/**
* Process the identifiers.
*
* @param type The type of thing these identifiers are.
* @param input The creation input.
* @param ids The identifiers.
* @param unmatchedIds A collector for unmatched identifiers.
* @param tempBag The temporary bag to add results to.
* @throws IOException If we can't from the request.
* @throws ClassNotFoundException if the type is not valid.
* @throws InterMineException If something goes wrong building the bag.
* @throws ObjectStoreException If there is a problem on the database level.
*/
protected void processIdentifiers(
final String type,
ListCreationInput input,
final Set<String> ids,
final Set<String> unmatchedIds,
final InterMineBag tempBag)
throws IOException, ClassNotFoundException, InterMineException, ObjectStoreException {
final Collection<String> addIssues = input.getAddIssues();
String line;
final StrMatcher matcher = getMatcher();
final BufferedReader r = getReader(request);
try {
while ((line = r.readLine()) != null) {
final StrTokenizer st = new StrTokenizer(line, matcher, StrMatcher.doubleQuoteMatcher());
while (st.hasNext()) {
final String token = st.nextToken();
ids.add(token);
}
if (ids.size() >= BAG_QUERY_MAX_BATCH_SIZE) {
addIdsToList(ids, tempBag, type, input.getExtraValue(), unmatchedIds, addIssues);
ids.clear();
}
}
} finally {
if (r != null) {
r.close();
}
}
if (ids.size() > 0) {
addIdsToList(ids, tempBag, type, input.getExtraValue(), unmatchedIds, addIssues);
}
}
/**
* Parses the pattern string, which should be a comma separated list of regular expressions, each
* of which may be surrounded with double quotes.
*
* @param inStr pattern string to parse
* @return list of pattern regular expressions
*/
private ArrayList<Pattern> parsePatternString(String inStr) {
ArrayList<Pattern> result = new ArrayList<Pattern>();
StrTokenizer tokenizer = new StrTokenizer(inStr, ',', '"');
tokenizer.setIgnoreEmptyTokens(true);
tokenizer.setIgnoredMatcher(StrMatcher.charSetMatcher(" \t\n\r"));
while (tokenizer.hasNext()) {
String tok = (String) tokenizer.next();
Pattern pat = Pattern.compile(tok);
result.add(pat);
}
return result;
}
@Override
protected void processInput(String keggGeneId, String keggGeneRecord, Resource output) {
Map<String, String> recordSections = KeggUtils.getSectionsFromKeggRecord(keggGeneRecord);
StrTokenizer tokenizer = new StrTokenizer();
if (recordSections.containsKey(PATHWAY_RECORD_SECTION)) {
for (String line : recordSections.get(PATHWAY_RECORD_SECTION).split("\\r?\\n")) {
String keggPathwayId = tokenizer.reset(line).nextToken();
Resource keggPathwayNode =
LSRNUtils.createInstance(
output.getModel(), LSRNUtils.getClass(LSRN.Namespace.KEGG_PATHWAY), keggPathwayId);
output.addProperty(SIO.is_participant_in, keggPathwayNode);
}
}
}
@Override
public void init(RuleConfiguration ruleConfig) {
minOrderValue =
new Money(
BigDecimal.valueOf(
Double.valueOf(
ruleConfig.getConfigItemValue(RuleConfigConstants.MIN_ORDER_VALUE))));
fixedRsOff =
new Money(
BigDecimal.valueOf(
Double.valueOf(
ruleConfig.getConfigItemValue(RuleConfigConstants.FIXED_DISCOUNT_RS_OFF))));
StrTokenizer strTokCategories =
new StrTokenizer(ruleConfig.getConfigItemValue(RuleConfigConstants.CATEGORY_LIST), ",");
categories = StringToIntegerList.convert((List<String>) strTokCategories.getTokenList());
StrTokenizer strTokClients =
new StrTokenizer(ruleConfig.getConfigItemValue(RuleConfigConstants.CLIENT_LIST), ",");
client_list = StringToIntegerList.convert((List<String>) strTokClients.getTokenList());
log.info(
"minOrderValue : " + minOrderValue.toString() + ", fixedRsOff : " + fixedRsOff.toString());
}
/**
* Converts a String back to connection parameters.
*
* @param input String from configuration
* @return JDBC connection parameters
*/
protected static Properties propertiesFromString(String input) {
if (input != null && !input.isEmpty()) {
Properties result = new Properties();
StrTokenizer propertyTokenizer = StrTokenizer.getCSVInstance(input);
StrTokenizer valueTokenizer = StrTokenizer.getCSVInstance();
valueTokenizer.setDelimiterChar('=');
while (propertyTokenizer.hasNext()) {
valueTokenizer.reset(propertyTokenizer.nextToken());
String[] values = valueTokenizer.getTokenArray();
if (values.length == 2) {
result.put(values[0], values[1]);
}
}
return result;
} else {
return null;
}
}
protected static <T> void parseEDISegmentFields(
EDIMessage ediMessage, Object segment, String segmentLine)
throws IllegalAccessException, InvocationTargetException, ClassNotFoundException,
ConversionException, InstantiationException {
if (LOG.isDebugEnabled()) {
LOG.debug("Before Field Values: " + ReflectionToStringBuilder.toString(segment));
LOG.debug("Segment Values: " + segmentLine);
}
// now, tokenize the line, and set the fields.
StrTokenizer tokenizer = new StrTokenizer(segmentLine, ediMessage.elementDelimiter());
tokenizer.setEmptyTokenAsNull(true);
tokenizer.setIgnoreEmptyTokens(false);
// move past the initial tag.
tokenizer.next();
Iterator<Field> fieldIterator =
Arrays.asList(segment.getClass().getDeclaredFields()).iterator();
while (tokenizer.hasNext() && fieldIterator.hasNext()) {
Field field = fieldIterator.next();
String val = tokenizer.nextToken();
// check field to see if it is a component of regular field type.
if (field.isAnnotationPresent(EDIComponent.class)) {
EDIComponent ediComponent = field.getAnnotation(EDIComponent.class);
Collection obj = CollectionFactory.newInstance(field.getType());
Class objType = getCollectionType(field);
char componentDelimiter =
ediComponent.delimiter() == Character.UNASSIGNED
? ediMessage.componentDelimiter()
: ediComponent.delimiter();
// parse each element to the collection.
StrTokenizer componentTokenizer = new StrTokenizer(val, componentDelimiter);
componentTokenizer.setEmptyTokenAsNull(true);
componentTokenizer.setIgnoreEmptyTokens(false);
while (componentTokenizer.hasNext()) {
String component = componentTokenizer.nextToken();
Object fieldObj =
objType.cast(FieldAwareConverter.convertFromString(objType, field, component));
obj.add(fieldObj);
}
BeanUtils.setProperty(segment, field.getName(), obj);
} else {
if (val == null) {
LOG.debug(" " + field.getName() + " -> null");
continue;
}
// try and populate the field.
try {
Object fieldObj = FieldAwareConverter.convertFromString(field.getType(), field, val);
LOG.debug(" " + field.getName() + " -> " + val);
BeanUtils.setProperty(segment, field.getName(), fieldObj);
} catch (Exception e) {
throw new ConversionException(
"Exception setting: "
+ segment.getClass()
+ "."
+ field.getName()
+ " with value: "
+ val,
e);
}
}
}
if (LOG.isDebugEnabled()) {
LOG.debug("After Field Values: " + ReflectionToStringBuilder.toString(segment));
}
}
protected static <T> void processSegmentGroup(
EDIMessage ediMessage,
T object,
Queue<String> lookAhead,
SegmentIterator segmentIterator,
FieldMatch fm)
throws InstantiationException, IllegalAccessException, InvocationTargetException,
ClassNotFoundException, ConversionException, EDIMessageException {
LOG.debug("Object: " + ReflectionToStringBuilder.toString(object));
LOG.debug("Field: " + fm.getField().getName());
Class<?> segmentGroupClass = getEDISegmentOrGroupType(fm.getField());
if (!segmentGroupClass.isAnnotationPresent(EDISegmentGroup.class)) {
throw new EDIMessageException("Segment Group should have annotation.");
}
LOG.debug("Segment Group Type: " + segmentGroupClass);
String line = fm.getLine();
EDISegmentGroup es = segmentGroupClass.getAnnotation(EDISegmentGroup.class);
if (StringUtils.equals(es.header(), line)) {
// feed line.
} else {
LOG.debug("Adding to Look Ahead: " + line);
lookAhead.add(line);
}
if (Collection.class.isAssignableFrom(fm.getField().getType())) {
Collection obj = CollectionFactory.newInstance(fm.getField().getType());
BeanUtils.setProperty(object, fm.getField().getName(), obj);
String segmentTag = getSegmentTag(fm.getField(), true);
while (true) {
LOG.debug("Looping to collect Collection of Segment Groups");
// parse the group...
Field[] fields = segmentGroupClass.getDeclaredFields();
final List<Field> fieldsList = Arrays.asList(fields);
ListIterator<Field> fieldIterator = fieldsList.listIterator(0);
Object collectionObj = segmentGroupClass.newInstance();
while (fieldIterator.hasNext() && (segmentIterator.hasNext() || lookAhead.size() > 0)) {
if (startOfNewRecursiveObject(fieldIterator, fieldsList, segmentGroupClass)) {
String next = lookAhead.size() > 0 ? lookAhead.remove() : segmentIterator.next();
lookAhead.add(next);
if (!isSegmentHeirarchicalLevelAndDeeperLevel(
line, next, Character.toString(ediMessage.elementDelimiter()))) {
LOG.debug("Reaching new instance of list.");
break;
}
}
parseEDISegmentOrSegmentGroup(
ediMessage, collectionObj, fieldIterator, lookAhead, segmentIterator);
}
obj.add(collectionObj);
// look to next line...
String nextLine = lookAhead.size() > 0 ? lookAhead.remove() : segmentIterator.next();
// get the first element of the line.
StrTokenizer nextLineTokenizer = new StrTokenizer(nextLine, ediMessage.elementDelimiter());
if (StringUtils.equals(segmentTag, nextLineTokenizer.nextToken())) {
if (!isSegmentEqual(line, nextLine, Character.toString(ediMessage.elementDelimiter()))) {
LOG.debug("Reaching new collection");
lookAhead.add(nextLine);
break;
}
LOG.debug("Might be a repeat..");
LOG.debug("Next line: " + line);
lookAhead.add(nextLine);
} else {
lookAhead.add(nextLine);
break;
}
// now, look ahead to see whether the next line is of the same
// object type..
if (!segmentIterator.hasNext() && lookAhead.size() == 0) {
break;
}
}
} else {
Field[] fields = segmentGroupClass.getDeclaredFields();
Iterator<Field> fieldIterator = Arrays.asList(fields).iterator();
Object obj = segmentGroupClass.newInstance();
while (fieldIterator.hasNext() && (segmentIterator.hasNext() || lookAhead.size() > 0)) {
parseEDISegmentOrSegmentGroup(ediMessage, obj, fieldIterator, lookAhead, segmentIterator);
}
BeanUtils.setProperty(object, fm.getField().getName(), obj);
}
// look at next...
if (StringUtils.isNotBlank(es.header())) {
line = lookAhead.size() > 0 ? lookAhead.remove() : segmentIterator.next();
if (StringUtils.endsWith(es.footer(), line)) {
// feed line.
LOG.debug("Popping footer off of the line iterator.");
} else {
lookAhead.add(line);
}
}
}
static List<double[]> readDataBodyAsDouble(
BufferedReader br, int[] remappedColumns, int mappedColumnCount)
throws NumberFormatException, IOException {
List<double[]> rows = new ArrayList<double[]>(500);
int colCount = 0; // Number of columns (minus one) - must match in each row
int curRow = 0; // Current row number - 0 based;
String s = null; // Single line from file
while ((s = br.readLine()) != null) {
StrTokenizer strTk = new StrTokenizer(s);
strTk.setIgnoreEmptyTokens(false);
strTk.setDelimiterChar('\t');
String src[] = strTk.getTokenArray();
// String src[] = s.split("\t");
if (isNonEmptyLine(src)) {
if (colCount == 0) {
colCount = src.length; // initialize column count
if (remappedColumns == null) {
// assign default mapping now
remappedColumns = new int[colCount];
for (int i = 0; i < colCount; i++) {
remappedColumns[i] = i;
}
mappedColumnCount = colCount;
}
} else {
if (src.length != colCount) {
for (int i = 0; i < src.length; i++) {
System.out.println(i + ": " + src[i]);
}
throw new IllegalStateException(
"Parse Error: Row "
+ curRow
+ " has "
+ src.length
+ " columns, previous columns had "
+ colCount
+ " columns.");
}
}
double[] row = new double[mappedColumnCount];
for (int i = 0; i < src.length; i++) {
if (remappedColumns[i] != -1) {
if (src[i].length() > 0) {
// Simple hack to ignore text columns in the predict.txt file
try {
row[remappedColumns[i]] = Double.parseDouble(src[i]);
} catch (Exception e) {
row[remappedColumns[i]] = Double.NaN;
}
} else {
row[remappedColumns[i]] = 0;
}
}
}
rows.add(row);
} else {
// ignore empty lines
}
curRow++;
}
log.debug("Found " + curRow + " rows in predict.txt file.");
return rows;
}
