fix: use new barcode normalization

pyproject.toml

@@ -74,7 +74,7 @@ lark = "~1.1.4"
 h5py = "~3.8.0"
 opencv-contrib-python-headless = "~4.10.0.84"
 toml = "~0.10.2"
-openfoodfacts = "1.1.1"
+openfoodfacts = "1.1.2"
 imagehash = "~4.3.1"
 peewee-migrate = "~1.12.2"
 diskcache = "~5.6.3"

robotoff/app/api.py

@@ -18,6 +18,7 @@
 import requests
 from falcon.media.validators import jsonschema
 from openfoodfacts import OCRResult
+from openfoodfacts.images import extract_barcode_from_url
 from openfoodfacts.ocr import OCRParsingException, OCRResultGenerationException
 from openfoodfacts.types import COUNTRY_CODE_TO_NAME, Country
 from PIL import Image
@@ -62,12 +63,7 @@
     batch_insert,
     db,
 )
-from robotoff.off import (
-    OFFAuthentication,
-    generate_image_path,
-    generate_json_ocr_url,
-    get_barcode_from_url,
-)
+from robotoff.off import OFFAuthentication, generate_image_path, generate_json_ocr_url
 from robotoff.prediction import image_classifier, ingredient_list
 from robotoff.prediction.category import predict_category
 from robotoff.prediction.langid import predict_lang
@@ -543,7 +539,7 @@ class OCRPredictionPredictorResource:
     def on_get(self, req: falcon.Request, resp: falcon.Response):
         ocr_url = req.get_param("ocr_url", required=True)
         server_type = get_server_type_from_req(req)
-        barcode = get_barcode_from_url(ocr_url)
+        barcode = extract_barcode_from_url(ocr_url)
         prediction_types = req.get_param_as_list(
             "prediction_types",
             default=DEFAULT_OCR_PREDICTION_TYPES,

robotoff/cli/insights.py

@@ -8,10 +8,10 @@
 import dacite
 import orjson
 import tqdm
+from openfoodfacts.images import extract_barcode_from_path
 from openfoodfacts.ocr import OCRResult
 
 from robotoff.insights.extraction import DEFAULT_OCR_PREDICTION_TYPES
-from robotoff.off import get_barcode_from_path
 from robotoff.prediction.ocr import extract_predictions
 from robotoff.prediction.ocr.core import ocr_content_iter
 from robotoff.types import Prediction, PredictionType, ProductIdentifier, ServerType
@@ -83,7 +83,7 @@ def generate_from_ocr_archive(
         if source_image is None:
             continue
 
-        barcode: Optional[str] = get_barcode_from_path(source_image)
+        barcode = extract_barcode_from_path(source_image)
 
         if barcode is None:
             logger.warning("cannot extract barcode from source: %s", source_image)

robotoff/cli/main.py

@@ -527,10 +527,11 @@ def run_object_detection_model(
     from urllib.parse import urlparse
 
     import tqdm
+    from openfoodfacts.images import extract_barcode_from_url
     from peewee import JOIN
 
     from robotoff.models import ImageModel, ImagePrediction, db
-    from robotoff.off import generate_image_url, get_barcode_from_url
+    from robotoff.off import generate_image_url
     from robotoff.utils import text_file_iter
     from robotoff.workers.queues import enqueue_job, low_queue
     from robotoff.workers.tasks.import_image import (
@@ -588,7 +589,7 @@ def run_object_detection_model(
 
     if typer.confirm(f"{len(image_urls)} jobs are going to be launched, confirm?"):
         for image_url in tqdm.tqdm(image_urls, desc="image"):
-            barcode = get_barcode_from_url(image_url)
+            barcode = extract_barcode_from_url(image_url)
             if barcode is None:
                 raise RuntimeError()
             enqueue_job(

robotoff/off.py

@@ -80,23 +80,6 @@ def get_source_from_url(ocr_url: str) -> str:
     return url_path
 
 
-def get_barcode_from_url(url: str) -> Optional[str]:
-    url_path = urlparse(url).path
-    return get_barcode_from_path(url_path)
-
-
-def get_barcode_from_path(path: str) -> Optional[str]:
-    barcode = ""
-
-    for parent in Path(path).parents:
-        if parent.name.isdigit():
-            barcode = parent.name + barcode
-        else:
-            break
-
-    return barcode or None
-
-
 def _generate_file_path(product_id: ProductIdentifier, image_id: str, suffix: str):
     splitted_barcode = split_barcode(product_id.barcode)
     return f"/{'/'.join(splitted_barcode)}/{image_id}{suffix}"

tests/unit/test_off.py

@@ -1,27 +1,6 @@
-from typing import Optional
-
 import pytest
 
-from robotoff.off import get_barcode_from_url, get_source_from_url
-
-
-@pytest.mark.parametrize(
-    "url,output",
-    [
-        (
-            "https://world.openfoodfacts.org/images/products/541/012/672/6954/1.jpg",
-            "5410126726954",
-        ),
-        (
-            "https://world.openfoodfacts.org/images/products/541/012/672/6954/1.json",
-            "5410126726954",
-        ),
-        ("https://world.openfoodfacts.org/images/products/invalid/1.json", None),
-        ("https://world.openfoodfacts.org/images/products/252/535.bk/1.jpg", None),
-    ],
-)
-def test_get_barcode_from_url(url: str, output: Optional[str]):
-    assert get_barcode_from_url(url) == output
+from robotoff.off import get_source_from_url
 
 
 @pytest.mark.parametrize(