Implementing a data generator to create batch of images for feature extraction in HIstomicsML2

FeatureExtraction_batch.xml

<?xml version="1.0" encoding="UTF-8"?>
<executable>
  <category>HistomicsTK</category>
  <title>Create HistomicsML Dataset</title>
  <description>Create HistomicsML files</description>
  <version>0.1.0</version>
  <documentation-url>https://digitalslidearchive.github.io/HistomicsTK/</documentation-url>
  <license>Apache 2.0</license>
  <contributor>Deepak Roy Chittajallu (Kitware), Sanghoon Lee (Emory University)</contributor>
  <acknowledgements>This work is part of the HistomicsTK project.</acknowledgements>
  <parameters>
    <label>IO</label>
    <description>Input/output parameters</description>
    <string>
      <name>inputPCAModel</name>
      <label>Input PCA fitted model</label>
      <description>Input PCA Model</description>
      <longflag>inputPCAModel</longflag>
    </string>
    <string>
      <name>projectName</name>
      <label>project Name</label>
      <description>project Name</description>
      <longflag>projectName</longflag>
      <default>/example</default>
    </string>
  </parameters>
  <parameters advanced="true">
    <label>Color Normalization</label>
    <description>Color Normalization parameters</description>
    <double-vector>
      <name>reference_mu_lab</name>
      <label>Reference Mean LAB</label>
      <description>Mean of reference image in LAB color space for Reinhard color normalization</description>
      <longflag>reference_mu_lab</longflag>
      <default>8.63234435, -0.11501964, 0.03868433</default>
    </double-vector>
    <double-vector>
      <name>reference_std_lab</name>
      <label>Reference Stddev LAB</label>
      <description>Standard deviation of reference image in LAB color space for Reinhard color normalization</description>
      <longflag>reference_std_lab</longflag>
      <default>0.57506023, 0.10403329, 0.01364062</default>
    </double-vector>
  </parameters>
  <parameters advanced="true">
    <label>WSI Analysis</label>
    <description>Whole-slide image analysis (WSI) parameters</description>
    <double>
      <name>max_tile_size</name>
      <label>Max Analysis Tile Size</label>
      <description>Tile size for blockwise analysis</description>
      <longflag>max_tile_size</longflag>
      <default>16384</default>
    </double>
    <double>
      <name>max_mag</name>
      <label>Max Analysis Magnification</label>
      <description>The magnification at which the analysis should be performed.</description>
      <longflag>max_mag</longflag>
      <default>40</default>
    </double>
    <double>
      <name>min_fgnd_frac</name>
      <label>Minimum foreground fraction</label>
      <description>The minimum amount of foreground that must be present in a tile for it to be analyzed</description>
      <longflag>min_fgnd_frac</longflag>
      <constraints>
      	<minimum>0</minimum>
      	<maximum>1</maximum>
      </constraints>
      <default>0.001</default>
    </double>
    <double>
      <name>sample_fraction</name>
      <label>Fraction of pixels to sample</label>
      <description>Fraction of pixels to sample for normalization</description>
      <longflag>sample_fraction</longflag>
      <constraints>
      	<minimum>0</minimum>
      	<maximum>1</maximum>
      </constraints>
      <default>0.1</default>
    </double>
  </parameters>
<parameters advanced="true">
    <label>SuperpixelParameters</label>
    <description>Superpixel parameters</description>
    <integer>
      <name>superpixelSize</name>
      <longflag>superpixelSize</longflag>
      <label>Superpixel size</label>
      <description>Patch size for superpixel region</description>
      <constraints>
      	<minimum>8</minimum>
      	<maximum>256</maximum>
      </constraints>
      <default>64</default>
    </integer>
    <integer>
      <name>patchSize</name>
      <longflag>patchSize</longflag>
      <label>Patch size</label>
      <description>Patch size for superpixel region</description>
      <constraints>
      	<minimum>8</minimum>
      	<maximum>512</maximum>
      </constraints>
      <default>128</default>
    </integer>
    <integer>
      <name>patchSizeResized</name>
      <longflag>patchSizeResized</longflag>
      <label>Resized patch size</label>
      <description>patchSize resized for superpixel region</description>
      <default>224</default>
    </integer>
    <integer>
      <name>pca_dim</name>
      <longflag>pca_dim</longflag>
      <label>PCA dimension</label>
      <description>PCA dimension</description>
      <default>64</default>
    </integer>
    <integer>
      <name>fcn</name>
      <longflag>fcn</longflag>
      <label>fully connected network</label>
      <description>size of fully connected network</description>
      <default>4096</default>
    </integer>
    <double>
      <longflag>pca_sample_scale</longflag>
      <label>pca_sample_scale</label>
      <description>PCA sampling scale</description>
      <default>0.1</default>
    </double>
    <integer>
      <name>compactness</name>
      <longflag>compactness</longflag>
      <label>Compactness</label>
      <description>Compactness of SLIC algorithm</description>
      <constraints>
      	<minimum>0.01</minimum>
      	<maximum>100</maximum>
      </constraints>
      <default>50</default>
    </integer>
    <integer>
      <name>min_fgnd_superpixel</name>
      <longflag>min_fgnd_superpixel</longflag>
      <label>Minimum foreground pixels in a superpixel</label>
      <description>Minimum number of foreground pixels in a superpixel </description>
      <default>10</default>
    </integer>
    <double>
      <longflag>rg_ratio_superpixel</longflag>
      <label>Red and green ratio for superpixel</label>
      <description>Red and green ratio for superpixel</description>
      <default>1.2</default>
    </double>
    <double>
      <longflag>min_var_superpixel</longflag>
      <label>Minumum variance of superpixel</label>
      <description>Minumum variance of superpixel</description>
      <constraints>
      	<minimum>0</minimum>
      	<maximum>1</maximum>
      </constraints>
      <default>0.0015</default>
    </double>
  </parameters>
  <parameters advanced="true">
    <label>HistomicsMLParameters</label>
    <description>HistomicsML parameters</description>
    <integer>
      <name>columnSize</name>
      <longflag>columnSize</longflag>
      <label>Column size</label>
      <description>HistomicsML dataset column size</description>
      <default>1</default>
    </integer>
    <integer>
      <name>channelSize</name>
      <longflag>channelSize</longflag>
      <label>Channel size</label>
      <description>HistomicsML color channel size</description>
      <default>3</default>
    </integer>
  </parameters>
  <parameters advanced="true">
    <label>Dask</label>
    <description>Dask parameters</description>
    <string>
      <name>scheduler</name>
      <label>Scheduler Address</label>
      <description>Address of a dask scheduler in the format '127.0.0.1:8786'.  Not passing this parameter sets up a dask cluster on the local machine.  'multiprocessing' uses Python multiprocessing.  'multithreading' uses Python multiprocessing in threaded mode.</description>
      <longflag>scheduler</longflag>
      <default></default>
    </string>
    <integer>
      <name>num_workers</name>
      <label>Number of workers</label>
      <description>Number of dask workers to start while setting up a local cluster internally. If a negative value is specified then the number of workers is set to number of cpu cores on the machine minus the number of workers specified.</description>
      <longflag>num_workers</longflag>
      <default>-1</default>
    </integer>
    <integer>
      <name>num_threads_per_worker</name>
      <label>Number of threads per worker</label>
      <description>Number of threads to use per worker while setting up a local cluster internally. Must be a positive integer >= 1.</description>
      <longflag>num_threads_per_worker</longflag>
      <default>1</default>
    </integer>
  </parameters>
  <parameters advanced="true">
    <label>GPU</label>
    <description>GPU parameters</description>
    <string>
      <name>gpus</name>
      <label>gpus</label>
      <description>gpu id</description>
      <longflag>gpus</longflag>
      <default>0</default>
    </string>
    <integer>
      <name>batch_size</name>
      <label>batch_size</label>
      <description>batch_size</description>
      <longflag>batch_size</longflag>
      <default>32</default>
    </integer>
  </parameters>
</executable>