DaySale

Upload a daily sales report (.htm, .html, or .pdf) and get back clean, ready-to-copy numbers: Today Sale, Opening Balance, and the CHEERS breakdowns.

One Flask app does everything — no separate backend, no build step for the UI.

Run it

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python application.py

Open http://127.0.0.1:2323 and drop in a report file.

Deploy on Vercel

Vercel detects the Flask app from wsgi.py and installs dependencies from requirements.txt. Import the repo in Vercel and deploy — no build command or extra settings are needed.

• Files in public/ are served from Vercel's CDN; Flask handles /api/* and the health check.
• Vercel rejects request bodies over 4.5 MB before they reach the app, so the UI and Flask both cap uploads at 4.5 MB.
• Set optional environment variables in Project Settings → Environment Variables.

API

• POST /api/parse — send a file as multipart field file, get { "success": true, "result": { ... } } back
• GET /healthz (or /api/test) — health check
• GET /* — serves the UI from public/

Settings

All optional — each one falls back to the default shown:

Variable	What it does	Default
DEPOT_NAME	Depot shown in reports	MDCL-2
DEPOT_PREFIX	Report prefix	Depot
PARSE_TIMEOUT_MS	Give up on a parse after this long	50000
RATE_LIMIT_WINDOW_MS / RATE_LIMIT_MAX	Requests allowed per window	60000 / 30

Project layout

daysale/
├── application.py   # Flask routes
├── wsgi.py          # Vercel entrypoint
├── lib/
│   ├── parse.py     # picks the right parser
│   ├── html_parser.py
│   ├── pdf_parser.py
│   ├── sales.py     # Today Sale numbers
│   ├── stock.py     # Opening Balance numbers
│   ├── cheers.py    # CHEERS numbers
│   └── config.py    # product codes and depot info
├── public/          # index.html, styles.css, ui.js, api.js, script.js, favicon.svg
└── tests/           # Playwright regression tests and fixtures

Tests

The browser regression suite covers upload, report tabs, copy, product search, the product table, errors, and mobile layout.

npm install
npx playwright install chromium
npm test

The tests start the Flask app on port 4323 with a higher rate limit and use the fixtures in tests/fixtures/. To watch the tests run, use npm run test:headed.

Making changes

You want to change...	Edit
How HTML reports are read	lib/html_parser.py
How PDF reports are read	lib/pdf_parser.py
The numbers or report wording	lib/sales.py, lib/stock.py, lib/cheers.py
Product codes or depot info	lib/config.py

Good to know

• The UI doesn't care whether the file came in as HTML or PDF — both produce the same result shape.
• PDFs need selectable text. OCR could be added inside lib/pdf_parser.py later without touching the UI.
• Uploads are capped at 4.5 MB in both the UI and Flask to match Vercel's request-body limit.
• Static asset links in public/index.html carry a ?v= query. Bump it after changing any file in public/ so browsers and the CDN pick up the new version.
• The rate limit is counted in memory, so it applies per process; on Vercel, each function instance keeps its own counter.