Chris@1855: /* -*- c-basic-offset: 4 indent-tabs-mode: nil -*-  vi:set ts=8 sts=4 sw=4: */
Chris@1855: 
Chris@1855: /*
Chris@1855:     Sonic Visualiser
Chris@1855:     An audio file viewer and annotation editor.
Chris@1855:     Centre for Digital Music, Queen Mary, University of London.
Chris@1855:     
Chris@1855:     This program is free software; you can redistribute it and/or
Chris@1855:     modify it under the terms of the GNU General Public License as
Chris@1855:     published by the Free Software Foundation; either version 2 of the
Chris@1855:     License, or (at your option) any later version.  See the file
Chris@1855:     COPYING included with this distribution for more information.
Chris@1855: */
Chris@1855: 
Chris@1855: #include "TextTest.h"
Chris@1855: 
Chris@1855: #include "base/Debug.h"
Chris@1855: #include "base/StringBits.h"
Chris@1855: 
Chris@1855: #include <QFile>
Chris@1855: #include <QXmlInputSource>
Chris@1855: 
Chris@1855: bool
Chris@1855: TextTest::isApparentTextDocument(FileSource source)
Chris@1855: {
Chris@1855:     // Return true if the document can be opened and contains some
Chris@1855:     // sort of text, either UTF-8 (so it could be Turtle) or another
Chris@1855:     // encoding that is recognised as XML
Chris@1855: 
Chris@1855:     if (!source.isAvailable()) {
Chris@1855:         SVDEBUG << "NOTE: TextTest::isApparentTextDocument: Failed to retrieve document from " << source.getLocation() << endl;
Chris@1855:         return false;
Chris@1855:     }
Chris@1855: 
Chris@1855:     QFile file(source.getLocalFilename());
Chris@1855:     if (!file.open(QFile::ReadOnly)) {
Chris@1855:         SVDEBUG << "NOTE: TextTest::isApparentTextDocument: Failed to open local file from " << source.getLocalFilename() << endl;
Chris@1855:         return false;
Chris@1855:     }
Chris@1855: 
Chris@1855:     QByteArray bytes = file.read(200);
Chris@1855: 
Chris@1859:     if (StringBits::isValidUtf8(std::string(bytes.data()), true)) {
Chris@1855:         SVDEBUG << "NOTE: TextTest::isApparentTextDocument: Document appears to be UTF-8" << endl;
Chris@1855:         return true; // good enough to be worth trying to parse
Chris@1855:     }
Chris@1855: 
Chris@1855:     QXmlInputSource xmlSource;
Chris@1855:     xmlSource.setData(bytes); // guesses text encoding
Chris@1855: 
Chris@1855:     if (xmlSource.data().startsWith("<?xml")) {
Chris@1855:         SVDEBUG << "NOTE: TextTest::isApparentTextDocument: Document appears to be XML" << endl;
Chris@1855:         return true;
Chris@1855:     }
Chris@1855: 
Chris@1855:     SVDEBUG << "NOTE: TextTest::isApparentTextDocument: Document is not UTF-8 and is not XML, rejecting" << endl;
Chris@1855:     return false;
Chris@1855: }