// -----------------------------------------------------
// Description: Determine the rotation of the png file
// and rotate it whereby the side edge is 90 degrees.
// Return the angle of rotation.
// -----------------------------------------------------
double autoAlignRotate(String fileName, String destinationDirectory) {
String fileName_woext = (fileName.substring(0, fileName.lastIndexOf('.')));
Spawn.execute(
Settings.Programs.CONVERT.path(),
"-deskew",
"40%",
"+repage",
destinationDirectory + File.separator + fileName_woext + File.separator + fileName,
destinationDirectory + File.separator + fileName_woext + File.separator + fileName);
return 0.0;
}
// -----------------------------------------------------
// Description: Normalize the file to have the origin
// be the top left of the form non white space. We
// are gettting rid of the whitespace borders.
// -----------------------------------------------------
void normalizeByTrimming(String fileName, String destinationDirectory) {
String fileName_woext = (fileName.substring(0, fileName.lastIndexOf('.')));
Spawn.execute(
Settings.Programs.CONVERT.path(),
"-trim",
"-fuzz",
"5%",
"+repage",
destinationDirectory + File.separator + fileName_woext + File.separator + fileName,
destinationDirectory
+ File.separator
+ fileName_woext
+ File.separator
+ fileName_woext
+ "_trim.png");
pageX = 0;
pageY = 0;
}
// -----------------------------------------------------
// Description: Using the specified template, extract
// the template sections into PNG and TXT corresponding
// parts.
// -----------------------------------------------------
void extractWithTemplate(String templateName, String fileName, String destinationDirectory) {
// from template
int intOriginX = 0;
int intOriginY = 0;
// from template
int intX = 0;
int intY = 0;
int intWidth = 0;
int intHeight = 0;
// calculated
int viaTemplateX = 0;
int viaTemplateY = 0;
int viaTemplatePlusWidth = 0;
int viaTemplatePlusHeight = 0;
System.out.println("fileName = " + fileName);
String fileName_woext = fileName.substring(0, fileName.lastIndexOf('.'));
String pathPlusFileName_woext =
destinationDirectory + File.separator + fileName_woext + File.separator + fileName_woext;
System.out.println("pathPlusFileName_woext = " + pathPlusFileName_woext);
try {
File fXmlFile = new File(templateLocation, templateName);
DocumentBuilderFactory dbFactory = DocumentBuilderFactory.newInstance();
DocumentBuilder dBuilder = dbFactory.newDocumentBuilder();
Document doc = dBuilder.parse(fXmlFile);
doc.getDocumentElement().normalize();
System.out.println("Root element :" + doc.getDocumentElement().getNodeName());
NodeList nList = doc.getElementsByTagName("section");
Node nNode;
for (int temp = 0; temp < nList.getLength(); temp++) {
nNode = nList.item(temp);
if (nNode.getNodeType() == Node.ELEMENT_NODE) {
Element eElement = (Element) nNode;
String myX = getTagValue("x", eElement);
String myY = getTagValue("y", eElement);
String myWidth = getTagValue("width", eElement);
String myHeight = getTagValue("height", eElement);
String mySuffix = getTagValue("suffix", eElement);
intX = Integer.parseInt(myX);
intY = Integer.parseInt(myY);
intWidth = Integer.parseInt(myWidth);
intHeight = Integer.parseInt(myHeight);
viaTemplateX = intX - (intOriginX - pageX);
viaTemplateY = intY - (intOriginY - pageY);
viaTemplatePlusWidth = viaTemplateX + intWidth;
viaTemplatePlusHeight = viaTemplateY + intHeight;
Spawn.execute(
Settings.Programs.CONVERT.path(),
pathPlusFileName_woext + "_trim.png",
"-crop",
String.format("%dx%d+%d+%d", intWidth, intHeight, viaTemplateX, viaTemplateY),
"+repage",
pathPlusFileName_woext + "_" + mySuffix + ".png");
Spawn.execute(
Settings.Programs.CONVERT.path(),
pathPlusFileName_woext + "_trim.png",
"-fill",
"none",
"-stroke",
"red",
"-strokewidth",
"3",
"-draw",
String.format(
"rectangle %d,%d %d,%d",
viaTemplateX, viaTemplateY, viaTemplatePlusWidth, viaTemplatePlusHeight),
"+repage",
pathPlusFileName_woext + "_draw.png");
Spawn.execute(
Settings.Programs.TESSERACT.path(),
pathPlusFileName_woext + "_" + mySuffix + ".png",
pathPlusFileName_woext + "_" + mySuffix);
}
}
} catch (Exception e) {
e.printStackTrace();
}
}
// -----------------------------------------------------
// Description: Determine if a template exists for the
// given file and return true on success.
// -----------------------------------------------------
boolean templateExistFor(String fileName, String destinationDirectory) {
templateIdentified = false;
// from template
int intOriginX = 0;
int intOriginY = 0;
// from template
int intX = 0;
int intY = 0;
int intWidth = 0;
int intHeight = 0;
// calculated
int viaTemplateX = 0;
int viaTemplateY = 0;
int viaTemplatePlusWidth = 0;
int viaTemplatePlusHeight = 0;
String fileName_woext = fileName.substring(0, fileName.lastIndexOf('.'));
String pathPlusFileName_woext =
destinationDirectory + File.separator + fileName_woext + File.separator + fileName_woext;
try {
File folder = new File(templateLocation);
File[] listOfFiles = folder.listFiles();
long startTime = System.currentTimeMillis();
for (File file : listOfFiles) {
if (file.isFile()) {
templateName = "";
DocumentBuilderFactory dbFactory = DocumentBuilderFactory.newInstance();
DocumentBuilder dBuilder = dbFactory.newDocumentBuilder();
Document doc = dBuilder.parse(file);
doc.getDocumentElement().normalize();
System.out.println("Root element :" + doc.getDocumentElement().getNodeName());
NodeList nList = doc.getElementsByTagName("condition");
for (int temp = 0; temp < nList.getLength(); temp++) {
Node nNode = nList.item(temp);
if (nNode.getNodeType() == Node.ELEMENT_NODE) {
Element eElement = (Element) nNode;
String myX = getTagValue("x", eElement);
String myY = getTagValue("y", eElement);
String myWidth = getTagValue("width", eElement);
String myHeight = getTagValue("height", eElement);
String mySuffix = getTagValue("suffix", eElement);
String myRequirements = getTagValue("required", eElement);
intX = Integer.parseInt(myX);
intY = Integer.parseInt(myY);
intWidth = Integer.parseInt(myWidth);
intHeight = Integer.parseInt(myHeight);
viaTemplateX = intX - (intOriginX - pageX);
viaTemplateY = intY - (intOriginY - pageY);
viaTemplatePlusWidth = viaTemplateX + intWidth;
viaTemplatePlusHeight = viaTemplateY + intHeight;
Spawn.execute(
Settings.Programs.CONVERT.path(),
pathPlusFileName_woext + "_trim.png",
"-crop",
String.format("%dx%d+%d+%d", intWidth, intHeight, viaTemplateX, viaTemplateY),
"+repage",
pathPlusFileName_woext + "_" + mySuffix + ".png");
Spawn.execute(
Settings.Programs.CONVERT.path(),
pathPlusFileName_woext + "_trim.png",
"-fill",
"none",
"-stroke",
"red",
"-strokewidth",
"3",
"-draw",
String.format(
"rectangle %d,%d %d,%d",
viaTemplateX, viaTemplateY, viaTemplatePlusWidth, viaTemplatePlusHeight),
"+repage",
pathPlusFileName_woext + "_draw.png");
Spawn.execute(
Settings.Programs.TESSERACT.path(),
pathPlusFileName_woext + "_" + mySuffix + ".png",
pathPlusFileName_woext + "_" + mySuffix);
String line = ""; // String that holds current file line
String accumulate = "";
try {
FileReader input = new FileReader(pathPlusFileName_woext + "_" + mySuffix + ".txt");
BufferedReader bufRead = new BufferedReader(input);
line = bufRead.readLine();
while (line != null) {
accumulate += line;
line = bufRead.readLine();
}
bufRead.close();
input.close();
} catch (IOException e) {
e.printStackTrace();
}
String[] requirements = myRequirements.split("#");
for (String requirement : requirements) {
if (requirement.equals(accumulate.trim())) {
templateName = file.getName();
templateIdentified = true;
return templateIdentified;
}
}
}
}
}
}
// record end time
long endTime = System.currentTimeMillis();
System.out.println(
"Template Search [" + fileName + "] " + (endTime - startTime) / 1000 + " seconds");
} catch (Exception e) {
e.printStackTrace();
}
return templateIdentified;
}