# Pokemon Discovery (pokemon-disco)

Scrapes Pokemon TCG card pack and tin products from Dollar General and generates a PDF product catalog with images and UPC-A barcodes.

## How It Works

Dollar General's Pokemon category page loads products dynamically via an internal API. A browser HAR capture contains the API responses with all product data. `disco.py` extracts products from the HAR file, filters for card packs and tins, downloads product images, generates UPC-A barcodes, and produces a LaTeX-based PDF catalog.

### Pipeline

```
HAR file → Extract API responses → Filter packs/tins → Download images
         → Generate UPC-A barcodes → Compile PDF catalog (pdflatex)
```

## Requirements

- Python 3.10+
- pdflatex (via `texlive-core` + `texlive-latexextra`)
- Python packages: `requests`, `beautifulsoup4`, `python-barcode`, `Pillow`

### Install (Arch / CachyOS)

```bash
sudo pacman -S texlive-basic texlive-latex texlive-latexextra texlive-fontsrecommended
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
```

## Usage

### Full run (scrape + PDF)

```bash
source venv/bin/activate
python disco.py
```

### Scrape only (output JSON)

```bash
python disco.py --scrape-only
```

### PDF only (from existing JSON)

```bash
python disco.py --pdf-only pokemon_tcg_products_YYYYMMDD_HHMMSS.json
```

## Output

```
pokemon_tcg_products_YYYYMMDD_HHMMSS.json    Product data
catalog_output/
├── pokemon_catalog_YYYYMMDD_HHMMSS.pdf      PDF catalog
├── pokemon_catalog_YYYYMMDD_HHMMSS.tex      LaTeX source
├── images/                                   Product images (PNG)
└── barcodes/                                 UPC-A barcodes (PNG)
```

### PDF Layout

**Page 1 — Manifest:** table of all products with SKU, price, and stock count.

**Product pages:**

```
Product Name
Stock status                              Price
SKU: XXXXXXXX                   UPC: XXXXXXXXXXXX

┌─────────────────────────────┐
│                             │
│       Product Image         │
│                             │
└─────────────────────────────┘

┌─────────────────────────────┐
│      UPC-A Barcode          │
└─────────────────────────────┘
```

## Capturing a HAR File

The HAR file provides product data from Dollar General's internal API. To capture one:

1. Open your browser (Brave, Chrome, Firefox)
2. Open DevTools → **Network** tab
3. Visit `https://www.dollargeneral.com/c/toys/pokemon?q=`
4. Wait for products to load, toggle any filters you want
5. Right-click in the Network tab → **Save all as HAR**
6. Place the `.har` file in the project root

`disco.py` looks for any `.har` file matching the default name pattern. Edit the `HAR_FILE` constant at the top of `disco.py` if your filename differs.

## Files

| File | Purpose |
|------|---------|
| `disco.py` | Main tool — scrape, filter, generate PDF |
| `scraper.py` | Reference site scraper (HTML + Selenium/Brave) |
| `requirements.txt` | Python dependencies |
| `*.har` | Browser HAR capture with API data |

## API Details (Reference)

The product data comes from this internal API:

```
POST https://dggo.dollargeneral.com/omni/api/v2/category/search/provider
Content-Type: application/json
Authorization: Bearer <session-token>

{
  "StoreNbr": 17506,
  "Id": 723960,          // Pokemon category
  "PageSize": 24,
  "Filters": {
    "soldAtStore": true,
    "inStock": false      // false = include out of stock
  }
}
```

Response contains `ItemList.Items[]` with fields: `Description`, `UPC`, `Price`, `Image`, `AvailableQty`, `rootSV` (internal ID → SKU).

The bearer token is session-scoped and short-lived. `disco.py` sidesteps this by reading the API responses directly from a HAR capture.