LDiHK

BUSINESS VIDEO

The BUSINESS VIDEO is at:

https://youtube.com/shorts/t1Bc-_SUxZg

Hosted Site

The website is hosted at:

https://ldihk.xyz/

The hosted frontend is configured to use the live Render backend API:

https://ldihk-api.onrender.com

Health check:

https://ldihk-api.onrender.com/health

The hosted demo runs with the frontend mock API disabled and sends browser API traffic to the Render service. Local development instructions are below.

Quick Start

Set up the local Python environment:

uv venv .venv
uv sync

Run the tests:

uv run python -m unittest discover -s backend/tests

Process the local Google Takeout export:

uv run python backend/scripts/process_youtube_usage.py

This writes:

data/processed/users/local_user/youtube_usage.v1.json

Build the v3 SQLite usage store:

uv run python backend/scripts/import_youtube_usage_sql.py

This writes:

data/processed/users/local_user/youtube_usage.v3.sqlite

Enrich v3 video durations when a YouTube Data API key is configured:

cp .env.example .env
# Set YOUTUBE_API_KEY in .env
uv run python backend/scripts/enrich_youtube_durations.py

Start the read-only API:

uv run flask --app backend.app run --host 127.0.0.1 --port 8000

Available endpoints:

GET http://127.0.0.1:8000/health
GET http://127.0.0.1:8000/api/users/local_user/youtube-usage
GET http://127.0.0.1:8000/api/v2/users/local_user/youtube-usage/temporal
POST http://127.0.0.1:8000/api/v3/query
POST http://127.0.0.1:8000/api/query
POST http://127.0.0.1:8000/api/imports
GET http://127.0.0.1:8000/api/imports/{import_id}

If uv cannot write to its default cache in a restricted environment, prefix commands with:

uv --cache-dir .uv-cache ...

Frontend

The pulled frontend lives in Frontend/.

cd Frontend
npm install
cp .env.example .env
npm run dev

For the real backend, set:

PUBLIC_API_URL=http://127.0.0.1:8000

For local browser development, the backend CORS allowlist includes Astro's default origin:

http://localhost:4321

Leave PUBLIC_API_URL empty to use the frontend's local Astro mock API routes.

Technical Diagram

flowchart LR
  subgraph Ingestion["Ingestion"]
    IG["Instagram usage data"]
    YT["YouTube usage data"]
    TT["TikTok / Douyin usage data"]
    RAW["Raw usage data"]
    IG --> RAW
    YT --> RAW
    TT --> RAW
  end

  subgraph Preprocessing["Preprocessing"]
    direction TB

    CLEAN["Clean and validate data"] --> STRUCTURE["Convert to structured format"] --> STANDARD["Standardized data structure"]
    STANDARD --> DB[("SQL database")]
    DB --> AI["AI analysis / augmentation"]
    AI --> FEATURES["Derived metrics and risk signals"]
  end

  subgraph Dashboard["Analysis Dashboard"]
    TEMPORAL["Temporal usage graph"]
    POP["Population comparison"]
    HEALTHY["Comparison with healthy / average usage"]
    RISK["Risk-factor trends over time"]
    CORR["Correlation analysis across groups"]
  end

  RAW --> CLEAN
  FEATURES --> TEMPORAL
  FEATURES --> POP
  FEATURES --> HEALTHY
  FEATURES --> RISK
  FEATURES --> CORR

Loading