dryark/strawberry
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

LyricsProvider: Improve parsing from HTML function

Browse Source
This commit is contained in:
Jonas Kvinge
2023-01-20 22:45:05 +01:00
parent 536fe637aa
commit b982a6a762
1 changed files with 21 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
src/lyrics/lyricsprovider.cpp
View File

@@ -40,26 +40,19 @@ QString LyricsProvider::ParseLyricsFromHTML(const QString &content, const QRegul
QRegularExpressionMatch rematch = lyrics_start.match(content, start_idx); QRegularExpressionMatch rematch = lyrics_start.match(content, start_idx);
if (!rematch.hasMatch()) break; if (!rematch.hasMatch()) break;
const qint64 start_lyrics_tag_idx = rematch.capturedEnd(); const qint64 start_lyrics_idx = rematch.capturedEnd();
qint64 end_lyrics_idx = -1;
// Find the index of the end tag. // Find the index of the end tag.
qint64 start_tag_idx = 0; qint64 idx = start_lyrics_idx;
qint64 end_tag_idx = 0; QRegularExpressionMatch rematch_start_tag;
qint64 end_tag_length = 0; QRegularExpressionMatch rematch_end_tag;
qint64 idx = start_lyrics_tag_idx;
int tags = 1; int tags = 1;
do { do {
QRegularExpressionMatch rematch_start_tag = QRegularExpression(start_tag).match(content, idx); rematch_start_tag = QRegularExpression(start_tag).match(content, idx);
start_tag_idx = rematch_start_tag.hasMatch() ? rematch_start_tag.capturedStart() : -1; const qint64 start_tag_idx = rematch_start_tag.hasMatch() ? rematch_start_tag.capturedStart() : -1;
QRegularExpressionMatch rematch_end_tag = QRegularExpression(end_tag).match(content, idx); rematch_end_tag = QRegularExpression(end_tag).match(content, idx);
if (rematch_end_tag.hasMatch()) { const qint64 end_tag_idx = rematch_end_tag.hasMatch() ? rematch_end_tag.capturedStart() : -1;
end_tag_idx = rematch_end_tag.capturedStart();
end_tag_length = rematch_end_tag.capturedLength();
}
else {
end_tag_idx = -1;
end_tag_length = 0;
}
if (rematch_start_tag.hasMatch() && start_tag_idx <= end_tag_idx) { if (rematch_start_tag.hasMatch() && start_tag_idx <= end_tag_idx) {
++tags; ++tags;
idx = start_tag_idx + rematch_start_tag.capturedLength(); idx = start_tag_idx + rematch_start_tag.capturedLength();
@@ -67,21 +60,26 @@ QString LyricsProvider::ParseLyricsFromHTML(const QString &content, const QRegul
else if (rematch_end_tag.hasMatch()) { else if (rematch_end_tag.hasMatch()) {
--tags; --tags;
idx = end_tag_idx + rematch_end_tag.capturedLength(); idx = end_tag_idx + rematch_end_tag.capturedLength();
if (tags == 0) {
end_lyrics_idx = rematch_end_tag.capturedStart();
start_idx = rematch_end_tag.capturedEnd();
}
} }
} }
while (tags > 0 || end_tag_idx >= start_tag_idx); while (tags > 0 && (rematch_start_tag.hasMatch() || rematch_end_tag.hasMatch()));
start_idx = end_tag_idx + end_tag_length; if (end_lyrics_idx != -1 && start_lyrics_idx < end_lyrics_idx) {
if (end_tag_idx > 0 || start_lyrics_tag_idx < end_tag_idx) {
if (!lyrics.isEmpty()) { if (!lyrics.isEmpty()) {
lyrics.append("\n"); lyrics.append("\n");
} }
lyrics.append(content.mid(start_lyrics_tag_idx, end_tag_idx - start_lyrics_tag_idx) lyrics.append(content.mid(start_lyrics_idx, end_lyrics_idx - start_lyrics_idx)
.replace(QRegularExpression("<br[^>]+>"), "\n") .replace(QRegularExpression("<br[^>]*>"), "\n")
.remove(QRegularExpression("<[^>]*>")) .remove(QRegularExpression("<[^>]*>"))
.trimmed()); .trimmed());
} }
else {
start_idx = -1;
}
} }
while (start_idx > 0 && multiple); while (start_idx > 0 && multiple);
@@ -93,4 +91,3 @@ QString LyricsProvider::ParseLyricsFromHTML(const QString &content, const QRegul
return Utilities::DecodeHtmlEntities(lyrics); return Utilities::DecodeHtmlEntities(lyrics);
} }