Creating your own OCR Validation Control plugin

You can create your own OCR Validation plugin.

Conditions:

• The new OCR Validation class must implement the "OCRTemplateControlParser" interface.
• The new OCR Validation class must be declared under the package "com.openkm.ocr.template.controlparser".
• The new OCR Validation class must be annotated with "@PluginImplementation".

OCR TemplateControlParser interface:

package com.openkm.plugin.ocr.template;

import net.xeoh.plugins.base.Plugin;

import com.openkm.dao.bean.OCRTemplateControlField;

/**
 * OCRTemplateControlParser
 * 
 * 
 */
public interface OCRTemplateControlParser extends Plugin {
	boolean parse(OCRTemplateControlField otcf, String text) throws OCRTemplateException, OCRParserEmptyValueException;
	String getName();
	boolean isPatternRequired();
	String info();
}

Methods description

Example:

StringSimilarParser class:

package com.openkm.ocr.template.controlparser;

import java.util.regex.Matcher;
import java.util.regex.Pattern;

import net.xeoh.plugins.base.annotations.PluginImplementation;

import com.openkm.core.DatabaseException;
import com.openkm.dao.ConfigDAO;
import com.openkm.dao.bean.OCRTemplateControlField;
import com.openkm.ocr.template.OCRParserEmptyValueException;
import com.openkm.ocr.template.OCRTemplateControlParser;
import com.openkm.ocr.template.OCRTemplateException;
import com.openkm.util.AlgorithmUtils;

/**
 * StringParser
 *
 */
@PluginImplementation
public class StringSimilarParser implements OCRTemplateControlParser {
	@Override
	public boolean parse(OCRTemplateControlField otf, String text) throws OCRTemplateException, OCRParserEmptyValueException {
		if (text == null || text.equals("")) {
			return false;
		} else {
			if (otf.getPattern() == null || otf.getPattern().equals("")) {
				int distance = AlgorithmUtils.computeLevenshteinDistance(text, otf.getValue());
				try {
					return distance<ConfigDAO.getInteger("ocr.control.similar.distance", 5);
				} catch (DatabaseException e) {
					throw new OCRTemplateException(e);
				}
			} else {
				Pattern pattern = Pattern.compile("(" + otf.getPattern() + ")", Pattern.UNICODE_CASE);
				Matcher matcher = pattern.matcher(text);
				if (matcher.find() && matcher.groupCount() == 1) {
					if (otf.getValue().equals("")) {
						return true;
					} else {
						int distance = AlgorithmUtils.computeLevenshteinDistance(matcher.group(), otf.getValue());
						try {
							return distance<ConfigDAO.getInteger("ocr.control.similar.distance", 5);
						} catch (DatabaseException e) {
							throw new OCRTemplateException(e);
						}
					}
				} else {
					return false;
				}
			}
		}
	}

	@Override
	public String getName() {
		return "Similar String";
	}

	@Override
	public boolean isPatternRequired() {
		return false;
	}

	@Override
	public String info() {
		StringBuffer sb = new StringBuffer();
		sb.append("Empty pattern case: true if extracted text similar field value.\n");
		sb.append("Pattern with value field empty: true if pattern found.\n");
		sb.append("Pattern with value field: true if pattern found similar value.\n");
		return null;
	}
}