Mercurial > dive4elements > river
view flys-backend/src/main/java/de/intevation/flys/importer/parsers/SedimentYieldParser.java @ 4723:83304d3651ec
DA66Parser: Whitespace cosmetics.
author | Felix Wolfsteller <felix.wolfsteller@intevation.de> |
---|---|
date | Thu, 27 Dec 2012 16:14:15 +0100 |
parents | 1c5370683769 |
children | e8e092c00791 2435968b30b1 |
line wrap: on
line source
package de.intevation.flys.importer.parsers; import java.io.File; import java.io.IOException; import java.text.NumberFormat; import java.text.ParseException; import java.util.ArrayList; import java.util.List; import java.util.regex.Matcher; import java.util.regex.Pattern; import org.apache.log4j.Logger; import de.intevation.flys.importer.ImportGrainFraction; import de.intevation.flys.importer.ImportSedimentYield; import de.intevation.flys.importer.ImportSedimentYieldValue; import de.intevation.flys.importer.ImportTimeInterval; import de.intevation.flys.importer.ImportUnit; import de.intevation.flys.model.GrainFraction; public class SedimentYieldParser extends LineParser { private static final Logger log = Logger.getLogger(SedimentYieldParser.class); public static final NumberFormat nf = NumberFormat.getInstance(DEFAULT_LOCALE); public static final String FRAKTION_START = "Fraktion:"; public static final String FRACTION_COARSE_STR = "_Grobkorn.csv"; public static final String FRACTION_FINE_MIDDLE_STR = "_Fein-Mittel-Kies.csv"; public static final String FRACTION_SAND = "_Sand.csv"; public static final String FRACTION_SUSP_SAND = "_susp_Sand.csv"; public static final String FRACTION_SUSP_SAND_BED = "_bettbild_Anteil_susp_Sand.csv"; public static final String FRACTION_SUSP_SAND_BED_EPOCH = "_susp_Sand_bettbildAnteil.csv"; public static final String FRACTION_SUSPENDED_SEDIMENT = "_Schwebstoff.csv"; public static final String FRACTION_TOTAL = "_gesamt.csv"; public static final Pattern TIMEINTERVAL_SINGLE = Pattern.compile("\\D*([0-9]+?)\\D*"); public static final Pattern TIMEINTERVAL_EPOCH = Pattern.compile("\\D*([0-9]+?)\\s*-\\s*([0-9]+?)\\D*"); public static final Pattern META_FRACTION = Pattern.compile("^Fraktion: (.*)"); public static final Pattern META_UNIT = Pattern.compile("^Einheit: \\[(.*)\\].*"); public static final Pattern META_COLUMN_NAMES = Pattern.compile("^Fluss-km.*"); public static final Pattern META_GRAIN_FRACTION_A = Pattern.compile("\\D*(([0-9]+?,[0-9]+?)\\s*-|([0-9]++)\\s*-)(([0-9]+?,[0-9]+?)|([0-9]++))\\s*([a-zA-Z]+?)\\W*\\D*"); public static final Pattern META_GRAIN_FRACTION_B = Pattern.compile("(<|>){1}\\s*(\\w++)\\s*(([0-9]+?,[0-9]+?)\\s*-|([0-9]++)\\s*-)(([0-9]+?,[0-9]+?)|([0-9]++))\\s*([a-zA-Z]+?)"); public static final Pattern META_GRAIN_FRACTION_C = Pattern.compile("(<|>){1}\\s*((([0-9]+?,[0-9]+?)|([0-9]++))\\s*(\\w+))"); protected List<ImportSedimentYield> sedimentYields; protected ImportSedimentYield[] current; protected ImportGrainFraction grainFraction; protected ImportUnit unit; protected String description; protected String[] columnNames; public SedimentYieldParser() { sedimentYields = new ArrayList<ImportSedimentYield>(); } @Override public void parse(File file) throws IOException { description = file.getName(); super.parse(file); } @Override protected void reset() { current = null; grainFraction = null; unit = null; } @Override protected void finish() { if (current != null) { for (ImportSedimentYield isy: current) { sedimentYields.add(isy); } } description = null; } @Override protected void handleLine(int lineNum, String line) { if (line.startsWith(START_META_CHAR)) { handleMetaLine(stripMetaLine(line)); } else { handleDataLine(line); } } protected void handleMetaLine(String line) { if (handleMetaUnit(line)) { return; } else if (handleMetaFraction(line)) { return; } else if (handleColumnNames(line)) { return; } else { log.warn("SYP: Unknown meta line: '" + line + "'"); } } protected boolean handleMetaUnit(String line) { Matcher m = META_UNIT.matcher(line); if (m.matches()) { unit = new ImportUnit(m.group(1)); return true; } return false; } public boolean handleMetaFraction(String line) { Matcher m = META_FRACTION.matcher(line); if (m.matches()) { String tmp = m.group(1); this.grainFraction = buildGrainFraction(tmp); return true; } else if (line.startsWith(FRAKTION_START)) { String newLine = line.replace(FRAKTION_START, "").trim(); if (newLine.length() == 0) { log.debug("Found total grain fraction."); this.grainFraction = new ImportGrainFraction(GrainFraction.TOTAL); return true; } } return false; } public boolean handleColumnNames(String line) { Matcher m = META_COLUMN_NAMES.matcher(line); if (m.matches()) { columnNames = line.split(SEPERATOR_CHAR); initializeSedimentYields(); return true; } return false; } protected void handleDataLine(String line) { String[] vals = line.split(SEPERATOR_CHAR); if (vals == null || vals.length < columnNames.length-1) { log.warn("SYP: skip invalid data line: '" + line + "'"); return; } try { Double km = nf.parse(vals[0]).doubleValue(); for (int i = 1, n = columnNames.length-1; i < n; i++) { String curVal = vals[i]; if (curVal != null && curVal.length() > 0) { current[i-1].addValue(new ImportSedimentYieldValue( km, nf.parse(vals[i]).doubleValue() )); } } } catch (ParseException pe) { log.warn("SYP: Error while parsing numbers in '" + line + "':", pe); } } private void initializeSedimentYields() { // skip first column (Fluss-km) and last column (Hinweise) current = new ImportSedimentYield[columnNames.length-2]; for (int i = 0, n = columnNames.length; i < n-2; i++) { current[i] = new ImportSedimentYield(this.description); current[i].setTimeInterval(getTimeInterval(columnNames[i+1])); current[i].setUnit(unit); current[i].setGrainFraction(grainFraction); } } private ImportTimeInterval getTimeInterval(String column) { try { Matcher a = TIMEINTERVAL_EPOCH.matcher(column); if (a.matches()) { int yearA = nf.parse(a.group(1)).intValue(); int yearB = nf.parse(a.group(2)).intValue(); return new ImportTimeInterval( getDateFromYear(yearA), getDateFromYear(yearB) ); } Matcher b = TIMEINTERVAL_SINGLE.matcher(column); if (b.matches()) { int year = nf.parse(b.group(1)).intValue(); return new ImportTimeInterval(getDateFromYear(year)); } log.warn("SYP: Unknown time interval string: '" + column + "'"); } catch (ParseException pe) { log.warn("SYP: Error while parsing years: " + column, pe); } return null; } private ImportGrainFraction buildGrainFraction(String gfStr) { Matcher a = META_GRAIN_FRACTION_A.matcher(gfStr); if (a.matches()) { String lowerA = a.group(2); String lowerB = a.group(3); String upperA = a.group(4); String upperB = a.group(5); String unitStr = a.group(7); String lower = lowerA != null ? lowerA : lowerB; String upper = upperA != null ? upperA : upperB; try { return new ImportGrainFraction( getGrainFractionTypeName(this.description), nf.parse(lower).doubleValue(), nf.parse(upper).doubleValue(), new ImportUnit(unitStr) ); } catch (ParseException pe) { log.warn("SYP: Error while parsing ranges of: '" + gfStr + "'"); } } Matcher b = META_GRAIN_FRACTION_B.matcher(gfStr); if (b.matches()) { String lowerA = b.group(4); String lowerB = b.group(5); String upperA = b.group(6); String upperB = b.group(7); String unitStr = b.group(9); String lower = lowerA != null ? lowerA : lowerB; String upper = upperA != null ? upperA : upperB; try { return new ImportGrainFraction( getGrainFractionTypeName(this.description), nf.parse(lower).doubleValue(), nf.parse(upper).doubleValue(), new ImportUnit(unitStr) ); } catch (ParseException pe) { log.warn("SYP: Error while parsing ranges of: '" + gfStr + "'"); } } Matcher c = META_GRAIN_FRACTION_C.matcher(gfStr); if (c.matches()) { String oper = c.group(1); String valueStr = c.group(3); String unitStr = c.group(6); try { Double value = nf.parse(valueStr).doubleValue(); if (oper.equals(">")) { return new ImportGrainFraction( getGrainFractionTypeName(this.description), value, null, new ImportUnit(unitStr) ); } else { return new ImportGrainFraction( getGrainFractionTypeName(this.description), null, value, new ImportUnit(unitStr) ); } } catch (ParseException pe) { log.warn("SYP: Error while parsing ranges of: '" + gfStr + "'"); } } log.warn("SYP: Unknow grain fraction: '" + gfStr + "'"); return null; } public static String getGrainFractionTypeName(String filename) { if (filename.endsWith(FRACTION_COARSE_STR)) { return GrainFraction.COARSE; } else if (filename.endsWith(FRACTION_FINE_MIDDLE_STR)) { return GrainFraction.FINE_MIDDLE; } else if (filename.endsWith(FRACTION_SAND) && !filename.endsWith(FRACTION_SUSP_SAND)) { return GrainFraction.SAND; } else if (filename.endsWith(FRACTION_SUSP_SAND) && !filename.endsWith(FRACTION_SUSP_SAND_BED)) { return GrainFraction.SUSP_SAND; } else if (filename.endsWith(FRACTION_SUSP_SAND_BED) || filename.endsWith(FRACTION_SUSP_SAND_BED_EPOCH)) { return GrainFraction.SUSP_SAND_BED; } else if (filename.endsWith(FRACTION_SUSPENDED_SEDIMENT)) { return GrainFraction.SUSPENDED_SEDIMENT; } else if (filename.endsWith(FRACTION_TOTAL)) { return GrainFraction.TOTAL; } else { log.warn("SYP: Unknown grain fraction '" + filename + "'"); return "unknown"; } } public List<ImportSedimentYield> getSedimentYields() { return sedimentYields; } } // vim:set ts=4 sw=4 si et sta sts=4 fenc=utf8 :