added check if letter and build Hashmap counter for existing words
parent
21546dd91d
commit
b0cc3a5ce9
|
@ -2,40 +2,100 @@ package domain;
|
|||
|
||||
import org.apache.pdfbox.pdmodel.PDDocument;
|
||||
import org.apache.pdfbox.text.PDFTextStripper;
|
||||
|
||||
import javax.swing.*;
|
||||
import javax.swing.filechooser.FileNameExtensionFilter;
|
||||
import java.io.File;
|
||||
|
||||
import java.awt.*;
|
||||
import java.io.*;
|
||||
import java.util.HashMap;
|
||||
import java.util.Map;
|
||||
|
||||
|
||||
public class FileManager {
|
||||
File inputFile;
|
||||
String originalPath;
|
||||
String goalPath;
|
||||
|
||||
public void loadFile() {
|
||||
try{
|
||||
public FileManager() {
|
||||
originalPath = "quelle.pdf";
|
||||
goalPath = "ziel.txt";
|
||||
inputFile = null;
|
||||
}
|
||||
|
||||
public OutputStream loadFilePath() {
|
||||
InputStream in;
|
||||
OutputStream out = null;
|
||||
|
||||
try {
|
||||
in = new FileInputStream(originalPath);
|
||||
out = new FileOutputStream(goalPath);
|
||||
|
||||
byte[] buffer = new byte[1024];
|
||||
int gelesen;
|
||||
|
||||
while ((gelesen = in.read(buffer)) > -1) {
|
||||
out.write(buffer, 0, gelesen);
|
||||
}
|
||||
|
||||
in.close();
|
||||
out.close();
|
||||
return out;
|
||||
}
|
||||
catch (IOException e) {
|
||||
e.printStackTrace();
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
public File loadFileGUI() {
|
||||
try {
|
||||
JFileChooser fileChooser = new JFileChooser();
|
||||
fileChooser.setFileFilter(new FileNameExtensionFilter("PDF Files", "pdf"));
|
||||
|
||||
int result = fileChooser.showOpenDialog(null);
|
||||
|
||||
if (result == JFileChooser.APPROVE_OPTION) {
|
||||
File selectedFile = fileChooser.getSelectedFile();
|
||||
System.out.println("Ausgewählte Datei: " + selectedFile.getAbsolutePath());
|
||||
inputFile = fileChooser.getSelectedFile();
|
||||
}
|
||||
return inputFile;
|
||||
} catch (HeadlessException e) {
|
||||
throw new RuntimeException(e);
|
||||
}
|
||||
}
|
||||
|
||||
try (PDDocument document = PDDocument.load(selectedFile)) {
|
||||
PDFTextStripper pdfStripper = new PDFTextStripper();
|
||||
String text = pdfStripper.getText(document);
|
||||
String regex = "[,\\s\\.:/]";
|
||||
String[] words = text.split(regex);
|
||||
for(String word : words) {
|
||||
System.out.println(word);
|
||||
public boolean processFile(File inputFile){
|
||||
boolean success = false;
|
||||
try {
|
||||
PDDocument document = null;
|
||||
HashMap<String, Integer> filteredWords = new HashMap<>();
|
||||
if(inputFile != null) {
|
||||
document = PDDocument.load(inputFile);
|
||||
PDFTextStripper pdfStripper = new PDFTextStripper();
|
||||
String text = pdfStripper.getText(document);
|
||||
String splittedText = "[,\\s\\.:/!§$%&/()=?+*~#.;_<>^°]";
|
||||
String[] textWords = text.split(splittedText);
|
||||
for(String word : textWords){
|
||||
// if(Character.isLetter(word.charAt(0)) && Character.isLetter((word.charAt(word.length() - 1)))) {
|
||||
if (filteredWords.containsKey(word)) {
|
||||
filteredWords.compute(word, (k, counter) -> counter + 1);
|
||||
}
|
||||
// }
|
||||
else {
|
||||
filteredWords.put(word, 0);
|
||||
}
|
||||
}
|
||||
for(Map.Entry e : filteredWords.entrySet()){
|
||||
System.out.println(e.getKey() + " = " + e.getValue());
|
||||
}
|
||||
if (document != null) {
|
||||
document.close();
|
||||
}
|
||||
}
|
||||
else {
|
||||
System.out.println("Dateiauswahl abgebrochen.");
|
||||
}
|
||||
|
||||
} catch (Exception e) {
|
||||
success = true;
|
||||
} catch (Exception e){
|
||||
e.printStackTrace();
|
||||
}
|
||||
return success;
|
||||
}
|
||||
|
||||
public void saveFile(){}
|
||||
|
|
|
@ -1,4 +1,30 @@
|
|||
package facade;
|
||||
|
||||
import domain.FileManager;
|
||||
import domain.PictureManager;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.OutputStream;
|
||||
|
||||
public class WordCloudManager {
|
||||
FileManager fileManager;
|
||||
PictureManager pictureManager;
|
||||
|
||||
public WordCloudManager() {
|
||||
fileManager = new FileManager();
|
||||
pictureManager = new PictureManager();
|
||||
}
|
||||
|
||||
public boolean loadFileGUI() {
|
||||
boolean success = false;
|
||||
File inputFile = fileManager.loadFileGUI();
|
||||
success = fileManager.processFile(inputFile);
|
||||
return success;
|
||||
}
|
||||
|
||||
public void loadFilePath() {
|
||||
OutputStream inputFile = fileManager.loadFilePath();
|
||||
// fileManager.processFile(null, inputFile);
|
||||
}
|
||||
}
|
||||
|
|
|
@ -1,30 +1,52 @@
|
|||
package tui;
|
||||
|
||||
import domain.FileManager;
|
||||
import facade.WordCloudManager;
|
||||
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.IOException;
|
||||
import java.util.Scanner;
|
||||
|
||||
public class TUI {
|
||||
boolean isRunning;
|
||||
|
||||
public TUI(){
|
||||
isRunning = true;
|
||||
tui();
|
||||
}
|
||||
|
||||
public void tui() {
|
||||
Scanner scan = new Scanner(System.in);
|
||||
int option = scan.nextInt();
|
||||
while(isRunning) {
|
||||
System.out.println("Welcome to Word Cloud.\nMenu:\n\n(0) Load File from main path\n(1) Load File with Gui" +
|
||||
"\n(2) Save File\n(3) Show Picture\n(4) Exit");
|
||||
int option = scan.nextInt();
|
||||
WordCloudManager wcm = new WordCloudManager();
|
||||
|
||||
switch (option) {
|
||||
case (0):
|
||||
//Load File
|
||||
FileManager fileManager = new FileManager();
|
||||
fileManager.loadFile();
|
||||
case (1):
|
||||
//Save File
|
||||
case (2):
|
||||
//Show Picture
|
||||
case (3):
|
||||
//Exit
|
||||
switch (option) {
|
||||
case (0):
|
||||
//Load File Path
|
||||
wcm.loadFilePath();
|
||||
break;
|
||||
case (1):
|
||||
//Load File GUI
|
||||
if(wcm.loadFileGUI()) {
|
||||
System.out.println("File loaded successful!\n");
|
||||
}
|
||||
break;
|
||||
case (2):
|
||||
//Save Picture
|
||||
break;
|
||||
case (3):
|
||||
//Show Picture
|
||||
break;
|
||||
case (4):
|
||||
//Exit
|
||||
isRunning = false;
|
||||
System.out.println("Close Program!");
|
||||
break;
|
||||
}
|
||||
}
|
||||
scan.close();
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in New Issue