Search in sources :

Example 1 with Attribute

use of datawave.query.attributes.Attribute in project datawave by NationalSecurityAgency.

the class LimitFields method apply.

@Override
public Entry<Key, Document> apply(Entry<Key, Document> entry) {
    // key is the limited field name with _ORIGINAL_COUNT appended,
    // value will be set to the original count of that field in the document
    Map<String, Integer> limitedFieldCounts = new HashMap<>();
    Document document = entry.getValue();
    Map<String, String> hitTermMap = this.getHitTermMap(document);
    Multimap<String, Attribute<? extends Comparable<?>>> reducedMap = LinkedListMultimap.create();
    Map<String, Integer> countForFieldMap = Maps.newHashMap();
    // maps from the key with NO grouping context to a multimap of
    // key WITH grouping context to attributes:
    // DIRECTION : [DIRECTION.1 : [over,under], DIRECTION.2 : [sideways,down]]
    LoadingCache<String, Multimap<String, Attribute<? extends Comparable<?>>>> hits = CacheBuilder.newBuilder().build(new CacheLoader<String, Multimap<String, Attribute<? extends Comparable<?>>>>() {

        public Multimap<String, Attribute<? extends Comparable<?>>> load(String key) {
            return LinkedListMultimap.create();
        }
    });
    // maps from the key with NO grouping context to a multimap of
    // key WITH grouping context to attributes:
    // DIRECTION : [DIRECTION.1 : [over,under], DIRECTION.2 : [sideways,down]]
    @SuppressWarnings("serial") LoadingCache<String, Multimap<String, Attribute<? extends Comparable<?>>>> misses = CacheBuilder.newBuilder().build(new CacheLoader<String, Multimap<String, Attribute<? extends Comparable<?>>>>() {

        public Multimap<String, Attribute<? extends Comparable<?>>> load(String key) {
            return LinkedListMultimap.create();
        }
    });
    for (Map.Entry<String, Attribute<? extends Comparable<?>>> de : document.entrySet()) {
        String keyWithGrouping = de.getKey();
        String keyNoGrouping = keyWithGrouping;
        // if we have grouping context on, remove the grouping context
        if (keyNoGrouping.indexOf('.') != -1) {
            keyNoGrouping = keyNoGrouping.substring(0, keyNoGrouping.indexOf('.'));
        }
        // limit value for _ANYFIELD_
        if (this.limitFieldsMap.containsKey("_ANYFIELD_") && this.limitFieldsMap.containsKey(keyNoGrouping) == false) {
            this.limitFieldsMap.put(keyNoGrouping, this.limitFieldsMap.get("_ANYFIELD_"));
            log.trace("added " + keyNoGrouping + " - " + this.limitFieldsMap.get(keyNoGrouping) + " to the limitFieldsMap because of the _ANYFIELD_ entry");
        }
        if (this.limitFieldsMap.containsKey(keyNoGrouping)) {
            // look for the key without the grouping context
            if (log.isTraceEnabled())
                log.trace("limitFieldsMap contains " + keyNoGrouping);
            Attribute<?> attr = de.getValue();
            // used below if you un-comment to get all hits
            int limit = this.limitFieldsMap.get(keyNoGrouping);
            if (attr instanceof Attributes) {
                Attributes attrs = (Attributes) attr;
                Set<Attribute<? extends Comparable<?>>> attrSet = attrs.getAttributes();
                for (Attribute<? extends Comparable<?>> value : attrSet) {
                    manageHitsAndMisses(keyWithGrouping, keyNoGrouping, value, hitTermMap, hits, misses, countForFieldMap);
                }
            } else {
                manageHitsAndMisses(keyWithGrouping, keyNoGrouping, attr, hitTermMap, hits, misses, countForFieldMap);
            }
        }
    }
    for (String keyNoGrouping : countForFieldMap.keySet()) {
        int limit = this.limitFieldsMap.get(keyNoGrouping);
        Multimap<String, Attribute<? extends Comparable<?>>> hitMap = hits.getUnchecked(keyNoGrouping);
        for (String keyWithGrouping : hitMap.keySet()) {
            for (Attribute<? extends Comparable<?>> value : hitMap.get(keyWithGrouping)) {
                // if(limit <= 0) break; // comment this line if you want to get ALL hits even if the limit is exceeded
                reducedMap.put(keyWithGrouping, value);
                limit--;
            }
        }
        Multimap<String, Attribute<? extends Comparable<?>>> missMap = misses.getUnchecked(keyNoGrouping);
        for (String keyWithGrouping : missMap.keySet()) {
            for (Attribute<? extends Comparable<?>> value : missMap.get(keyWithGrouping)) {
                if (limit <= 0)
                    break;
                reducedMap.put(keyWithGrouping, value);
                limit--;
            }
        }
        if (log.isTraceEnabled()) {
            log.trace("reducedMap:" + reducedMap);
            log.trace("mapOfHits:" + hits.asMap());
            log.trace("mapOfMisses:" + misses.asMap());
        }
        // only generate an original count if a field was reduced
        if (countForFieldMap.get(keyNoGrouping) > this.limitFieldsMap.get(keyNoGrouping)) {
            limitedFieldCounts.put(keyNoGrouping + ORIGINAL_COUNT_SUFFIX, countForFieldMap.get(keyNoGrouping));
        }
    }
    // mutate the document with the changes collected in the above loop
    applyCounts(document, limitedFieldCounts);
    Map<String, Multimap<String, Attribute<? extends Comparable<?>>>> toRemove = Maps.newLinkedHashMap();
    toRemove.putAll(hits.asMap());
    toRemove.putAll(misses.asMap());
    makeReduction(document, toRemove, reducedMap);
    return entry;
}
Also used : HashMap(java.util.HashMap) Attribute(datawave.query.attributes.Attribute) Attributes(datawave.query.attributes.Attributes) Document(datawave.query.attributes.Document) LinkedListMultimap(com.google.common.collect.LinkedListMultimap) Multimap(com.google.common.collect.Multimap) HashMap(java.util.HashMap) Map(java.util.Map)

Example 2 with Attribute

use of datawave.query.attributes.Attribute in project datawave by NationalSecurityAgency.

the class GroupingTransform method flatten.

/**
 * <pre>
 * flush used the countingMap:
 * [[MALE, 16],
 * [MALE, 20],
 * [40, MALE],
 * [40, MALE],
 * [MALE, 22] x 2,
 * [FEMALE, 18],
 * [MALE, 24],
 * [20, MALE],
 * [30, MALE],
 * [FEMALE, 18],
 * [34, MALE]]
 *
 * to create documents list: [
 * {AGE=16, COUNT=1, GENDER=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {COUNT=1, ETA=20, GENERE=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {COUNT=1, ETA=40, GENERE=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {AGE=40, COUNT=1, GENDER=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {COUNT=2, ETA=22, GENERE=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {AGE=18, COUNT=1, GENDER=FEMALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {COUNT=1, ETA=24, GENERE=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {AGE=20, COUNT=1, GENDER=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {AGE=30, COUNT=1, GENDER=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {COUNT=1, ETA=18, GENERE=FEMALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false,
 * {AGE=34, COUNT=1, GENDER=MALE}:20130101_0 test%00;-d5uxna.msizfm.-oxy0iu: [ALL] 1356998400000 false]
 *
 * which is then flattened to just one document with the fields and counts correlated with a grouping context suffix:
 *
 * {
 * AGE.0=16, GENDER.0=MALE, COUNT.0=1,
 * ETA.1=20, GENERE.1=MALE, COUNT.1=1,
 * ETA.2=40, GENERE.2=MALE, COUNT.2=1,
 * AGE.3=40, GENDER.3=MALE, COUNT.3=1,
 * ETA.4=22, GENERE.4=MALE, COUNT.4=2,
 * AGE.5=18, GENDER.5=FEMALE, COUNT.5=1,
 * ETA.6=24, GENERE.6=MALE, COUNT.6=1,
 * AGE.7=20, GENDER.7=MALE, COUNT.7=1,
 * AGE.8=30, GENDER.8=MALE, COUNT.8=1,
 * ETA.9=18, GENERE.9=FEMALE, COUNT.9=1,
 * AGE.A=34, GENDER.A=MALE, COUNT.A=1,
 * }
 * </pre>
 *
 * The Attributes, which have had their visibilities merged, are copied into normal TypeAttributes for serialization to the webserver.
 *
 * @param documents
 */
private void flatten(List<Document> documents) {
    log.trace("flatten {}", documents);
    Document theDocument = new Document(documents.get(documents.size() - 1).getMetadata(), true);
    int context = 0;
    Set<ColumnVisibility> visibilities = new HashSet<>();
    for (Document document : documents) {
        log.trace("document: {}", document);
        for (Entry<String, Attribute<? extends Comparable<?>>> entry : document.entrySet()) {
            String name = entry.getKey();
            visibilities.add(entry.getValue().getColumnVisibility());
            Attribute<? extends Comparable<?>> attribute = entry.getValue();
            attribute.setColumnVisibility(entry.getValue().getColumnVisibility());
            // call copy() on the GroupingTypeAttribute to get a plain TypeAttribute
            // instead of a GroupingTypeAttribute that is package protected and won't serialize
            theDocument.put(name + "." + Integer.toHexString(context).toUpperCase(), (TypeAttribute) attribute.copy(), true, false);
        }
        context++;
    }
    ColumnVisibility combinedVisibility = combine(visibilities);
    log.trace("combined visibilities: {} to {}", visibilities, combinedVisibility);
    theDocument.setColumnVisibility(combinedVisibility);
    documents.clear();
    log.trace("flattened document: {}", theDocument);
    documents.add(theDocument);
}
Also used : Attribute(datawave.query.attributes.Attribute) TypeAttribute(datawave.query.attributes.TypeAttribute) ColumnVisibility(org.apache.accumulo.core.security.ColumnVisibility) Document(datawave.query.attributes.Document) HashSet(java.util.HashSet) LinkedHashSet(java.util.LinkedHashSet)

Example 3 with Attribute

use of datawave.query.attributes.Attribute in project datawave by NationalSecurityAgency.

the class QueryLogicTestHarness method assertLogicResults.

// =============================================
// assert methods
/**
 * Determines if the correct results were obtained for a query.
 *
 * @param logic
 *            key/value response data
 * @param expected
 *            list of key values expected within response data
 * @param checkers
 *            list of additional validation methods
 */
public void assertLogicResults(BaseQueryLogic<Map.Entry<Key, Value>> logic, Collection<String> expected, List<DocumentChecker> checkers) {
    Set<String> actualResults = new HashSet<>();
    if (log.isDebugEnabled()) {
        log.debug("    ======  expected id(s)  ======");
        for (String e : expected) {
            log.debug("id(" + e + ")");
        }
    }
    for (Map.Entry<Key, Value> entry : logic) {
        if (FinalDocumentTrackingIterator.isFinalDocumentKey(entry.getKey())) {
            continue;
        }
        final Document document = this.deserializer.apply(entry).getValue();
        // check all of the types to ensure that all are keepers as defined in the
        // AttributeFactory class
        int count = 0;
        for (Attribute<? extends Comparable<?>> attribute : document.getAttributes()) {
            if (attribute instanceof TimingMetadata) {
            // ignore
            } else if (attribute instanceof Attributes) {
                Attributes attrs = (Attributes) attribute;
                Collection<Class<?>> types = new HashSet<>();
                for (Attribute<? extends Comparable<?>> attr : attrs.getAttributes()) {
                    count++;
                    if (attr instanceof TypeAttribute) {
                        Type<? extends Comparable<?>> type = ((TypeAttribute<?>) attr).getType();
                        if (Objects.nonNull(type)) {
                            types.add(type.getClass());
                        }
                    }
                }
                Assert.assertEquals(AttributeFactory.getKeepers(types), types);
            } else {
                count++;
            }
        }
        // ignore empty documents (possible when only passing FinalDocument back)
        if (count == 0) {
            continue;
        }
        // parse the document
        String extractedResult = this.parser.parse(entry.getKey(), document);
        log.debug("result(" + extractedResult + ") key(" + entry.getKey() + ") document(" + document + ")");
        // verify expected results
        Assert.assertNotNull("extracted result", extractedResult);
        Assert.assertFalse("duplicate result(" + extractedResult + ") key(" + entry.getKey() + ")", actualResults.contains(extractedResult));
        actualResults.add(extractedResult);
        // perform any custom assert checks on document
        for (final DocumentChecker check : checkers) {
            check.assertValid(document);
        }
    }
    log.info("total records found(" + actualResults.size() + ") expected(" + expected.size() + ")");
    // ensure that the complete expected result set exists
    if (expected.size() > actualResults.size()) {
        final Set<String> notFound = new HashSet<>(expected);
        notFound.removeAll(actualResults);
        for (final String m : notFound) {
            log.error("missing result(" + m + ")");
        }
    } else if (expected.size() < actualResults.size()) {
        final Set<String> extra = new HashSet<>(actualResults);
        extra.removeAll(expected);
        for (final String r : extra) {
            log.error("unexpected result(" + r + ")");
        }
    }
    Assert.assertEquals("results do not match expected", expected.size(), actualResults.size());
    Assert.assertTrue("expected and actual values do not match", expected.containsAll(actualResults));
    Assert.assertTrue("expected and actual values do not match", actualResults.containsAll(expected));
}
Also used : Set(java.util.Set) HashSet(java.util.HashSet) Attribute(datawave.query.attributes.Attribute) TypeAttribute(datawave.query.attributes.TypeAttribute) Attributes(datawave.query.attributes.Attributes) TimingMetadata(datawave.query.attributes.TimingMetadata) Document(datawave.query.attributes.Document) Type(datawave.data.type.Type) TypeAttribute(datawave.query.attributes.TypeAttribute) Value(org.apache.accumulo.core.data.Value) Collection(java.util.Collection) Map(java.util.Map) Key(org.apache.accumulo.core.data.Key) HashSet(java.util.HashSet)

Example 4 with Attribute

use of datawave.query.attributes.Attribute in project datawave by NationalSecurityAgency.

the class ResponseFieldChecker method assertValid.

/**
 * Verifies the query response document contains all of the return fields.
 *
 * @param doc
 *            query response document
 */
@Override
public void assertValid(final Document doc) {
    for (final String field : this.fields) {
        final Attribute val = doc.get(field);
        Assert.assertNotNull("missing return field(" + field + ")", val);
        if (val instanceof Attributes) {
            Attributes multiAttr = (Attributes) val;
            for (Attribute attr : multiAttr.getAttributes()) {
                Assert.assertNotNull("missing metadata for field(" + field + ")", attr.getMetadata());
            }
        } else {
            Assert.assertNotNull("missing metadata for field(" + field + ")", val.getMetadata());
        }
    }
    for (final String field : this.missing) {
        final Attribute val = doc.get(field);
        Assert.assertNull("blacklisted return field(" + field + ")", val);
    }
}
Also used : Attribute(datawave.query.attributes.Attribute) Attributes(datawave.query.attributes.Attributes)

Example 5 with Attribute

use of datawave.query.attributes.Attribute in project datawave by NationalSecurityAgency.

the class EventDataQueryExpressionVisitor method extractTypes.

public static Set<Type> extractTypes(AttributeFactory attrFactory, String fieldName, String fieldValue, Key key) {
    final Set<Type> types = new HashSet<>();
    final Queue<Attribute<?>> attrQueue = new LinkedList<>();
    attrQueue.add(attrFactory.create(fieldName, fieldValue, key, true));
    Attribute<?> attr;
    while ((attr = attrQueue.poll()) != null) {
        if (TypeAttribute.class.isAssignableFrom(attr.getClass())) {
            TypeAttribute dta = (TypeAttribute) attr;
            Type t = dta.getType();
            types.add(t);
        } else if (AttributeBag.class.isAssignableFrom(attr.getClass())) {
            attrQueue.addAll(((AttributeBag<?>) attr).getAttributes());
        } else {
            log.warn("Unexpected attribute type when extracting type: " + attr.getClass().getCanonicalName());
        }
    }
    return types;
}
Also used : NoOpType(datawave.data.type.NoOpType) Type(datawave.data.type.Type) TypeAttribute(datawave.query.attributes.TypeAttribute) Attribute(datawave.query.attributes.Attribute) TypeAttribute(datawave.query.attributes.TypeAttribute) LinkedList(java.util.LinkedList) HashSet(java.util.HashSet) AttributeBag(datawave.query.attributes.AttributeBag)

Aggregations

Attribute (datawave.query.attributes.Attribute)21 Document (datawave.query.attributes.Document)17 Attributes (datawave.query.attributes.Attributes)14 Key (org.apache.accumulo.core.data.Key)13 HashSet (java.util.HashSet)11 Value (org.apache.accumulo.core.data.Value)8 TypeAttribute (datawave.query.attributes.TypeAttribute)6 Map (java.util.Map)6 Entry (java.util.Map.Entry)6 Cardinality (datawave.query.attributes.Cardinality)4 QueryImpl (datawave.webservice.query.QueryImpl)4 GenericQueryConfiguration (datawave.webservice.query.configuration.GenericQueryConfiguration)4 Set (java.util.Set)4 Type (datawave.data.type.Type)3 Multimap (com.google.common.collect.Multimap)2 Content (datawave.query.attributes.Content)2 FieldValueCardinality (datawave.query.attributes.FieldValueCardinality)2 PreNormalizedAttribute (datawave.query.attributes.PreNormalizedAttribute)2 TimingMetadata (datawave.query.attributes.TimingMetadata)2 DatawaveKey (datawave.query.data.parsers.DatawaveKey)2