/*
- Copyright (C) 2014 Carl Hetherington <cth@carlh.net>
+ Copyright (C) 2014-2015 Carl Hetherington <cth@carlh.net>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
*/
+/** @file src/subrip_reader.cc
+ * @brief SubripReader class.
+ */
+
#include "subrip_reader.h"
#include "exceptions.h"
+#include "util.h"
#include <boost/algorithm/string.hpp>
#include <boost/lexical_cast.hpp>
+#include <boost/regex.hpp>
+#include <boost/bind.hpp>
#include <cstdio>
#include <vector>
+#include <iostream>
using std::string;
using std::vector;
+using std::list;
+using std::cout;
+using std::hex;
+using std::stringstream;
using boost::lexical_cast;
+using boost::to_upper;
+using boost::optional;
+using boost::function;
using namespace sub;
+/** @param s Subtitle string encoded in UTF-8 */
+SubripReader::SubripReader (string const & s)
+{
+ stringstream str (s);
+ this->read (boost::bind (&get_line_stringstream, &str));
+}
+
+/** @param f Subtitle file encoded in UTF-8 */
SubripReader::SubripReader (FILE* f)
+{
+ this->read (boost::bind (&get_line_file, f));
+}
+
+void
+SubripReader::read (function<optional<string> ()> get_line)
{
enum {
COUNTER,
CONTENT
} state = COUNTER;
- char buffer[256];
-
- TimePair from;
- TimePair to;
+ RawSubtitle rs;
- string line;
- int line_number = 0;
+ /* This reader extracts no information about where the subtitle
+ should be on screen, so its reference is TOP_OF_SUBTITLE.
+ */
+ rs.vertical_position.line = 0;
+ rs.vertical_position.reference = TOP_OF_SUBTITLE;
- while (!feof (f)) {
- char* r = fgets (buffer, sizeof (buffer), f);
- if (r == 0 || feof (f)) {
+ while (true) {
+ optional<string> line = get_line ();
+ if (!line) {
break;
}
- line = string (buffer);
- trim_right_if (line, boost::is_any_of ("\n\r"));
+ trim_right_if (*line, boost::is_any_of ("\n\r"));
+ remove_unicode_bom (line);
switch (state) {
case COUNTER:
{
- if (line.empty ()) {
+ if (line->empty ()) {
/* a blank line at the start is ok */
break;
}
state = METADATA;
+
+ /* Reset stuff that should not persist across separate subtitles */
+ rs.bold = false;
+ rs.italic = false;
+ rs.underline = false;
+ rs.vertical_position.line = 0;
}
break;
case METADATA:
{
vector<string> p;
- boost::algorithm::split (p, line, boost::algorithm::is_any_of (" "));
+
+ /* Further trim this line, removing spaces from the end */
+ trim_right_if (*line, boost::is_any_of (" "));
+
+ boost::algorithm::split (p, *line, boost::algorithm::is_any_of (" "));
if (p.size() != 3 && p.size() != 7) {
- throw SubripError (line, "a time/position line");
+ throw SubripError (*line, "a time/position line");
}
- from = convert_time (p[0]);
- to = convert_time (p[2]);
+ rs.from = convert_time (p[0]);
+ rs.to = convert_time (p[2]);
/* XXX: should not ignore coordinate specifications */
-
+
state = CONTENT;
break;
}
case CONTENT:
- if (line.empty ()) {
+ if (line->empty ()) {
state = COUNTER;
- line_number = 0;
} else {
- convert_line (line, line_number, from, to);
- line_number++;
+ convert_line (*line, rs);
+ rs.vertical_position.line = rs.vertical_position.line.get() + 1;
}
break;
}
}
}
-TimePair
+Time
SubripReader::convert_time (string t)
{
vector<string> a;
vector<string> b;
boost::algorithm::split (b, a[2], boost::is_any_of (","));
- return TimePair (
- MetricTime (
- lexical_cast<int> (a[0]),
- lexical_cast<int> (a[1]),
- lexical_cast<int> (b[0]),
- lexical_cast<int> (b[1])
- )
+ return Time::from_hms (
+ lexical_cast<int> (a[0]),
+ lexical_cast<int> (a[1]),
+ lexical_cast<int> (b[0]),
+ lexical_cast<int> (b[1])
);
}
void
-SubripReader::convert_line (string t, int line_number, TimePair from, TimePair to)
+SubripReader::convert_line (string t, RawSubtitle& p)
{
enum {
TEXT,
TAG
} state = TEXT;
-
+
string tag;
- RawSubtitle p;
- p.font = "Arial";
- p.font_size.set_points (48);
- p.from = from;
- p.to = to;
- p.vertical_position.line = line_number;
- p.vertical_position.reference = TOP_OF_SUBTITLE;
-
+ list<Colour> colours;
+ colours.push_back (Colour (1, 1, 1));
+
/* XXX: missing <font> support */
/* XXX: nesting of tags e.g. <b>foo<i>bar<b>baz</b>fred</i>jim</b> might
not work, I think.
} else if (tag == "/u") {
maybe_content (p);
p.underline = false;
+ } else if (boost::starts_with (tag, "font")) {
+ maybe_content (p);
+ boost::regex re (".*color=\"#([0123456789abcdef]+)\"");
+ boost::smatch match;
+ if (boost::regex_search (tag, match, re) && string (match[1]).size() == 6) {
+ p.colour = Colour::from_rgb_hex (match[1]);
+ colours.push_back (p.colour);
+ }
+ } else if (tag == "/font") {
+ maybe_content (p);
+ colours.pop_back ();
+ p.colour = colours.back ();
}
tag.clear ();
state = TEXT;
maybe_content (p);
}
+/* Push p into _subs if it has some text, and clear the text out of p */
void
SubripReader::maybe_content (RawSubtitle& p)
{