|
| 1 | +package com.iusenko.subrip; |
| 2 | + |
| 3 | +import java.io.BufferedReader; |
| 4 | +import java.io.InputStream; |
| 5 | +import java.io.InputStreamReader; |
| 6 | +import java.util.ArrayList; |
| 7 | +import java.util.List; |
| 8 | +import java.util.regex.Pattern; |
| 9 | + |
| 10 | +class SubRipParser { |
| 11 | + |
| 12 | + private static final Pattern phraseNumberPattern = Pattern.compile("^\\d+"); |
| 13 | + private static final Pattern timePattern = Pattern.compile("\\d+(:\\d+)+,\\d+\\s+-->\\s+\\d+(:\\d+)+,\\d+"); |
| 14 | + private BufferedReader in; |
| 15 | + |
| 16 | + protected SubRipParser() { |
| 17 | + // for testing purpose |
| 18 | + } |
| 19 | + |
| 20 | + public SubRipParser(InputStream in) { |
| 21 | + if (in == null) { |
| 22 | + throw new IllegalArgumentException(); |
| 23 | + } |
| 24 | + this.in = new BufferedReader(new InputStreamReader(in)); |
| 25 | + } |
| 26 | + |
| 27 | + public List<Phrase> getPhrases() throws Exception { |
| 28 | + List<Phrase> phrases = new ArrayList<Phrase>(); |
| 29 | + String line; |
| 30 | + StringBuilder buffer = new StringBuilder(); |
| 31 | + Phrase phrase = new Phrase(); |
| 32 | + |
| 33 | + while ((line = in.readLine()) != null) { |
| 34 | + line = line.trim(); |
| 35 | + |
| 36 | + if (isPhraseNumber(line)) { |
| 37 | + phrase = new Phrase(); |
| 38 | + phrase.setNumber(Integer.parseInt(line)); |
| 39 | + buffer = new StringBuilder(); |
| 40 | + continue; |
| 41 | + } |
| 42 | + |
| 43 | + if (isBlank(line)) { |
| 44 | + phrase.setText(buffer.toString().trim()); |
| 45 | + boolean fromBlank = isBlank(phrase.getFromTime()); |
| 46 | + boolean toBlank = isBlank(phrase.getToTime()); |
| 47 | + boolean textBlank = isBlank(phrase.getText()); |
| 48 | + if (!fromBlank && !toBlank && !textBlank) { |
| 49 | + phrases.add(phrase); |
| 50 | + } |
| 51 | + continue; |
| 52 | + } |
| 53 | + |
| 54 | + if (isTime(line)) { |
| 55 | + phrase.setFromTime(getStartTime(line)); |
| 56 | + phrase.setToTime(getEndTime(line)); |
| 57 | + continue; |
| 58 | + } |
| 59 | + |
| 60 | + if (isSentence(line)) { |
| 61 | + buffer.append(" ").append(line).append("\n"); |
| 62 | + } |
| 63 | + } |
| 64 | + return phrases; |
| 65 | + } |
| 66 | + |
| 67 | + protected final boolean isSentence(String line) { |
| 68 | + if (isBlank(line) || isTime(line) || isPhraseNumber(line)) { |
| 69 | + return false; |
| 70 | + } |
| 71 | + return true; |
| 72 | + } |
| 73 | + |
| 74 | + protected final boolean isTime(String line) { |
| 75 | + return timePattern.matcher(line).matches(); |
| 76 | + } |
| 77 | + |
| 78 | + protected final boolean isPhraseNumber(String line) { |
| 79 | + return phraseNumberPattern.matcher(line).matches(); |
| 80 | + } |
| 81 | + |
| 82 | + protected final String getStartTime(String rawString) { |
| 83 | + return getTime(rawString, true); |
| 84 | + } |
| 85 | + |
| 86 | + protected final String getEndTime(String rawString) { |
| 87 | + return getTime(rawString, false); |
| 88 | + } |
| 89 | + |
| 90 | + protected final String getTime(String rawString, boolean getStartTime) { |
| 91 | + int index = rawString.indexOf("-->"); |
| 92 | + if (index == -1) { |
| 93 | + return ""; |
| 94 | + } |
| 95 | + |
| 96 | + rawString = rawString.replaceAll(",", "."); |
| 97 | + if (getStartTime) { |
| 98 | + return rawString.substring(0, index).trim(); |
| 99 | + } else { |
| 100 | + return rawString.substring(index + 3).trim(); |
| 101 | + } |
| 102 | + } |
| 103 | + |
| 104 | + private boolean isBlank(String string) { |
| 105 | + return string == null || "".equals(string.trim()); |
| 106 | + } |
| 107 | +} |
0 commit comments