syaffers

<configuration>
    <property>
        <name>mapreduce.framework.name</name>
        <value>yarn</value>
    </property>
    <property>
        <name>yarn.app.mapreduce.am.env</name>
        <value>HADOOP_MAPRED_HOME=/opt/hadoop-3.2.2</value>
    </property>
    <property>

syaffers

<configuration>
    <property>
        <name>yarn.resourcemanager.hostname</name>
        <value>your_hostname</value>
    </property>
    <property>
        <name>yarn.nodemanager.aux-services</name>
        <value>mapreduce_shuffle</value>
    </property>
    <property>

syaffers

class EntirePalette:
    # ... omitted

    def recolor(self, image):
        image_shape = image.shape
        image = self._preprocess(image)
        image_colors = np.unique(image, axis=0)

        self_da = da.from_array(
            self.source_pixels.astype(np.long), chunks=(self.chunk_size, 3)

syaffers

typedef struct {
   unsigned char r, g, b;
} RGB;

double ColourDistance(RGB e1, RGB e2)
{
  long rmean = ( (long)e1.r + (long)e2.r ) / 2;
  long r = (long)e1.r - (long)e2.r;
  long g = (long)e1.g - (long)e2.g;
  long b = (long)e1.b - (long)e2.b;

syaffers

src = np.array(imread('carles-rabada-DD1fSz2HF1s-unsplash-small.png'))[:, :, :3]
trn = np.array(imread('olia-gozha-9A_peGrSbZc-unsplash-small.png'))[:, :, :3]
src = np.unique(src.reshape(-1, 3), axis=0)
trn = np.unique(trn.reshape(-1, 3), axis=0)
self_da = da.from_array(src.astype(np.long), chunks=(1024, 3))
other_da = da.from_array(trn.reshape(-1, 1, 3).astype(np.long),
                         chunks=(1024, 1, 3))
rgb_da = self_da - other_da
display(self_da)
display(other_da)

syaffers

class EntirePalette:
    def __init__(self, chunk_size=1024):
        self.chunk_size = chunk_size
        self.source_pixels = None

    def _preprocess(self, image):
        # ... omitted, same as KMeansReducedPalette

    def fit(self, image):
        self.source_pixels = np.unique(self._preprocess(image), axis=0)

syaffers

class KMeansReducedPalette:
    # ... omitted

    def random_neighborhood_walk_recolor(self, image, max_steps):
        original_shape = image.shape
        image = self._preprocess(image)
        recolor = image.copy()
        centroid_idxs = self.kmeans.predict(image)

        for ci in range(self.num_colors):

syaffers

class KMeansReducedPalette:
    # ... omitted

    def fit(self, image):
        image_cpy = image.copy()
        self.source_pixels = self._preprocess(image_cpy)
        self.kmeans.fit(self.source_pixels)

        self.centroid_nearest_pixels = []

syaffers

class KMeansReducedPalette:
    # ... omitted

    def random_walk_recolor(self, image, max_steps):
        original_shape = image.shape
        image = self._preprocess(image)
        centroids = self.kmeans.predict(image)
        start = np.round(self.kmeans.cluster_centers_[centroids])

        diff = np.zeros(image.shape)

syaffers

class UniqueKMeansReducedPalette(KMeansReducedPalette):
    def __init__(self, num_colors):
        super().__init__(num_colors)

    def fit(self, image):
        image_cpy = image.copy()
        pixels = self._preprocess(image_cpy)
        super().fit(np.unique(pixels, axis=0))