Fix resizing

dan/datasets/extract/extract.py

@@ -6,7 +6,6 @@ import pickle
 import random
 from collections import defaultdict
 from concurrent.futures import Future, ThreadPoolExecutor
-from functools import cached_property
 from pathlib import Path
 from typing import Dict, List, Optional, Union
 from uuid import UUID
@@ -91,23 +90,26 @@ class ArkindexExtractor:
         # Image download tasks to process
         self.tasks: List[Dict[str, str]] = []
 
-    @cached_property
-    def max_resize(self):
-        # We keep the aspect ratio so we only use on dimension, the biggest one
-        if self.max_width > self.max_height:
-            return f"{self.max_width},"
-        return f",{self.max_height}"
-
     def get_iiif_size_arg(self, width: int, height: int) -> str:
         if (self.max_width is None or width <= self.max_width) and (
             self.max_height is None or height <= self.max_height
         ):
             return IIIF_FULL_SIZE
 
-        # Resizing if the image is bigger than the wanted size.
-        if self.max_width and self.max_height:
-            return self.max_resize
-        return f"{self.max_width or ''},{self.max_height or ''}"
+        bigger_width = self.max_width and width >= self.max_width
+        bigger_height = self.max_height and height >= self.max_height
+
+        if bigger_width and bigger_height:
+            # Resize to the biggest dim to keep aspect ratio
+            # Only resize width is bigger than max size
+            # This ratio tells which dim needs the biggest shrinking
+            ratio = width * self.max_height / (height * self.max_width)
+            return f"{self.max_width}," if ratio > 1 else f",{self.max_height}"
+        elif bigger_width:
+            return f"{self.max_width},"
+        # Only resize height is bigger than max size
+        elif bigger_height:
+            return f",{self.max_height}"
 
     def build_iiif_url(self, polygon, image_url):
         bbox = polygon_to_bbox(json.loads(str(polygon)))

tests/test_extract.py

@@ -12,7 +12,7 @@ import pytest
 from PIL import Image, ImageChops
 
 from dan.datasets.extract.exceptions import NoEndTokenError
-from dan.datasets.extract.extract import ArkindexExtractor
+from dan.datasets.extract.extract import IIIF_FULL_SIZE, ArkindexExtractor
 from dan.datasets.extract.utils import EntityType, insert_token, remove_spaces
 from dan.utils import parse_tokens
 from tests import FIXTURES
@@ -36,6 +36,24 @@ def filter_tokens(keys):
     return {key: value for key, value in TOKENS.items() if key in keys}
 
 
+@pytest.mark.parametrize(
+    "max_width, max_height, width, height, resize",
+    (
+        (1000, 2000, 900, 800, IIIF_FULL_SIZE),
+        (1000, 2000, 1100, 800, "1000,"),
+        (1000, 2000, 1100, 2800, ",2000"),
+        (1000, 2000, 2000, 3000, "1000,"),
+    ),
+)
+def test_get_iiif_size_arg(max_width, max_height, width, height, resize):
+    assert (
+        ArkindexExtractor(max_width=max_width, max_height=max_height).get_iiif_size_arg(
+            width=width, height=height
+        )
+        == resize
+    )
+
+
 @pytest.mark.parametrize(
     "text,offset,length,expected",
     (