feat: Added origin extraction

.env

@@ -8,8 +8,8 @@ PO_LOCAL_NET=po_default
 
 # User uid - in dev align this with your own user uid / gid
 # see https://gist.github.com/alexgarel/6e6158ee869d6db2192e0441fd58576e
-# OFF_UID=1000
-# OFF_GID=1000
+ OFF_UID=1000
+ OFF_GID=1000
 
 # Robotoff
 TAG=latest

Makefile

@@ -2,7 +2,7 @@
 
 # nice way to have our .env in environment for use in makefile
 # see https://lithic.tech/blog/2020-05/makefile-dot-env
-# Note: this will mask environment variable as opposed to docker-compose priority
+# Note: this will mask environment variable as opposed to docker compose priority
 # yet most developper should'nt bump into this
 ifneq (,$(wildcard ./.env))
     -include .env
@@ -14,8 +14,8 @@ NAME = "robotoff"
 ENV_FILE ?= .env
 MOUNT_POINT ?= /mnt
 HOSTS=127.0.0.1 robotoff.openfoodfacts.localhost
-DOCKER_COMPOSE=docker-compose --env-file=${ENV_FILE}
-DOCKER_COMPOSE_TEST=COMPOSE_PROJECT_NAME=robotoff_test PO_LOCAL_NET=po_test docker-compose --env-file=${ENV_FILE}
+DOCKER_COMPOSE=docker compose --env-file=${ENV_FILE}
+DOCKER_COMPOSE_TEST=COMPOSE_PROJECT_NAME=robotoff_test PO_LOCAL_NET=po_test docker compose --env-file=${ENV_FILE}
 
 .DEFAULT_GOAL := dev
 # avoid target corresponding to file names, to depends on them
@@ -50,7 +50,7 @@ edit_etc_hosts:
 # Docker Compose #
 #----------------#
 up:
-# creates a docker network and runs docker-compose
+# creates a docker network and runs docker compose
 	@echo "🥫 Building and starting containers …"
 	docker network create po_default || true  
 	${DOCKER_COMPOSE} up -d --build 2>&1
@@ -76,7 +76,7 @@ livecheck:
 	docker/docker-livecheck.sh
 
 log:
-	@echo "🥫 Reading logs (docker-compose) …"
+	@echo "🥫 Reading logs (docker compose) …"
 	${DOCKER_COMPOSE} logs -f
 
 #------------#

doc/references/api.md

@@ -163,3 +163,19 @@ Generate spellcheck corrections. Either the barcode or the text to correct must
   }
 }
 ```
+
+### Orgin prediction [/predict/origin]
+
+#### Predict origin from OCR JSON [GET]
+
+- ocr_url (str, required) - the url of the OCR JSON
+
+- Response 200 (application/json)
+
+```json
+[{
+  "origin": "en:france",
+  "same_for_all_ingredients": False, //if true, all ingredients have the same origin
+  "concerned_ingredients": ["en:quinoa"] // if same_for_all_ingredients is True, this field is null
+}]
+```

robotoff/app/api.py

@@ -365,6 +365,41 @@ def on_get(self, req: falcon.Request, resp: falcon.Response):
         resp.media = {"nutrients": [p.to_dict() for p in predictions]}
 
 
+class OriginPredictorResource:
+    def on_get(self, req: falcon.Request, resp: falcon.Response):
+        ocr_url = req.get_param("ocr_url", required=True)
+
+        if not ocr_url.endswith(".json"):
+            raise falcon.HTTPBadRequest("a JSON file is expected")
+
+        barcode = get_barcode_from_url(ocr_url)
+
+        if barcode is None:
+            raise falcon.HTTPBadRequest(f"invalid OCR URL: {ocr_url}")
+
+        try:
+            predictions = extract_ocr_predictions(
+                barcode, ocr_url, [PredictionType.origin]
+            )
+
+        except requests.exceptions.RequestException:
+            resp.media = {
+                "error": "download_error",
+                "error_description": "an error occurred during OCR JSON download",
+            }
+            return
+
+        except OCRParsingException as e:
+            logger.error(e)
+            resp.media = {
+                "error": "invalid_ocr",
+                "error_description": "an error occurred during OCR parsing",
+            }
+            return
+
+        resp.media = {"origin": [p.to_dict() for p in predictions]}
+
+
 class OCRInsightsPredictorResource:
     def on_get(self, req: falcon.Request, resp: falcon.Response):
         ocr_url = req.get_param("ocr_url", required=True)
@@ -1172,6 +1207,7 @@ def on_get(self, req: falcon.Request, resp: falcon.Response):
 api.add_route("/api/v1/insights/annotate", AnnotateInsightResource())
 api.add_route("/api/v1/predict/ingredients/spellcheck", IngredientSpellcheckResource())
 api.add_route("/api/v1/predict/nutrient", NutrientPredictorResource())
+api.add_route("/api/v1/predict/origin", OriginPredictorResource())
 api.add_route("/api/v1/predict/ocr_insights", OCRInsightsPredictorResource())
 api.add_route("/api/v1/predict/category", CategoryPredictorResource())
 api.add_route("/api/v1/products/dataset", UpdateDatasetResource())

robotoff/insights/extraction.py

@@ -22,6 +22,7 @@
     PredictionType.store,
     PredictionType.packaging,
     PredictionType.nutrient,
+    PredictionType.origin,
     PredictionType.nutrient_mention,
     PredictionType.image_lang,
     PredictionType.image_orientation,

robotoff/prediction/ocr/core.py

@@ -20,6 +20,7 @@
 from .label import find_labels
 from .location import find_locations
 from .nutrient import find_nutrient_mentions, find_nutrient_values
+from .origin import origin_parser
 from .packager_code import find_packager_codes
 from .packaging import find_packaging
 from .product_weight import find_product_weight
@@ -41,6 +42,7 @@
     PredictionType.trace: find_traces,
     PredictionType.nutrient: find_nutrient_values,
     PredictionType.nutrient_mention: find_nutrient_mentions,
+    PredictionType.origin: origin_parser.find_origin,
     PredictionType.brand: find_brands,
     PredictionType.store: find_stores,
     PredictionType.packaging: find_packaging,