@Test
public void testPhraseLabelsExactMatch() {
createDocuments(
"data is cool", "data is cool", "data is cool", "data is cool", "data cool", "data cool");
final int[][] expectedDocumentIndices = new int[][] {new int[] {0, 1}};
documentAssigner.exactPhraseAssignment = true;
documentAssigner.minClusterSize = 2;
check(expectedDocumentIndices, 0);
}
@Test
public void testStemmedSingleWordLabelConflation() {
createDocuments("cat", "cat", "cat", "cat", "cats", "cats", "cats", "cats");
final int[][] expectedDocumentIndices = new int[][] {new int[] {0, 1, 2, 3}};
documentAssigner.minClusterSize = 1;
check(expectedDocumentIndices, -1);
}
@Test
public void testSingleWordLabels() {
createDocuments("coal is", "coal is", "mining", "mining");
final int[][] expectedDocumentIndices = new int[][] {new int[] {0}, new int[] {1}};
documentAssigner.minClusterSize = 1;
check(expectedDocumentIndices, -1);
}
@Test
public void testMinClusterSize() {
createDocuments("test coal", "test coal", "coal test . mining", "coal test . mining");
final int[][] expectedDocumentIndices =
new int[][] {
new int[] {0, 1},
new int[] {0, 1},
new int[] {0, 1},
new int[] {0, 1}
};
documentAssigner.minClusterSize = 2;
check(expectedDocumentIndices, 2);
}
private void check(int[][] expectedDocumentIndices, int expectedFirstPhraseIndex) {
runPreprocessing();
documentAssigner.assign(context);
assertThat(context.allLabels.firstPhraseIndex)
.as("allLabels.firstPhraseIndex")
.isEqualTo(expectedFirstPhraseIndex);
assertThat(context.allLabels.documentIndices)
.as("allLabels.documentIndices")
.hasSize(expectedDocumentIndices.length);
for (int i = 0; i < expectedDocumentIndices.length; i++) {
assertThat(context.allLabels.documentIndices[i].asIntLookupContainer().toArray())
.as("allLabels.documentIndices[" + i + "]")
.isEqualTo(expectedDocumentIndices[i]);
}
}
@Test
public void testStemmedPhraseLabelConflation() {
createDocuments(
"cat horse",
"cat horse",
"cats horse",
"cats horse",
"cat horses",
"cat horses",
"cats horses",
"cats horses");
final int[][] expectedDocumentIndices =
new int[][] {
new int[] {0, 1, 2, 3},
new int[] {0, 1, 2, 3},
new int[] {0, 1, 2, 3}
};
documentAssigner.minClusterSize = 1;
check(expectedDocumentIndices, 2);
}
