Week #2

Week #2 #

Detailed Requirements Elaboration #

This week’s primary focus was to develop a low-fidelity wireframe for the project’s frontend interface, analyze the structural design of databases, outline the ML (Machine Learning) core functionality, and map the backend workflow. Additionally, the task included analyzing and creating a user flow to ensure seamless navigation and usability.

Prioritized backlog #

The backlog link: Backlog
As a Kanban board we create a GitHub Project

Project specific progress #

Design #

Low-Fidelity Wireframe Creation #

Establish a foundational layout for user interfaces and workflows before committing to visual details.

Figma Prototype: Developed a low-fidelity wireframe focusing on:

  • Core Screens:
    • Login / Register,
    • Home screen
    • Settings / Profile
    • screens with other functional
      • Dictionary
      • Diary
      • Lesson screens
      • Statistic screen
      • Calendar screen

Visual Design & Theming #

Define a comprehensive visual identity to enhance usability and brand alignment.

  • Design sketches based on low-fidelity prototypes
  • Color Theme: Selected a primary palette (Primary orange, white, and black, purple, blue as secondary colors)

Data Analysis #

Reverse Engineering & Functional Specification #

Conducted reverse engineering of competitor platforms:

  • Platforms Studied: Duolingo, Wordich, RE:Word, Babbel, Quizlet
  • Objective: Systematically evaluate strengths to inform our feature prioritization.

Duolingo #

  • Typical exercises:

    • Filling in the correct word in a phrase
    • Text (dialogue), answering questions about it or filling in blanks
    • Listening to/reading a task and selecting/typing/voicing the answer
    • Constructing sentences from words

  • Functional features:

    • Full sentence pronunciation with new words
    • Users can tap a word to see its translation and hear pronunciation
    • Rating and reward system

RE:Word #

  • Typical exercises:

    • Before starting, users see words with swipe options <-- don't know> <know --> to build a word base
    • Typing a specific word / selecting from options / seeing translation

  • Functional features:

    • Words are added and categorized manually by users

Wordich #

  • Typical exercises:

    • User receives a spoken word, English spelling - translation, example sentence with the word and its translation
    • User listens to the word and types it
    • User receives a word and selects 1 of 4 translation options (English -> Russian and vice versa)

  • Functional features:

    • All functionality is implemented through a Telegram bot that regularly sends notifications and reminders

Babbel #

  • Typical exercises:

    • Selecting picture-word flashcards
    • Filling words into dialogues

  • Functional features: none

Quizlet #

  • Typical exercises:

    • Flashcard - translation, followed by memorization exercises

  • Functional features:

    • Users can create their own flashcards and browse community cards

Database Engineering #

Designed words Databse: database schema

Database

link to the DB

Frontend #

  • Planned user flow of application
  • Understand common principles
  • Started developing iOS and Android apps

Backend #

  • Rethought the structure of authorization through Google (Rethought the fields for user and preferences models)
  • Make full setup Google and email&password based authentication with JWT token generation
  • Added an admin panel for CRUD database operations via a graphical interface
  • The models in the code have been rewritten and new ones have been added with handlers of operations with them

ML #

Model Investigation & Selection #

  • Researched and evaluated multiple NLP architectures
  • Selected BERT with MLM (Masked Language Modeling)
  • Started developing of chosen model

DevOps #

Updated README #

Changes Pushed to Main:

  • Added two installation guides:
    1. Local Setup (recommended for TA)
    2. Full Deployment

Key Improvements:

Local Setup Guide #

  • Designed specifically for tests of the system locally via localhost
  • Includes:
    • One-command environment setup (Docker Compose)
    • Test data injection script
    • Health check endpoints for validation

Mobile Deployment #

  • Android
  • iOS

Updated README
Main site
Swagger (API documentation)
Admin Panel

Weekly commitments #

Individual contribution of each participant #

Danila Kochegarov (Team Lead & Backend Developer) #

  • Rethought the structure of authorization through Google (Rethought the fields for user and preferences models)
  • Organize work in the team (creating issues, communicate with each of the people on their topic, solve technical and theme, app logic issues)
  • Populate database with 5950 words from Oxford dictionary (word, part of speech, CEFR level)
  • Make full setup Google and email&password based authentication with JWT token generation

Issues links:

  1. User model revision
  2. Populate DB
  3. Google AUTH PR.
    Related issues:

Savva Ponomarev (iOS Developer & Product Manager) #

  • Created backlog for a week
  • Designed with Evgeniy low-fidelity design of the screens
  • Made skethes for design app
  • Set-up the iOS project structure, with fonts and colors
  • Wrote a report
  • Started Developing screens:
    • Home screen

link to Backlog
links to issues:
https://github.com/FluentlyOrg/Fluently-fork/issues/18
https://github.com/FluentlyOrg/Fluently-fork/issues/21
https://github.com/FluentlyOrg/Fluently-fork/issues/11

link to PR:
Home screen

George Selivanov (System Analyst) #

  • Conducted reverse engineering of competitor products to define core application functionality
  • Collaboratively designed the database schema with Danya
  • Researched and selected optimal data sources for parsing
  • Mapped user flows based on competitive analysis findings

links to issues:

Timofey Ivlev (DevOps Engineer) #

Updated README https://github.com/FluentlyOrg/Fluently-fork/issues/53

Anton Korotkov (ML Engineer) #

  • Researched and evaluated multiple NLP architectures
  • Researched pre-tuned models capable of future fine-tuning
  • Use BERT with MLM (Masked Language Modeling) concepts for:
    • Contextual understanding of word relationships
    • Native mask prediction capabilities
  • Suggested to use WordNet and ConceptNet

Daniil Tskhe (Backend Developer) #

  • Rethought the structure of authorization through Google
  • I have linked an admin panel for interacting with the database through a graphical interface
  • Rethought the fields for user and preferences models
  • Changed user and preferences models in the code and added operations with them
  • Added additional models and operations with them

Links to issues:
https://github.com/FluentlyOrg/Fluently-fork/issues/26 https://github.com/FluentlyOrg/Fluently-fork/issues/37 https://github.com/FluentlyOrg/Fluently-fork/issues/46 https://github.com/FluentlyOrg/Fluently-fork/issues/47 https://github.com/FluentlyOrg/Fluently-fork/issues/49

Evgeniy Bortsov (Android Developer) #

  • Made a Low-fidelity design for lesson part
  • Set-up Android project sructure
  • Created User flow diagram

links to issues:
https://github.com/FluentlyOrg/Fluently-fork/issues/16
https://github.com/FluentlyOrg/Fluently-fork/issues/18#issuecomment-2980762714

Plan for Next Week #

Design #

  • Finish UI design

Frontend #

  • Create all key screens for both OS
    • iOS
    • Android
  • Make navigation
  • Network layer for working with API

Backend #

  • configure authentication on swagger page
  • make telegram bot to learn words and see some stats
  • write tests for API
  • check test coverage via SonarQube

ML #

  • Develop smart system that will generate exercises for lessons

Data Analysis #

  • Recollect datasets
  • Parsing info from new sources

DevOps #

  • configure SonarQube
  • configure endpoint for SonarQube
  • Set up Prometheus+Grafana for monitoring API load, security issues
  • Set up Loki for logs aggregations
  • Set up a second deploy server with domain for testing

P.S. #

Acceptance criteria

Confirmation of the code’s operability #

We confirm that the code in the main branch:

  • In working condition.
  • Run via docker-compose (or another alternative described in the README.md).