use of com.ibm.cohort.cql.util.EqualsStringMatcher in project quality-measure-and-cohort-service by Alvearie.
the class ColumnRuleCreator method getDataRequirementsForContext.
/**
* Retrieve the merged set of data type and column filters for all CQL jobs that will
* be evaluated for a given aggregation context.
*
* @param context ContextDefinition whose CQL jobs will be interrogated for data requirements
* @return Map of data type to the fields in that datatype that are used by the CQL jobs
*/
public Map<String, Set<StringMatcher>> getDataRequirementsForContext(ContextDefinition context) {
Map<CqlLibraryDescriptor, Set<String>> expressionsByLibrary = new HashMap<>();
for (CqlEvaluationRequest request : requests) {
Set<String> expressions = expressionsByLibrary.computeIfAbsent(request.getDescriptor(), desc -> new HashSet<>());
request.getExpressions().stream().forEach(exp -> expressions.add(exp.getName()));
}
DataTypeRequirementsProcessor requirementsProcessor = new DataTypeRequirementsProcessor(cqlTranslator);
Map<String, Set<StringMatcher>> pathsByDataType = new HashMap<>();
for (Map.Entry<CqlLibraryDescriptor, Set<String>> entry : expressionsByLibrary.entrySet()) {
LOG.debug("Extracting data requirements for {}", entry.getKey());
DataTypeRequirementsProcessor.DataTypeRequirements requirements = requirementsProcessor.getDataRequirements(libraryProvider, entry.getKey(), entry.getValue());
Map<String, Set<StringMatcher>> newPaths = requirements.allAsStringMatcher();
newPaths.forEach((key, value) -> {
pathsByDataType.merge(key, value, (prev, current) -> {
prev.addAll(current);
return prev;
});
});
}
Set<StringMatcher> contextFields = pathsByDataType.computeIfAbsent(context.getPrimaryDataType(), dt -> new HashSet<>());
contextFields.add(new EqualsStringMatcher(context.getPrimaryKeyColumn()));
if (context.getRelationships() != null) {
for (Join join : context.getRelationships()) {
Set<StringMatcher> joinFields = pathsByDataType.get(join.getRelatedDataType());
if (joinFields != null) {
joinFields.add(new EqualsStringMatcher(join.getRelatedKeyColumn()));
joinFields.add(new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX));
// if the join key is not the primary key of the primary data table, then we need to add in the alternate key
if (join.getPrimaryDataTypeColumn() != null) {
contextFields.add(new EqualsStringMatcher(join.getPrimaryDataTypeColumn()));
}
if (join instanceof ManyToMany) {
ManyToMany manyToMany = (ManyToMany) join;
Set<StringMatcher> associationFields = pathsByDataType.computeIfAbsent(manyToMany.getAssociationDataType(), dt -> new HashSet<>());
associationFields.add(new EqualsStringMatcher(manyToMany.getAssociationOneKeyColumn()));
associationFields.add(new EqualsStringMatcher(manyToMany.getAssociationManyKeyColumn()));
}
if (join instanceof MultiManyToMany) {
ManyToMany with = ((MultiManyToMany) join).getWith();
while (with != null) {
Set<StringMatcher> relatedFields = pathsByDataType.computeIfAbsent(with.getRelatedDataType(), dt -> new HashSet<>());
relatedFields.add(new EqualsStringMatcher(with.getRelatedKeyColumn()));
relatedFields.add(new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX));
with = (with instanceof MultiManyToMany) ? ((MultiManyToMany) with).getWith() : null;
}
}
}
}
}
pathsByDataType.values().forEach((matcherSet -> {
matcherSet.add(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX));
}));
return pathsByDataType;
}
use of com.ibm.cohort.cql.util.EqualsStringMatcher in project quality-measure-and-cohort-service by Alvearie.
the class ColumnRuleCreatorTest method testGetFiltersForContextOnlyJoinColumns.
@Test
public void testGetFiltersForContextOnlyJoinColumns() throws Exception {
CqlToElmTranslator cqlTranslator = new CqlToElmTranslator();
cqlTranslator.registerModelInfo(new File("src/test/resources/alltypes/modelinfo/alltypes-modelinfo-1.0.0.xml"));
ObjectMapper mapper = new ObjectMapper();
CqlEvaluationRequests requests = mapper.readValue(new File("src/test/resources/alltypes/metadata/join-only.json"), CqlEvaluationRequests.class);
CqlLibraryProvider backingProvider = new ClasspathCqlLibraryProvider("alltypes.cql");
TranslatingCqlLibraryProvider cqlLibraryProvider = new TranslatingCqlLibraryProvider(backingProvider, cqlTranslator);
ColumnRuleCreator columnRuleCreator = new ColumnRuleCreator(requests.getEvaluations(), cqlTranslator, cqlLibraryProvider);
ContextDefinitions definitions = mapper.readValue(new File("src/test/resources/alltypes/metadata/context-definitions-related-column.json"), ContextDefinitions.class);
ContextDefinition context = definitions.getContextDefinitionByName("Patient");
Map<String, Set<StringMatcher>> actual = columnRuleCreator.getDataRequirementsForContext(context);
Map<String, Set<StringMatcher>> expected = new HashMap<>();
expected.put("A", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("id_col"), new EqualsStringMatcher("pat_id"))));
expected.put("B", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("string"), new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX))));
expected.put("C", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("pat_id"), new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX))));
assertEquals(expected, actual);
}
use of com.ibm.cohort.cql.util.EqualsStringMatcher in project quality-measure-and-cohort-service by Alvearie.
the class ColumnRuleCreatorTest method testGetFiltersForContext.
@Test
public void testGetFiltersForContext() throws Exception {
CqlToElmTranslator cqlTranslator = new CqlToElmTranslator();
cqlTranslator.registerModelInfo(new File("src/test/resources/alltypes/modelinfo/alltypes-modelinfo-1.0.0.xml"));
ObjectMapper mapper = new ObjectMapper();
CqlEvaluationRequests requests = mapper.readValue(new File("src/test/resources/alltypes/metadata/parent-child-jobs.json"), CqlEvaluationRequests.class);
CqlLibraryProvider backingProvider = new ClasspathCqlLibraryProvider("alltypes.cql");
TranslatingCqlLibraryProvider cqlLibraryProvider = new TranslatingCqlLibraryProvider(backingProvider, cqlTranslator);
ColumnRuleCreator columnRuleCreator = new ColumnRuleCreator(requests.getEvaluations(), cqlTranslator, cqlLibraryProvider);
ContextDefinitions definitions = mapper.readValue(new File("src/test/resources/alltypes/metadata/context-definitions.json"), ContextDefinitions.class);
ContextDefinition context = definitions.getContextDefinitionByName("Patient");
Map<String, Set<StringMatcher>> actual = columnRuleCreator.getDataRequirementsForContext(context);
Map<String, Set<StringMatcher>> expected = new HashMap<>();
expected.put("A", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("pat_id"), new EqualsStringMatcher("code_col"), new EqualsStringMatcher("boolean_col"))));
assertEquals(expected, actual);
}
use of com.ibm.cohort.cql.util.EqualsStringMatcher in project quality-measure-and-cohort-service by Alvearie.
the class ColumnRuleCreatorTest method testGetFiltersForContextWithMultiJoinColumns.
@Test
public void testGetFiltersForContextWithMultiJoinColumns() throws Exception {
CqlToElmTranslator cqlTranslator = new CqlToElmTranslator();
cqlTranslator.registerModelInfo(new File("src/test/resources/multiple-joins/modelinfo/omop-modelinfo-5.2.2.xml"));
ObjectMapper mapper = new ObjectMapper();
CqlEvaluationRequests requests = mapper.readValue(new File("src/test/resources/multiple-joins/metadata/cql-jobs.json"), CqlEvaluationRequests.class);
TranslatingCqlLibraryProvider cqlLibraryProvider = new TranslatingCqlLibraryProvider(new ClasspathCqlLibraryProvider("multiple-joins.cql"), cqlTranslator);
ColumnRuleCreator columnRuleCreator = new ColumnRuleCreator(requests.getEvaluations(), cqlTranslator, cqlLibraryProvider);
ContextDefinitions definitions = mapper.readValue(new File("src/test/resources/multiple-joins/metadata/context-definitions.json"), ContextDefinitions.class);
ContextDefinition context = definitions.getContextDefinitionByName("person");
Map<String, Set<StringMatcher>> actual = columnRuleCreator.getDataRequirementsForContext(context);
Map<String, Set<StringMatcher>> expected = new HashMap<>();
expected.put("person", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("person_id"))));
expected.put("vocabulary", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("vocabulary_id"), new EqualsStringMatcher("vocabulary_version"), new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX))));
expected.put("concept", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("concept_id"), new EqualsStringMatcher("concept_code"), new EqualsStringMatcher("concept_name"), new EqualsStringMatcher("vocabulary_id"), new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX))));
expected.put("observation", new HashSet<>(Arrays.asList(new EqualsStringMatcher(ContextRetriever.SOURCE_FACT_IDX), new EqualsStringMatcher("observation_concept_id"), new EqualsStringMatcher("person_id"), new EqualsStringMatcher(ContextRetriever.JOIN_CONTEXT_VALUE_IDX))));
assertEquals(expected, actual);
}
use of com.ibm.cohort.cql.util.EqualsStringMatcher in project quality-measure-and-cohort-service by Alvearie.
the class ColumnFilterFunctionTest method testColumnFiltering.
@Test
public void testColumnFiltering() {
String path = new File("src/test/resources/alltypes/testdata/test-A.parquet").toURI().toString();
Dataset<Row> baseline = spark.read().format("parquet").load(path);
assertEquals(12, baseline.schema().fields().length);
String colName = "boolean_col";
String regexColName = "code_col[0-9]*";
ColumnFilterFunction datasetTransformer = new ColumnFilterFunction(new HashSet<>(Arrays.asList(new EqualsStringMatcher(colName), new RegexStringMatcher(regexColName))));
Dataset<Row> filtered = datasetTransformer.apply(baseline);
Set<String> expectedNames = new HashSet<>(Arrays.asList(colName, "code_col", "code_col_system", "string_col", "code_col2"));
Set<String> actualNames = new HashSet<>(Arrays.asList(filtered.schema().fieldNames()));
assertEquals(expectedNames, actualNames);
}
Aggregations