Chatbot URL): Narrowed by US_2-1_D (Need)

Example Output of User Story 2

  1. User’s Selections and Entries:

    • US_1:
      • Language: Français
      • Role: Professeur
      • Country: France
      • Class: 3ème and 4ème
      • Subject: Maths, English, Histoire
    • US_2:
      • Class: 3ème
      • Subject: Histoire
      • Objective: Création de cours
      • Needs: des Données chiffrées
      • Response language: Français
      • Details: 2nde guerre mondiale

  2. Prompt Automatically Created:

    • Role: Teacher in France, teaches "Histoire" in "3ème" class, known for pedagogy
    • Task: Assist "Professeur" with "Création de cours", provide "des Données chiffrées" on "2nde guerre mondiale"
    • Instructions: Ask for clarifications, prepare step-by-step, write best answer
    • Answer Format: Language: Français, Structure: Paragraphs with headings in markdown
    • Important: Take time, be precise and pedagogical

Technical Considerations

  1. Data Storage:

    • No database
    • Selectors’ options and prompt templates hardcoded
    • User profile stored in browser’s local storage

  2. Code Quality:

    • Manifest v3
    • Load time within 2 seconds
    • 200 simultaneous users without performance degradation
    • Minimal memory and CPU usage
    • 99.9% uptime
    • Code reviews for compliance and best practices

  3. Code Error Handling:

    • Input validation for mandatory fields
    • Retry logic for network errors
    • Handle storage limits
    • Update selectors’ options correctly
    • Handle network errors for prompt creation
    • Loading time feedback and concurrency handling
    • Global error handler and user issue reporting
⚡