Add adjust_hue frontend function to paddle.vision.transforms (ivy-llc…

…#21540)

ivy/functional/frontends/paddle/vision/transforms.py

@@ -18,6 +18,108 @@ def to_tensor(pic, data_format="CHW"):
     return Tensor(array)
 
 
+# helpers
+def _get_image_c_axis(data_format):
+    if data_format.lower() == "chw":
+        return -3
+    elif data_format.lower() == "hwc":
+        return -1
+
+
+def _get_image_num_channels(img, data_format):
+    return ivy.shape(img)[_get_image_c_axis(data_format)]
+
+
+def _rgb_to_hsv(img):
+    maxc = ivy.max(img, axis=-3)
+    minc = ivy.min(img, axis=-3)
+
+    is_equal = ivy.equal(maxc, minc)
+    one_divisor = ivy.ones_like(maxc)
+    c_delta = maxc - minc
+    s = c_delta / ivy.where(is_equal, one_divisor, maxc)
+
+    r, g, b = img[0], img[1], img[2]
+    c_delta_divisor = ivy.where(is_equal, one_divisor, c_delta)
+
+    rc = (maxc - r) / c_delta_divisor
+    gc = (maxc - g) / c_delta_divisor
+    bc = (maxc - b) / c_delta_divisor
+
+    hr = ivy.where((maxc == r), bc - gc, ivy.zeros_like(maxc))
+    hg = ivy.where(
+        ((maxc == g) & (maxc != r)),
+        rc - bc + 2.0,
+        ivy.zeros_like(maxc),
+    )
+    hb = ivy.where(
+        ((maxc != r) & (maxc != g)),
+        gc - rc + 4.0,
+        ivy.zeros_like(maxc),
+    )
+
+    h = (hr + hg + hb) / 6.0 + 1.0
+    h = h - ivy.trunc(h)
+
+    return ivy.stack([h, s, maxc], axis=-3)
+
+
+def _hsv_to_rgb(img):
+    h, s, v = img[0], img[1], img[2]
+    f = h * 6.0
+    i = ivy.floor(f)
+    f = f - i
+    i = ivy.astype(i, ivy.int32) % 6
+
+    p = ivy.clip(v * (1.0 - s), 0.0, 1.0)
+    q = ivy.clip(v * (1.0 - s * f), 0.0, 1.0)
+    t = ivy.clip(v * (1.0 - s * (1.0 - f)), 0.0, 1.0)
+
+    mask = ivy.astype(
+        ivy.equal(
+            ivy.expand_dims(i, axis=-3),
+            ivy.reshape(ivy.arange(6, dtype=ivy.dtype(i)), (-1, 1, 1)),
+        ),
+        ivy.dtype(img),
+    )
+    matrix = ivy.stack(
+        [
+            ivy.stack([v, q, p, p, t, v], axis=-3),
+            ivy.stack([t, v, v, q, p, p], axis=-3),
+            ivy.stack([p, p, t, v, v, q], axis=-3),
+        ],
+        axis=-4,
+    )
+    return ivy.einsum("...ijk, ...xijk -> ...xjk", mask, matrix)
+
+
+@with_supported_dtypes({"2.5.1 and below": ("float32", "float64", "uint8")}, "paddle")
+@to_ivy_arrays_and_back
+def adjust_hue(img, hue_factor):
+    assert -0.5 <= hue_factor <= 0.5, "hue_factor should be in range [-0.5, 0.5]"
+
+    channels = _get_image_num_channels(img, "CHW")
+
+    if channels == 1:
+        return img
+    elif channels == 3:
+        if ivy.dtype(img) == "uint8":
+            img = ivy.astype(img, "float32") / 255.0
+
+        img_hsv = _rgb_to_hsv(img)
+        h, s, v = img_hsv[0], img_hsv[1], img_hsv[2]
+
+        h = h + hue_factor
+        h = h - ivy.floor(h)
+
+        img_adjusted = _hsv_to_rgb(ivy.stack([h, s, v], axis=-3))
+
+    else:
+        raise ValueError("channels of input should be either 1 or 3.")
+
+    return img_adjusted
+
+
 @with_unsupported_device_and_dtypes(
     {
         "2.5.1 and below": {

ivy_tests/test_ivy/test_frontends/test_paddle/test_vision/test_transforms.py

@@ -31,6 +31,45 @@ def test_paddle_to_tensor(
     )
 
 
+# adjust_hue
+@handle_frontend_test(
+    fn_tree="paddle.vision.transforms.adjust_hue",
+    dtype_and_x=helpers.dtype_and_values(
+        available_dtypes=helpers.get_dtypes("valid"),
+        min_value=0,
+        max_value=255,
+        min_num_dims=3,
+        max_num_dims=3,
+        min_dim_size=3,
+        max_dim_size=3,
+    ),
+    hue_factor=helpers.floats(min_value=-0.5, max_value=0.5),
+)
+def test_paddle_adjust_hue(
+    *,
+    dtype_and_x,
+    hue_factor,
+    on_device,
+    fn_tree,
+    frontend,
+    test_flags,
+    backend_fw,
+):
+    input_dtype, x = dtype_and_x
+    helpers.test_frontend_function(
+        input_dtypes=input_dtype,
+        backend_to_test=backend_fw,
+        frontend=frontend,
+        test_flags=test_flags,
+        fn_tree=fn_tree,
+        rtol=1e-3,
+        atol=1e-3,
+        on_device=on_device,
+        img=x[0],
+        hue_factor=hue_factor,
+    )
+
+
 @handle_frontend_test(
     fn_tree="paddle.vision.transforms.vflip",
     dtype_and_x=helpers.dtype_and_values(