fix: remove pdo and pgi regex

One ingredient of the full product can be pdo or cgi, it's not a good proxy
openfoodfacts · Aug 28, 2023 · 008bd9d · 008bd9d
1 parent 1e36c85
commit 008bd9d
Show file tree

Hide file tree

Showing 2 changed files with 0 additions and 25 deletions.
diff --git a/data/ocr/label_flashtext.txt b/data/ocr/label_flashtext.txt
@@ -118,8 +118,6 @@ en:vegan||vegan
 en:vegan||veganoveganistisch
 en:vegan||vegetalien
 en:vegan||végétalien
-en:pdo||appellation d'origine controlee
-en:pdo||appellation d'origine contrôlée
 en:french-beef||viande bovine francaise
 en:french-beef||viande bovine française
 en:french-pork||le porc francais

diff --git a/robotoff/prediction/ocr/label.py b/robotoff/prediction/ocr/label.py
@@ -72,19 +72,6 @@ def process_es_bio_label_code(match) -> str:
             field=OCRField.full_text_contiguous,
         ),
     ],
-    "en:pgi": [
-        OCRRegex(
-            re.compile(
-                r"indication g[ée]ographique prot[eé]g[eé]e|Indicazione geografica protetta|geschützte geografische angabe",
-                re.I,
-            ),
-            field=OCRField.full_text_contiguous,
-        ),
-        OCRRegex(
-            re.compile(r"(?<!\w)(?:IGP|BGA|PGI)(?!\w)"),
-            field=OCRField.full_text_contiguous,
-        ),
-    ],
     "fr:label-rouge": [
         OCRRegex(
             re.compile(r"d[ée]cret du 0?5[./]01[./]07", re.I),
@@ -95,16 +82,6 @@ def process_es_bio_label_code(match) -> str:
             field=OCRField.full_text_contiguous,
         ),
     ],
-    "en:pdo": [
-        OCRRegex(
-            re.compile(r"(?<!\w)(?:PDO|AOP|DOP)(?!\w)"),
-            field=OCRField.full_text_contiguous,
-        ),
-        OCRRegex(
-            re.compile(r"appellation d'origine prot[eé]g[eé]e", re.I),
-            field=OCRField.full_text_contiguous,
-        ),
-    ],
     "fr:aoc": [
         OCRRegex(
             re.compile(r"(?<!\w)(?:AOC)(?!\w)"), field=OCRField.full_text_contiguous