/**
* Factory method to create a SetList using the supplied list to retain order.
*
* <p>If the list contains duplicates, these are removed (first indexed one kept). A <code>HashSet
* </code> is used for the set behaviour.
*
* @param <E> the element type
* @param list the list to decorate, must not be null
* @return a new {@link SetUniqueList}
* @throws NullPointerException if list is null
* @since 4.0
*/
public static <E> SetUniqueList<E> setUniqueList(final List<E> list) {
if (list == null) {
throw new NullPointerException("List must not be null");
}
if (list.isEmpty()) {
return new SetUniqueList<E>(list, new HashSet<E>());
}
final List<E> temp = new ArrayList<E>(list);
list.clear();
final SetUniqueList<E> sl = new SetUniqueList<E>(list, new HashSet<E>());
sl.addAll(temp);
return sl;
}
/**
* Project all of the normals used to create the PoN into the reduced panel.
*
* @param pon Not {@code null}.
* @param ctx Spark context, {@code null} indicates no spark context available, which is
* supported.
* @return never {@code null}. Result will contain multiple columns in each of the
* ReadCountCollection attributes.
*/
public static TangentNormalizationResult tangentNormalizeNormalsInPoN(
final PoN pon, final JavaSparkContext ctx) {
// Get the list of sample names in a modifiable List
final List<String> sampleNamesCopy = new ArrayList<>(pon.getSampleNames());
logger.info("Loading normalized counts...");
final ReadCountCollection coverageProfile =
new ReadCountCollection(
SetUniqueList.setUniqueList(pon.getTargets()),
SetUniqueList.setUniqueList(sampleNamesCopy),
pon.getNormalizedCounts());
logger.info("Tangent normalizing the normals (normalized by target factors) ...");
// For each sample in the PoN, tangent normalize against the qc reduced PoN.
return TangentNormalizer.tangentNormalize(pon, coverageProfile, ctx);
}