/**
* Parses the predicate definitions from the CSV file.
*
* @param csvFile The buffered reader containing the contents of the CSV file we are trying parse.
* @param ds The destination datastore for the csv file.
* @return The next line to be parsed from the file.
* @throws IOException If unable to read from the csvFile.
*/
private String parseDefinitions(final BufferedReader csvFile, final Datastore db)
throws IOException {
// Keep parsing lines and putting them in the newly formed nominal
// variable until we get to a line indicating the end of file or a new
// variable section.
String line = csvFile.readLine();
while ((line != null) && Character.isDigit(line.charAt(0))) {
/*
* TODO Parsing predicates.
*
// Parse arguments - for predicate vocab element.
String[] token = line.split(":|(?<!\\\\)-");
PredicateVocabElement pve = new PredicateVocabElement(db,
this.stripEscChars(token[1]));
for (String arg : token[2].split(",")) {
pve.appendFormalArg(parseFormalArgument(arg, db));
}
db.addPredVE(pve);
*/
// Get the next line in the file for reading.
line = csvFile.readLine();
}
return line;
}
/**
* Method to invoke when we encounter a block of text in the CSV file that is the contents of a
* predicate variable.
*
* @param csvFile The csvFile we are currently parsing.
* @param var The variable that we will be adding cells too.
* @param arg The matrix template we are using when parsing individual matrix elements to put in
* the spreadsheet.
* @return The next line in the file that is not part of the block of text in the CSV file.
* @throws IOException If unable to read the file correctly.
*/
private String parseMatrixVariable(
final BufferedReader csvFile, final Variable var, final Argument arg) throws IOException {
String line = csvFile.readLine();
while ((line != null) && Character.isDigit(line.charAt(0))) {
// Split the line into tokens using a comma delimiter.
String[] tokens = line.split(",");
Cell newCell = var.createCell();
// Set the onset and offset from tokens in the line.
newCell.setOnset(tokens[DATA_ONSET]);
newCell.setOffset(tokens[DATA_OFFSET]);
// Strip the brackets from the first and last argument.
tokens[DATA_INDEX] = tokens[DATA_INDEX].substring(1, tokens[DATA_INDEX].length());
int end = tokens.length - 1;
tokens[end] = tokens[end].substring(0, tokens[end].length() - 1);
parseFormalArgs(tokens, DATA_INDEX, var.getVariableType(), (MatrixValue) newCell.getValue());
// Get the next line in the file for reading.
line = csvFile.readLine();
}
return line;
}
/**
* Method to invoke when we encounter a block of text in the CSV file that is the contents of a
* variable.
*
* @param csvFile The csvFile we are currently parsing.
* @param var The variable that we will be adding cells too.
* @param The populator to use when converting the contents of the cell into a datavalue that can
* be inserted into the spreadsheet.
* @return The next line in the file that is not part of the block of text in the CSV file.
* @throws IOException If unable to read the file correctly.
*/
private String parseEntries(
final BufferedReader csvFile, final Variable var, final EntryPopulator populator)
throws IOException {
// Keep parsing lines and putting them in the newly formed nominal
// variable until we get to a line indicating the end of file or a new
// variable section.
String line = csvFile.readLine();
while ((line != null) && Character.isDigit(line.charAt(0))) {
// Split the line into tokens using a comma delimiter.
String[] tokens = line.split(",");
// BugzID: 1075 - If the line ends with an escaped new line - add
// the next line to the current text field.
while ((line != null) && line.endsWith("\\") && !line.endsWith("\\\\")) {
line = csvFile.readLine();
String content = tokens[tokens.length - 1];
content = content.substring(0, content.length() - 1);
tokens[tokens.length - 1] = content + '\n' + line;
}
Cell newCell = var.createCell();
// Set the onset and offset from tokens in the line.
newCell.setOnset(tokens[DATA_ONSET]);
newCell.setOffset(tokens[DATA_OFFSET]);
populator.populate(tokens, newCell.getValue());
// Get the next line in the file for reading.
line = csvFile.readLine();
}
return line;
}
/**
* Method to invoke when we encounter a block of text that is a variable.
*
* @param csvFile The CSV file we are currently reading.
* @param line The line of the CSV file we are currently reading.
* @param db The data store we are populating with data from the CSV file.
* @return The next String that is not part of the currently variable that we are parsing.
* @throws IOException When we are unable to read from the csvFile.
* @throws UserWarningException When we are unable to create variables.
*/
private String parseVariable(
final BufferedReader csvFile, final String line, final Datastore ds, final String version)
throws IOException, UserWarningException {
// Determine the variable name and type.
String[] tokens = line.split("\\(");
String varName = this.stripEscChars(tokens[0].trim());
String varType = null;
String varComment = "";
boolean varVisible = true;
if (version.equals("#4")) {
String[] varArgs = tokens[1].split(",");
varType = varArgs[0];
varVisible = Boolean.parseBoolean(varArgs[1]);
varComment = varArgs[2].substring(0, varArgs[2].indexOf(")"));
} else if (version.equals("#3")) {
varType = tokens[1].substring(0, tokens[1].indexOf(","));
varVisible =
Boolean.parseBoolean(
tokens[1].substring(tokens[1].indexOf(",") + 1, tokens[1].indexOf(")")));
} else {
varType = tokens[1].substring(0, tokens[1].indexOf(")"));
}
// BugzID:1703 - Ignore old macshapa query variables, we don't have a
// reliable mechanisim for loading their predicates. Given problems
// between the untyped nature of macshapa and the typed nature of
// OpenSHAPA.
if (varName.equals("###QueryVar###")) {
String lineEater = csvFile.readLine();
while ((lineEater != null) && Character.isDigit(lineEater.charAt(0))) {
lineEater = csvFile.readLine();
}
return lineEater;
}
// Create variable to put cells within.
Argument.Type variableType = getVarType(varType);
Variable newVar = ds.createVariable(varName, variableType);
newVar.setHidden(!varVisible);
// Read text variable.
if (variableType == Argument.Type.TEXT) {
return parseEntries(csvFile, newVar, new PopulateText());
} else if (variableType == Argument.Type.NOMINAL) {
// Read nominal variable.
return parseEntries(csvFile, newVar, new PopulateNominal());
} /* else if (getVarType(varType)
== MatrixVocabElement.MatrixType.INTEGER) {
// Read integer variable.
return parseEntries(csvFile,
newVar.getLegacyVariable(),
new PopulateInteger(legacyDb));
} else if (getVarType(varType) == MatrixVocabElement.MatrixType.FLOAT) {
return parseEntries(csvFile,
newVar.getLegacyVariable(),
new PopulateFloat(legacyDb));
}*/ else if (variableType == Argument.Type.MATRIX) {
// Read matrix variable - Build vocab for matrix.
String[] vocabString = tokens[1].split("(?<!\\\\)-");
// Get the vocab element for the matrix and clean it up to be
// populated with arguments from the CSV file.
Argument newArg = newVar.getVariableType();
// MatrixVocabElement mve = legacyDb.getMatrixVE(varName);
// mve.deleteFormalArg(0);
// For each of the formal arguments in the file - parse it and
// create a formal argument in the matrix vocab element.
for (String arg : vocabString[1].split(",")) {
newArg.childArguments.add(parseFormalArgument(arg));
}
newVar.setVariableType(newArg);
return parseMatrixVariable(csvFile, newVar, newArg);
// Read predicate variable.
} /*else if (getVarType(varType)
== MatrixVocabElement.MatrixType.PREDICATE) {
return parsePredicateVariable(csvFile, newVar.getLegacyVariable());
}*/
throw new IllegalStateException("Unknown variable type.");
}
/**
* This method parses a CSV input stream and populates the database (and spreadsheet) with data.
* The caller is responsible for managing the stream.
*
* @param inStream The stream to deserialized when populating the database.
* @return populated database on sucess, null otherwise.
*/
public Datastore openAsCSV(final InputStream inStream) {
try {
LOGGER.event("open csv database from stream");
Datastore db = DatastoreFactory.newDatastore();
db.setTitleNotifier(OpenSHAPA.getApplication());
InputStreamReader isr = new InputStreamReader(inStream);
BufferedReader csvFile = new BufferedReader(isr);
// Read each line of the CSV file.
String line = csvFile.readLine();
// If we have a version identifier parse the file using the schema
// that matches that identifier.
if ("#4".equalsIgnoreCase(line)) {
// Version 4 includes a comment for columns.
line = parseDefinitions(csvFile, db);
while (line != null) {
line = parseVariable(csvFile, line, db, "#4");
}
} else if ("#3".equalsIgnoreCase(line)) {
// Version 3 includes column visible status after the column type
// Parse predicate definitions first.
line = parseDefinitions(csvFile, db);
while (line != null) {
line = parseVariable(csvFile, line, db, "#3");
}
} else if ("#2".equalsIgnoreCase(line)) {
// Parse predicate definitions first.
line = parseDefinitions(csvFile, db);
while (line != null) {
line = parseVariable(csvFile, line, db);
}
} else {
// Use the original schema to load the file - just variables,
// and no escape characters.
while (line != null) {
line = parseVariable(csvFile, line, db);
}
}
csvFile.close();
isr.close();
return db;
} catch (IOException e) {
LOGGER.error("Unable to read line from CSV file", e);
} catch (UserWarningException e) {
LOGGER.error("Unable to create new variable.", e);
}
// Error encountered - return null.
return null;
}