fix: use new barcode normalization

- change the way we split barcodes to create directory hierarchy
- use images subdomain for both images and OCRs

openfoodfacts/images.py

@@ -1,5 +1,4 @@
 import logging
-import re
 from pathlib import Path
 from typing import List, Optional, Tuple, Union
 from urllib.parse import urlparse
@@ -12,7 +11,6 @@
 logger = logging.getLogger(__name__)
 
 
-BARCODE_PATH_REGEX = re.compile(r"^(...)(...)(...)(.*)$")
 # Base URL of the public Open Food Facts S3 bucket
 AWS_S3_BASE_URL = "https://openfoodfacts-images.s3.eu-west-3.amazonaws.com/data"
 
@@ -43,15 +41,21 @@ def split_barcode(barcode: str) -> List[str]:
     if not barcode.isdigit():
         raise ValueError(f"unknown barcode format: {barcode}")
 
-    match = BARCODE_PATH_REGEX.fullmatch(barcode)
+    # Pad the barcode with zeros to ensure it has 13 digits
+    barcode = barcode.zfill(13)
+    # Split the first 9 digits of the barcode into 3 groups of 3 digits to
+    # get the first 3 folder names
+    splits = [barcode[0:3], barcode[3:6], barcode[6:9], barcode[9:13]]
 
-    splits = [x for x in match.groups() if x] if match else [barcode]
+    # use the rest of the barcode as the last folder name
+    if len(barcode) > 13:
+        splits.append(barcode[13:])
 
     if org_id is not None:
         # For the pro platform only, images and OCRs belonging to an org
         # are stored in a folder named after the org for all its products, ex:
         # https://images.pro.openfoodfacts.org/images/products/org-lea-nature/330/713/080/3004/1.jpg
-        splits.append(org_id)
+        splits.insert(0, org_id)
 
     return splits
 
@@ -103,9 +107,8 @@ def generate_json_ocr_url(
         Environment.org
     :return: the generated JSON URL
     """
-    return (
-        URLBuilder.static(flavor, environment)
-        + f"/images/products{generate_json_ocr_path(code, image_id)}"
+    return URLBuilder.image_url(
+        flavor, environment, generate_json_ocr_path(code, image_id)
     )
 
 
@@ -141,6 +144,17 @@ def extract_barcode_from_url(url: str) -> Optional[str]:
 
 
 def extract_barcode_from_path(path: str) -> Optional[str]:
+    """Extract a product barcode from an image/OCR path.
+
+    The barcode is normalized using the following rules:
+
+    - all leading zeros are stripped
+    - if the barcode is less than 8 digits, it is left-padded with zeros up to
+      8 digits
+    - if the barcode is more than 8 digits but less than 13 digits, it is
+      left-padded with zeros up to 13 digits
+    - if the barcode has 13 digits or more, it's returned as it
+    """
     barcode = ""
 
     for parent in Path(path).parents:
@@ -149,7 +163,18 @@ def extract_barcode_from_path(path: str) -> Optional[str]:
         else:
             break
 
-    return barcode or None
+    # Strip leading zeros
+    barcode = barcode.lstrip("0")
+
+    if not barcode:
+        return None
+
+    if len(barcode) <= 8:
+        barcode = barcode.zfill(8)
+        return barcode
+
+    barcode = barcode.zfill(13)
+    return barcode
 
 
 def extract_source_from_url(url: str) -> str:

openfoodfacts/utils.py

@@ -111,7 +111,7 @@ def static(flavor: Flavor, environment: Environment) -> str:
     @staticmethod
     def image_url(flavor: Flavor, environment: Environment, image_path: str) -> str:
         prefix = URLBuilder._get_url(
-            prefix="static", tld=environment.value, base_domain=flavor.get_base_domain()
+            prefix="images", tld=environment.value, base_domain=flavor.get_base_domain()
         )
         return prefix + f"/images/products{image_path}"
 

tests/test_images.py

@@ -2,7 +2,13 @@
 
 import pytest
 
-from openfoodfacts.images import extract_barcode_from_url, extract_source_from_url
+from openfoodfacts.images import (
+    extract_barcode_from_url,
+    extract_source_from_url,
+    generate_image_url,
+    generate_json_ocr_url,
+)
+from openfoodfacts.types import Environment, Flavor
 
 
 @pytest.mark.parametrize(
@@ -43,3 +49,103 @@ def test_get_barcode_from_url(url: str, output: Optional[str]):
 )
 def test_get_source_from_url(url: str, output: str):
     assert extract_source_from_url(url) == output
+
+
+@pytest.mark.parametrize(
+    "code,image_id,flavor,environment,expected",
+    [
+        (
+            "5410126726954",
+            "1",
+            Flavor.off,
+            Environment.org,
+            "https://images.openfoodfacts.org/images/products/541/012/672/6954/1.jpg",
+        ),
+        (
+            "6539",
+            "1",
+            Flavor.off,
+            Environment.org,
+            "https://images.openfoodfacts.org/images/products/000/000/000/6539/1.jpg",
+        ),
+        (
+            "12458465",
+            "2.400",
+            Flavor.obf,
+            Environment.net,
+            "https://images.openbeautyfacts.net/images/products/000/001/245/8465/2.400.jpg",
+        ),
+        (
+            "org-lea-nature/5410126726954",
+            "1",
+            Flavor.off_pro,
+            Environment.org,
+            "https://images.pro.openfoodfacts.org/images/products/org-lea-nature/541/012/672/6954/1.jpg",
+        ),
+    ],
+)
+def test_generate_image_url(code, image_id, flavor, environment, expected):
+    assert generate_image_url(code, image_id, flavor, environment) == expected
+
+
+@pytest.mark.parametrize(
+    "code,image_id,flavor,environment,expected",
+    [
+        (
+            "5410126726954",
+            "1",
+            Flavor.off,
+            Environment.org,
+            "https://images.openfoodfacts.org/images/products/541/012/672/6954/1.json",
+        ),
+        (
+            "6539",
+            "1",
+            Flavor.off,
+            Environment.org,
+            "https://images.openfoodfacts.org/images/products/000/000/000/6539/1.json",
+        ),
+        (
+            "org-lea-nature/5410126726954",
+            "1",
+            Flavor.off_pro,
+            Environment.org,
+            "https://images.pro.openfoodfacts.org/images/products/org-lea-nature/541/012/672/6954/1.json",
+        ),
+    ],
+)
+def test_generate_json_ocr_url(code, image_id, flavor, environment, expected):
+    assert generate_json_ocr_url(code, image_id, flavor, environment) == expected
+
+
+@pytest.mark.parametrize(
+    "url,expected",
+    [
+        (
+            "https://world.openfoodfacts.org/images/products/541/012/672/6954/1.jpg",
+            "5410126726954",
+        ),
+        (
+            "https://world.openbeautyfacts.net/images/products/000/000/001/6954/1.jpg",
+            "00016954",
+        ),
+        (
+            "https://world.openbeautyfacts.net/images/products/000/009/121/6954/1.jpg",
+            "91216954",
+        ),
+        (
+            "https://world.openbeautyfacts.net/images/products/000/019/121/6954/1.jpg",
+            "0000191216954",
+        ),
+        (
+            "https://world.openbeautyfacts.net/images/products/343/919/121/6954/1.jpg",
+            "3439191216954",
+        ),
+        (
+            "https://world.openbeautyfacts.net/images/products/343/919/121/6954862052/1.jpg",
+            "3439191216954862052",
+        ),
+    ],
+)
+def test_extract_barcode_from_url(url, expected):
+    assert extract_barcode_from_url(url) == expected