Initial commit.

ocado_grocy/images.py

@@ -0,0 +1,113 @@
+"""Copy saved Ocado product photos into Grocy's native product pictures."""
+import base64
+import hashlib
+import json
+from pathlib import Path
+import tomllib
+from urllib.parse import quote, urlsplit
+
+import click
+import requests
+
+from .grocy import Grocy, imported_products
+from .history import private_json
+from .product_metadata import read_metadata
+
+
+def image_type(data):
+    if data.startswith(b'\xff\xd8\xff'):
+        return 'jpg'
+    if data.startswith(b'\x89PNG\r\n\x1a\n'):
+        return 'png'
+    if data.startswith(b'RIFF') and data[8:12] == b'WEBP':
+        return 'webp'
+    raise ValueError('Image response is not JPEG, PNG or WebP')
+
+
+def file_path(filename):
+    return '/files/productpictures/' + quote(base64.b64encode(filename.encode()).decode(), safe='')
+
+
+def add_picture(client, downloader, product):
+    if product.get('picture_file_name'):
+        return 'existing_picture'
+    metadata = read_metadata(product.get('description'))
+    url = metadata.get('image_url')
+    if not url:
+        return 'no_image_url'
+    parsed = urlsplit(url)
+    if parsed.scheme != 'https' or parsed.hostname not in {'www.ocado.com', 'ocado.com'}:
+        raise ValueError('Expected an HTTPS Ocado image URL')
+    # Separate session: never send the Grocy API key to the image host.
+    response = downloader.get(url, timeout=30, allow_redirects=False)
+    response.raise_for_status()
+    if response.status_code != 200:
+        raise ValueError('Image URL did not return HTTP 200')
+    data = response.content
+    extension = image_type(data)
+    if len(data) > 10_000_000:
+        raise ValueError('Image exceeds 10 MB')
+    filename = f"ocado-{int(product['id'])}-{hashlib.sha256(data).hexdigest()[:16]}.{extension}"
+    current = client.get(f"/objects/products/{product['id']}")
+    if not current:
+        return 'deleted_product'
+    if current.get('picture_file_name'):
+        return 'existing_picture'
+    if read_metadata(current.get('description')).get('Ocado product ID') != metadata.get('Ocado product ID'):
+        raise ValueError('Product identity changed')
+    path = file_path(filename)
+    uploaded = client.session.put(client.url + path, data=data,
+                                  headers={'Content-Type':'application/octet-stream'},
+                                  timeout=client.timeout, allow_redirects=False)
+    uploaded.raise_for_status()
+    if uploaded.status_code not in (200, 201, 204):
+        raise ValueError('Unexpected upload response')
+    # Check the stored file before setting the product's picture reference.
+    stored = client.session.get(client.url + path, timeout=client.timeout, allow_redirects=False)
+    stored.raise_for_status()
+    if stored.status_code != 200 or stored.content != data:
+        raise ValueError('Uploaded image verification failed')
+    client.request('PUT', f"/objects/products/{product['id']}", {'picture_file_name':filename})
+    return 'updated'
+
+
+@click.command()
+@click.option('--config', type=click.Path(path_type=Path), default=Path('config.toml'))
+@click.option('--report', type=click.Path(path_type=Path), default=Path('history/images/report.json'))
+@click.option('--apply', is_flag=True, help='Upload and assign missing product pictures.')
+@click.option('--order-id', 'order_ids', multiple=True, help='Limit to products imported from these orders.')
+def main(config, report, apply, order_ids):
+    """Add Ocado images to journal-linked products; keep existing pictures."""
+    settings = tomllib.loads(config.read_text())
+    client = Grocy(**settings['grocy'])
+    state = config.parent / settings.get('import', {}).get('state_file', 'imports.sqlite3')
+    products = imported_products(client, state, order_ids)
+    return run_images(client, products, report, apply=apply)
+
+
+def run_images(client, products, report, *, apply=True, raise_errors=True):
+    downloader = requests.Session()
+    downloader.headers['User-Agent'] = 'ocado-grocy/0.1'
+    results = []
+    private_json(report, {'apply':apply, 'products':results})
+    for product in products:
+        row = {'product_id':product['id'], 'name':product['name']}
+        try:
+            if apply:
+                row['status'] = add_picture(client, downloader, product)
+            else:
+                row['status'] = 'existing_picture' if product.get('picture_file_name') else 'ready' if read_metadata(product.get('description')).get('image_url') else 'no_image_url'
+        except (requests.RequestException, ValueError) as exc:
+            row.update(status='error', error=str(exc))
+        results.append(row)
+        private_json(report, {'apply':apply, 'products':results})
+        click.echo(f"{product['id']} {row['status']}: {product['name']}")
+    counts = {s:sum(r['status']==s for r in results) for s in sorted({r['status'] for r in results})}
+    click.echo(json.dumps(counts))
+    if counts.get('error') and raise_errors:
+        raise click.ClickException(f"{counts['error']} image updates failed; see {report}")
+    return {'products':results, 'counts':counts}
+
+
+if __name__ == '__main__':
+    main()