DenysMedvid/pdf2htmlEX
1
0
Fork 0
mirror of https://github.com/pdf2htmlEX/pdf2htmlEX.git synced 2024-12-22 04:50:09 +00:00
Code

working on linear gradient

Browse Source
This commit is contained in:
Lu Wang 2012-10-03 02:19:40 +08:00
parent 6ae97943df
commit 2041c2d146
15 changed files with 261 additions and 636 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
CMakeLists.txt
View File

@ -20,14 +20,20 @@ include_directories(${POPPLER_INCLUDE_DIRS})
link_directories(${POPPLER_LIBRARY_DIRS})
set(PDF2HTMLEX_LIBS ${PDF2HTMLEX_LIBS} ${POPPLER_LIBRARIES})
find_path(CAIRO_OUTPUTDEV_INCLUDE_PATH CairoOutputDev.h PATHS
${POPPLER_INCLUDE_DIRS} NO_DEFAULT_PATH)
if(CAIRO_OUTPUTDEV_INCLUDE_PATH)
message("Found CairoOutputDev.h: ${POPPLER_INCLUDE_DIRS}/CairoOutputDev.h")
# disable CAIRO for now
if(0)
pkg_check_modules(POPPLER_CAIRO poppler-cairo>=0.20.0)
if(POPPLER_CAIRO_FOUND)
set(HAVE_CAIRO 1)
include_directories(${POPPLER_CAIRO_INCLUDE_DIRS})
link_directories(${POPPLER_CAIRO_LIBRARY_DIRS})
set(PDF2HTMLEX_LIBS ${PDF2HTMLEX_LIBS} ${POPPLER_CAIRO_LIBRARIES})
else()
set(HAVE_CAIRO 0)
endif()
else()
set(HAVE_CAIRO 0)
endif()
# fontforge starts using pkg-config 'correctly' since 2.0.0

4
src/CairoBackgroundRenderer.cc
View File

@ -20,6 +20,10 @@ void CairoBackgroundRenderer::drawChar(GfxState *state, double x, double y,
// CairoOutputDev::drawChar(state,x,y,dx,dy,originX,originY,code, nBytes, u, uLen);
}
void CairoBackgroundRenderer::render_page(PDFDoc * doc, int pageno, const std::string & filename)
{
}
} // namespace pdf2htmlEX
#endif // HAVE_CAIRO

4
src/HTMLRenderer/LineBuffer.cc
View File

@ -75,7 +75,7 @@ void HTMLRenderer::LineBuffer::flush(void)
for(auto iter = states.begin(); iter != states.end(); ++iter)
{
const auto & s = *iter;
max_ascent = max(max_ascent, s.ascent * s.draw_font_size);
max_ascent = max<double>(max_ascent, s.ascent * s.draw_font_size);
}
ostream & out = renderer->html_fout;
@ -155,7 +155,7 @@ void HTMLRenderer::LineBuffer::flush(void)
++ cur_offset_iter;
}
size_t next_text_idx = min(cur_state_iter->start_idx, cur_offset_iter->start_idx);
size_t next_text_idx = min<size_t>(cur_state_iter->start_idx, cur_offset_iter->start_idx);
outputUnicodes(out, (&text.front()) + cur_text_idx, next_text_idx - cur_text_idx);
cur_text_idx = next_text_idx;

202
src/HTMLRenderer/draw.cc
View File

@ -7,6 +7,12 @@
* 2012.10.01
*/
#include <algorithm>
#include <cmath>
#include <sstream>
#include <vector>
#include <iostream>
#include "HTMLRenderer.h"
#include "util.h"
#include "namespace.h"
@ -14,6 +20,14 @@
namespace pdf2htmlEX {
using std::swap;
using std::min;
using std::max;
using std::acos;
using std::asin;
using std::ostringstream;
using std::sqrt;
using std::vector;
using std::ostream;
static bool is_horizontal_line(GfxSubpath * path)
{
@ -44,6 +58,157 @@ static bool is_rectangle(GfxSubpath * path)
&& _equal(path->getY(3), path->getY(0)));
}
static void get_shading_bbox(GfxState * state, GfxShading * shading,
double & x1, double & y1, double & x2, double & y2)
{
// from SplashOutputDev.cc in poppler
if(shading->getHasBBox())
{
shading->getBBox(&x1, &y1, &x2, &y2);
}
else
{
state->getClipBBox(&x1, &y1, &x2, &y2);
Matrix ctm, ictm;
state->getCTM(&ctm);
ctm.invertTo(&ictm);
double x[4], y[4];
ictm.transform(x1, y1, &x[0], &y[0]);
ictm.transform(x2, y1, &x[1], &y[1]);
ictm.transform(x1, y2, &x[2], &y[2]);
ictm.transform(x2, y2, &x[3], &y[3]);
x1 = x2 = x[0];
y1 = y2 = y[0];
for(int i = 1; i < 4; ++i)
{
x1 = min<double>(x1, x[i]);
y1 = min<double>(y1, y[i]);
x2 = max<double>(x2, x[i]);
y2 = max<double>(y2, y[i]);
}
}
}
static double get_degree(double dx, double dy)
{
static const double PI = acos(-1.0);
double r = hypot(dx, dy);
double ang = acos(dx / r);
if(!_is_positive(dy))
ang = 2 * PI - ang;
return ang * 180.0 / PI;
}
class LinearGradient
{
public:
LinearGradient(GfxAxialShading * shading,
double x1, double y1, double x2, double y2);
void dumpto (ostream & out);
static void style_function (void * p, ostream & out)
{
static_cast<LinearGradient*>(p)->dumpto(out);
}
// TODO, add alpha
class ColorStop
{
public:
GfxRGB rgb;
double pos; // [0,1]
};
vector<ColorStop> stops;
double degree;
};
LinearGradient::LinearGradient (GfxAxialShading * shading,
double x1, double y1, double x2, double y2)
{
// coordinate for t = 0 and t = 1
double t0x, t0y, t1x, t1y;
shading->getCoords(&t0x, &t0y, &t1x, &t1y);
degree = get_degree(t1x - t0x, t1y - t0y);
// get the range of t in the box
// from GfxState.cc in poppler
double box_tmin, box_tmax;
{
double idx = t1x - t0x;
double idy = t1y - t0y;
double inv_len = 1.0 / (idx * idx + idy * idy);
idx *= inv_len;
idy *= inv_len;
// t of (x1,y1)
box_tmin = box_tmax = (x1 - t0x) * idx + (y1 - t0y) * idy;
double tdx = (x2 - x1) * idx;
if(tdx < 0)
box_tmin += tdx;
else
box_tmax += tdx;
double tdy = (y2 - y1) * idy;
if(tdy < 0)
box_tmin += tdy;
else
box_tmax += tdy;
}
// get the domain of t in the box
double domain_tmin = max<double>(box_tmin, shading->getDomain0());
double domain_tmax = min<double>(box_tmax, shading->getDomain1());
// TODO: better sampling
// TODO: check background color
{
stops.clear();
double tstep = (domain_tmax - domain_tmin) / 13.0;
for(double t = domain_tmin; t <= domain_tmax; t += tstep)
{
GfxColor color;
shading->getColor(t, &color);
ColorStop stop;
shading->getColorSpace()->getRGB(&color, &stop.rgb);
stop.pos = (t - box_tmin) / (box_tmax - box_tmin);
stops.push_back(stop);
}
}
}
void LinearGradient::dumpto (ostream & out)
{
out << "background-color:red;";
}
GBool HTMLRenderer::axialShadedFill(GfxState *state, GfxAxialShading *shading, double tMin, double tMax)
{
if(!(param->css_draw)) return gFalse;
double x1, y1, x2, y2;
get_shading_bbox(state, shading, x1, y1, x2, y2);
LinearGradient lg(shading, x1, y1, x2, y2);
// TODO: check background color
css_draw_rectangle(x1, y1, x2-x1, y2-y1, state->getCTM(),
nullptr, 0,
nullptr, nullptr,
LinearGradient::style_function, &lg);
return gTrue;
}
//TODO track state
//TODO connection style
void HTMLRenderer::css_draw(GfxState *state, bool fill)
@ -67,9 +232,9 @@ void HTMLRenderer::css_draw(GfxState *state, bool fill)
double lw = state->getLineWidth();
css_draw_rectangle(x1, y - lw/2, x2-x1, lw,
css_draw_rectangle(x1, y - lw/2, x2-x1, lw, state->getCTM(),
nullptr, 0,
nullptr, &stroke_color, state);
nullptr, &stroke_color);
}
else if (is_rectangle(subpath))
{
@ -108,34 +273,36 @@ void HTMLRenderer::css_draw(GfxState *state, bool fill)
w += lw[1];
}
css_draw_rectangle(x, y, w, h,
css_draw_rectangle(x, y, w, h, state->getCTM(),
lw, lw_count,
ps, pf, state);
ps, pf);
}
}
}
void HTMLRenderer::css_draw_rectangle(double x, double y, double w, double h,
void HTMLRenderer::css_draw_rectangle(double x, double y, double w, double h, const double * tm,
double * line_width_array, int line_width_count,
const GfxRGB * line_color, const GfxRGB * fill_color,
GfxState * state)
const GfxRGB * line_color, const GfxRGB * fill_color,
void (*style_function)(void *, ostream &), void * style_function_data)
{
close_text_line();
double ctm[6];
memcpy(ctm, state->getCTM(), sizeof(ctm));
double new_tm[6];
memcpy(new_tm, tm, sizeof(new_tm));
_transform(ctm, x, y);
_transform(new_tm, x, y);
double scale = 1.0;
{
double i1 = ctm[0] + ctm[2];
double i2 = ctm[1] + ctm[3];
scale = sqrt((i1 * i1 + i2 * i2) / 2.0);
static const double sqrt2 = sqrt(2.0);
double i1 = (new_tm[0] + new_tm[2]) / sqrt2;
double i2 = (new_tm[1] + new_tm[3]) / sqrt2;
scale = hypot(i1, i2);
if(_is_positive(scale))
{
for(int i = 0; i < 4; ++i)
ctm[i] /= scale;
new_tm[i] /= scale;
}
else
{
@ -143,7 +310,7 @@ void HTMLRenderer::css_draw_rectangle(double x, double y, double w, double h,
}
}
html_fout << "<div class=\"Cd t" << install_transform_matrix(ctm) << "\" style=\"";
html_fout << "<div class=\"Cd t" << install_transform_matrix(new_tm) << "\" style=\"";
if(line_color)
{
@ -174,6 +341,11 @@ void HTMLRenderer::css_draw_rectangle(double x, double y, double w, double h,
html_fout << "background-color:transparent;";
}
if(style_function)
{
style_function(style_function_data, html_fout);
}
html_fout << "bottom:" << _round(y) << "px;"
<< "left:" << _round(x) << "px;"
<< "width:" << _round(w * scale) << "px;"

2
src/HTMLRenderer/general.cc
View File

@ -102,7 +102,7 @@ void HTMLRenderer::process(PDFDoc *doc)
zoom = *min_element(zoom_factors.begin(), zoom_factors.end());
}
text_scale_factor1 = max(zoom, param->font_size_multiplier);
text_scale_factor1 = max<double>(zoom, param->font_size_multiplier);
text_scale_factor2 = zoom / text_scale_factor1;
}

6
src/HTMLRenderer/install.cc
View File

@ -215,9 +215,11 @@ long long HTMLRenderer::install_font_size(double font_size)
long long HTMLRenderer::install_transform_matrix(const double * tm)
{
TM m(tm);
Matrix m;
memcpy(m.m, tm, sizeof(m.m));
auto iter = transform_matrix_map.lower_bound(m);
if((iter != transform_matrix_map.end()) && (m == (iter->first)))
if((iter != transform_matrix_map.end()) && (_tm_equal(m.m, iter->first.m, 4)))
return iter->second;
long long new_tm_id = transform_matrix_map.size();

8
src/HTMLRenderer/link.cc
View File

@ -191,10 +191,10 @@ void HTMLRenderer::processLink(AnnotLink * al)
double x,y,w,h;
double x1, y1, x2, y2;
al->getRect(&x1, &y1, &x2, &y2);
x = min(x1, x2);
y = min(y1, y2);
w = max(x1, x2) - x;
h = max(y1, y2) - y;
x = min<double>(x1, x2);
y = min<double>(y1, y2);
w = max<double>(x1, x2) - x;
h = max<double>(y1, y2) - y;
double border_width = 0;
double border_top_bottom_width = 0;

20
src/HTMLRenderer/state.cc
View File

@ -12,6 +12,7 @@
* optimize lines using nested <span> (reuse classes)
*/
#include <cmath>
#include <algorithm>
#include "HTMLRenderer.h"
@ -22,6 +23,7 @@ namespace pdf2htmlEX {
using std::max;
using std::abs;
using std::hypot;
void HTMLRenderer::updateAll(GfxState * state)
{
@ -98,7 +100,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
if(!(new_font_info->id == cur_font_info->id))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
cur_font_info = new_font_info;
}
@ -146,7 +148,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
double new_draw_text_tm[6];
memcpy(new_draw_text_tm, cur_text_tm, sizeof(new_draw_text_tm));
double new_draw_text_scale = 1.0/text_scale_factor2 * sqrt(new_draw_text_tm[2] * new_draw_text_tm[2] + new_draw_text_tm[3] * new_draw_text_tm[3]);
double new_draw_text_scale = 1.0/text_scale_factor2 * hypot(new_draw_text_tm[2], new_draw_text_tm[3]);
double new_draw_font_size = cur_font_size;
if(_is_positive(new_draw_text_scale))
@ -168,13 +170,13 @@ void HTMLRenderer::check_state_change(GfxState * state)
if(!(_equal(new_draw_font_size, draw_font_size)))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
draw_font_size = new_draw_font_size;
cur_fs_id = install_font_size(draw_font_size);
}
if(!(_tm_equal(new_draw_text_tm, draw_text_tm, 4)))
{
new_line_state = max(new_line_state, NLS_DIV);
new_line_state = max<NewLineState>(new_line_state, NLS_DIV);
memcpy(draw_text_tm, new_draw_text_tm, sizeof(draw_text_tm));
cur_ttm_id = install_transform_matrix(draw_text_tm);
}
@ -236,7 +238,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
if(!merged)
{
new_line_state = max(new_line_state, NLS_DIV);
new_line_state = max<NewLineState>(new_line_state, NLS_DIV);
}
}
@ -247,7 +249,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
double new_letter_space = state->getCharSpace();
if(!_equal(cur_letter_space, new_letter_space))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
cur_letter_space = new_letter_space;
cur_ls_id = install_letter_space(cur_letter_space * draw_text_scale);
}
@ -260,7 +262,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
double new_word_space = state->getWordSpace();
if(!_equal(cur_word_space, new_word_space))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
cur_word_space = new_word_space;
cur_ws_id = install_word_space(cur_word_space * draw_text_scale);
}
@ -273,7 +275,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
state->getFillRGB(&new_color);
if(!((new_color.r == cur_color.r) && (new_color.g == cur_color.g) && (new_color.b == cur_color.b)))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
cur_color = new_color;
cur_color_id = install_color(&new_color);
}
@ -286,7 +288,7 @@ void HTMLRenderer::check_state_change(GfxState * state)
double new_rise = state->getRise();
if(!_equal(cur_rise, new_rise))
{
new_line_state = max(new_line_state, NLS_SPAN);
new_line_state = max<NewLineState>(new_line_state, NLS_SPAN);
cur_rise = new_rise;
cur_rise_id = install_rise(new_rise * draw_text_scale);
}

2
src/HTMLRenderer/text.cc
View File

@ -310,7 +310,7 @@ void HTMLRenderer::embed_font(const string & filepath, GfxFont * font, FontInfo
memset(width_list, -1, 0x10000 * sizeof(*width_list));
if(code2GID)
maxcode = min(maxcode, code2GID_len - 1);
maxcode = min<int>(maxcode, code2GID_len - 1);
bool is_truetype = is_truetype_suffix(suffix);
int max_key = maxcode;

567
src/HTMLRenderer/text.cc.orig
View File

@ -1,567 +0,0 @@
/*
* text.cc
*
* Handling text & font, and relative stuffs
*
* by WangLu
* 2012.08.14
*/
#include <iostream>
#include <algorithm>
#include <unordered_set>
#include <cctype>
#include <cmath>
#include <CharCodeToUnicode.h>
#include <fofi/FoFiTrueType.h>
#include "ffw.h"
#include "HTMLRenderer.h"
#include "namespace.h"
namespace pdf2htmlEX {
using std::unordered_set;
using std::min;
using std::all_of;
using std::floor;
using std::swap;
string HTMLRenderer::dump_embedded_font (GfxFont * font, long long fn_id)
{
Object obj, obj1, obj2;
Object font_obj, font_obj2, fontdesc_obj;
string suffix;
string filepath;
try
{
// mupdf consulted
string subtype;
auto * id = font->getID();
Object ref_obj;
ref_obj.initRef(id->num, id->gen);
ref_obj.fetch(xref, &font_obj);
ref_obj.free();
if(!font_obj.isDict())
{
cerr << "Font object is not a dictionary" << endl;
throw 0;
}
Dict * dict = font_obj.getDict();
if(dict->lookup("DescendantFonts", &font_obj2)->isArray())
{
if(font_obj2.arrayGetLength() == 0)
{
cerr << "Warning: empty DescendantFonts array" << endl;
}
else
{
if(font_obj2.arrayGetLength() > 1)
cerr << "TODO: multiple entries in DescendantFonts array" << endl;
if(font_obj2.arrayGet(0, &obj2)->isDict())
{
dict = obj2.getDict();
}
}
}
if(!dict->lookup("FontDescriptor", &fontdesc_obj)->isDict())
{
cerr << "Cannot find FontDescriptor " << endl;
throw 0;
}
dict = fontdesc_obj.getDict();
if(dict->lookup("FontFile3", &obj)->isStream())
{
if(obj.streamGetDict()->lookup("Subtype", &obj1)->isName())
{
subtype = obj1.getName();
if(subtype == "Type1C")
{
suffix = ".cff";
}
else if (subtype == "CIDFontType0C")
{
suffix = ".cid";
}
else
{
cerr << "Unknown subtype: " << subtype << endl;
throw 0;
}
}
else
{
cerr << "Invalid subtype in font descriptor" << endl;
throw 0;
}
}
else if (dict->lookup("FontFile2", &obj)->isStream())
{
suffix = ".ttf";
}
else if (dict->lookup("FontFile", &obj)->isStream())
{
suffix = ".pfa";
}
else
{
cerr << "Cannot find FontFile for dump" << endl;
throw 0;
}
if(suffix == "")
{
cerr << "Font type unrecognized" << endl;
throw 0;
}
obj.streamReset();
filepath = (char*)str_fmt("%s/f%llx%s", param->tmp_dir.c_str(), fn_id, suffix.c_str());
add_tmp_file(filepath);
ofstream outf(filepath, ofstream::binary);
if(!outf)
throw string("Cannot open file ") + filepath + " for writing";
char buf[1024];
int len;
while((len = obj.streamGetChars(1024, (Guchar*)buf)) > 0)
{
outf.write(buf, len);
}
outf.close();
obj.streamClose();
}
catch(int)
{
cerr << "Someting wrong when trying to dump font " << hex << fn_id << dec << endl;
}
obj2.free();
obj1.free();
obj.free();
fontdesc_obj.free();
font_obj2.free();
font_obj.free();
return filepath;
}
void HTMLRenderer::embed_font(const string & filepath, GfxFont * font, FontInfo & info, bool get_metric_only)
{
if(param->debug)
{
cerr << "Embed font: " << filepath << " " << info.id << endl;
}
ffw_load_font(filepath.c_str());
int * code2GID = nullptr;
int code2GID_len = 0;
int maxcode = 0;
Gfx8BitFont * font_8bit = nullptr;
GfxCIDFont * font_cid = nullptr;
string suffix = get_suffix(filepath);
for(auto iter = suffix.begin(); iter != suffix.end(); ++iter)
*iter = tolower(*iter);
/*
* if parm->tounicode is 0, try the provided tounicode map first
*/
info.use_tounicode = (is_truetype_suffix(suffix) || (param->tounicode >= 0));
info.has_space = false;
const char * used_map = nullptr;
info.em_size = ffw_get_em_size();
if(get_metric_only)
return;
used_map = preprocessor.get_code_map(hash_ref(font->getID()));
/*
* Step 1
* dump the font file directly from the font descriptor and put the glyphs into the correct slots
*
* for 8bit + nonTrueType
* re-encoding the font using a PostScript encoding list (glyph id <-> glpyh name)
*
* for 8bit + TrueType
* sort the glpyhs as the original order, and later will map GID (instead of char code) to Unicode
*
* for CID + nonTrueType
* Flatten the font
*
* for CID Truetype
* same as 8bitTrueType, except for that we have to check 65536 charcodes
*/
if(!font->isCIDFont())
{
font_8bit = dynamic_cast<Gfx8BitFont*>(font);
maxcode = 0xff;
if(is_truetype_suffix(suffix))
{
ffw_reencode_glyph_order();
FoFiTrueType *fftt = nullptr;
if((fftt = FoFiTrueType::load((char*)filepath.c_str())) != nullptr)
{
code2GID = font_8bit->getCodeToGIDMap(fftt);
code2GID_len = 256;
delete fftt;
}
}
else
{
// move the slot such that it's consistent with the encoding seen in PDF
unordered_set<string> nameset;
bool name_conflict_warned = false;
memset(cur_mapping2, 0, 0x100 * sizeof(char*));
for(int i = 0; i < 256; ++i)
{
if(!used_map[i]) continue;
auto cn = font_8bit->getCharName(i);
if(cn == nullptr)
{
continue;
}
else
{
if(nameset.insert(string(cn)).second)
{
cur_mapping2[i] = cn;
}
else
{
if(!name_conflict_warned)
{
name_conflict_warned = true;
//TODO: may be resolved using advanced font properties?
cerr << "Warning: encoding confliction detected in font: " << hex << info.id << dec << endl;
}
}
}
}
ffw_reencode_raw2(cur_mapping2, 256, 0);
}
}
else
{
font_cid = dynamic_cast<GfxCIDFont*>(font);
maxcode = 0xffff;
if(is_truetype_suffix(suffix))
{
ffw_reencode_glyph_order();
GfxCIDFont * _font = dynamic_cast<GfxCIDFont*>(font);
// code2GID has been stored for embedded CID fonts
code2GID = _font->getCIDToGID();
code2GID_len = _font->getCIDToGIDLen();
}
else
{
ffw_cidflatten();
}
}
/*
* Step 2
* map charcode (or GID for CID truetype)
* generate an Consortium encoding file and let fontforge handle it.
*
* - Always map to Unicode for 8bit TrueType fonts and CID fonts
*
* - For 8bit nonTruetype fonts:
* Try to calculate the correct Unicode value from the glyph names, unless param->always_apply_tounicode is set
*
*
* Also fill in the width_list, and set widths accordingly
*/
{
unordered_set<int> codeset;
bool name_conflict_warned = false;
auto ctu = font->getToUnicode();
memset(cur_mapping, -1, 0x10000 * sizeof(*cur_mapping));
memset(width_list, -1, 0x10000 * sizeof(*width_list));
if(code2GID)
maxcode = min(maxcode, code2GID_len - 1);
bool is_truetype = is_truetype_suffix(suffix);
int max_key = maxcode;
/*
* Traverse all possible codes
*/
bool retried = false; // avoid infinite loop
for(int i = 0; i <= maxcode; ++i)
{
if(!used_map[i])
continue;
/*
* Skip glyphs without names (only for non-ttf fonts)
*/
if(!is_truetype && (font_8bit != nullptr)
&& (font_8bit->getCharName(i) == nullptr))
{
continue;
}
int k = i;
if(code2GID)
{
if((k = code2GID[i]) == 0) continue;
}
if(k > max_key)
max_key = k;
Unicode u, *pu=&u;
if(info.use_tounicode)
{
int n = ctu ? (ctu->mapToUnicode(i, &pu)) : 0;
u = check_unicode(pu, n, i, font);
}
else
{
u = unicode_from_font(i, font);
}
if(u == ' ')
info.has_space = true;
if(codeset.insert(u).second)
{
cur_mapping[k] = u;
}
else
{
// collision detected
if(param->tounicode == 0)
{
// in auto mode, just drop the tounicode map
if(!retried)
{
cerr << "ToUnicode CMap is not valid and got dropped" << endl;
retried = true;
codeset.clear();
info.use_tounicode = false;
memset(cur_mapping, -1, 0x10000 * sizeof(*cur_mapping));
memset(width_list, -1, 0x10000 * sizeof(*width_list));
i = -1;
continue;
}
}
if(!name_conflict_warned)
{
name_conflict_warned = true;
//TODO: may be resolved using advanced font properties?
cerr << "Warning: encoding confliction detected in font: " << hex << info.id << dec << endl;
}
}
if(font_8bit)
{
width_list[k] = (int)floor(font_8bit->getWidth(i) * info.em_size + 0.5);
}
else
{
char buf[2];
buf[0] = (i >> 8) & 0xff;
buf[1] = (i & 0xff);
width_list[k] = (int)floor(font_cid->getWidth(buf, 2) * info.em_size + 0.5);
}
}
ffw_reencode_raw(cur_mapping, max_key + 1, 1);
ffw_set_widths(width_list, max_key + 1);
if(ctu)
ctu->decRefCnt();
}
/*
* Step 3
*
* Generate the font as desired
*
*/
string cur_tmp_fn = (char*)str_fmt("%s/__tmp_font1%s", param->tmp_dir.c_str(), param->font_suffix.c_str());
add_tmp_file(cur_tmp_fn);
string other_tmp_fn = (char*)str_fmt("%s/__tmp_font2%s", param->tmp_dir.c_str(), param->font_suffix.c_str());
add_tmp_file(other_tmp_fn);
ffw_save(cur_tmp_fn.c_str());
ffw_close();
/*
* Step 4
* Font Hinting
*/
bool hinted = false;
// Call external hinting program if specified
if(param->external_hint_tool != "")
{
hinted = (system((char*)str_fmt("%s \"%s\" \"%s\"", param->external_hint_tool.c_str(), cur_tmp_fn.c_str(), other_tmp_fn.c_str())) == 0);
}
// Call internal hinting procedure if specified
if((!hinted) && (param->auto_hint))
{
ffw_load_font(cur_tmp_fn.c_str());
ffw_auto_hint();
ffw_save(other_tmp_fn.c_str());
ffw_close();
hinted = true;
}
if(hinted)
{
swap(cur_tmp_fn, other_tmp_fn);
}
/*
* Step 5
* Generate the font
* Reload to retrieve/fix accurate ascent/descent
*/
string fn = (char*)str_fmt("%s/f%llx%s",
(param->single_html ? param->tmp_dir : param->dest_dir).c_str(),
info.id, param->font_suffix.c_str());
if(param->single_html)
add_tmp_file(fn);
ffw_load_font(cur_tmp_fn.c_str());
ffw_metric(&info.ascent, &info.descent);
ffw_save(fn.c_str());
ffw_close();
}
void HTMLRenderer::drawString(GfxState * state, GooString * s)
{
if(s->getLength() == 0)
return;
auto font = state->getFont();
if((font == nullptr) || (font->getWMode()))
{
return;
}
//hidden
if((state->getRender() & 3) == 3)
{
return;
}
// see if the line has to be closed due to state change
check_state_change(state);
prepare_line(state);
// Now ready to output
// get the unicodes
char *p = s->getCString();
int len = s->getLength();
double dx = 0;
double dy = 0;
double dxerr = 0;
double dx1,dy1;
double ox, oy;
int nChars = 0;
int nSpaces = 0;
int uLen;
CharCode code;
Unicode *u = nullptr;
while (len > 0) {
auto n = font->getNextChar(p, len, &code, &u, &uLen, &dx1, &dy1, &ox, &oy);
if(!(_equal(ox, 0) && _equal(oy, 0)))
{
cerr << "TODO: non-zero origins" << endl;
}
bool is_space = false;
if (n == 1 && *p == ' ')
{
++nSpaces;
is_space = true;
}
if(is_space && (param->space_as_offset))
{
// ignore horiz_scaling, as it's merged in CTM
line_buf.append_offset((dx1 * cur_font_size + cur_letter_space + cur_word_space) * draw_scale);
}
else
{
if((param->decompose_ligature) && (uLen > 1) && all_of(u, u+uLen, isLegalUnicode))
{
line_buf.append_unicodes(u, uLen);
}
else
{
if(cur_font_info->use_tounicode)
{
Unicode uu = check_unicode(u, uLen, code, font);
line_buf.append_unicodes(&uu, 1);
}
else
{
Unicode uu = unicode_from_font(code, font);
line_buf.append_unicodes(&uu, 1);
}
}
}
dx += dx1;
dy += dy1;
++nChars;
p += n;
len -= n;
}
double hs = state->getHorizScaling();
// horiz_scaling is merged into ctm now,
// so the coordinate system is ugly
dx = (dx * cur_font_size + nChars * cur_letter_space + nSpaces * cur_word_space) * hs;
dy *= cur_font_size;
cur_tx += dx;
cur_ty += dy;
draw_tx += dx + dxerr * cur_font_size * hs;
draw_ty += dy;
}
} // namespace pdf2htmlEX

4
src/Preprocessor.cc
View File

@ -82,8 +82,8 @@ void Preprocessor::drawChar(GfxState *state, double x, double y,
void Preprocessor::startPage(int pageNum, GfxState *state)
{
max_width = max(max_width, state->getPageWidth());
max_height = max(max_height, state->getPageHeight());
max_width = max<double>(max_width, state->getPageWidth());
max_height = max<double>(max_height, state->getPageHeight());
}
const char * Preprocessor::get_code_map (long long font_id) const

25
src/include/CairoBackgroundRenderer.h
View File

@ -1,6 +1,6 @@
/*
* Splash Background renderer
* Render all those things not supported as Image, with Splash
* Cairo Background renderer
* Render all those things not supported as Image, with Cairo
*
* Copyright (C) 2012 Lu Wang <coolwanglu@gmail.com>
*/
@ -9,27 +9,32 @@
#ifndef CAIRO_BACKGROUND_RENDERER_H__
#define CAIRO_BACKGROUND_RENDERER_H__
#include <splash/SplashBitmap.h>
#include <SplashOutputDev.h>
#include <CairoOutputDev.h>
#include "Param.h"
namespace pdf2htmlEX {
// Based on BackgroundRenderer from poppler
class SplashBackgroundRenderer : public SplashOutputDev
class CairoBackgroundRenderer : public CairoOutputDev
{
public:
static const SplashColor white;
SplashBackgroundRenderer()
:SplashOutputDev(splashModeRGB8, 4, gFalse, (SplashColorPtr)&white, gTrue, gTrue)
CairoBackgroundRenderer(const Param * param)
:CairoOutputDev()
, param(param)
{ }
virtual ~SplashBackgroundRenderer() { }
virtual ~CairoBackgroundRenderer() { }
virtual void drawChar(GfxState *state, double x, double y,
double dx, double dy,
double originX, double originY,
CharCode code, int nBytes, Unicode *u, int uLen);
void render_page(PDFDoc * doc, int pageno, const std::string & filename);
protected:
const Param * param;
};
}

16
src/include/HTMLRenderer.h
View File

@ -79,6 +79,12 @@ class HTMLRenderer : public OutputDev
// Does this device use drawChar() or drawString()?
virtual GBool useDrawChar() { return gFalse; }
// Does this device use functionShadedFill(), axialShadedFill(), and
// radialShadedFill()? If this returns false, these shaded fills
// will be reduced to a series of other drawing operations.
virtual GBool useShadedFills(int type) { return type == 2; }
// Does this device use beginType3Char/endType3Char? Otherwise,
// text in Type 3 fonts will be drawn with drawChar/drawString.
virtual GBool interpretType3Chars() { return gFalse; }
@ -125,6 +131,7 @@ class HTMLRenderer : public OutputDev
virtual void stroke(GfxState *state) { css_draw(state, false); }
virtual void fill(GfxState *state) { css_draw(state, true); }
virtual GBool axialShadedFill(GfxState *state, GfxAxialShading *shading, double tMin, double tMax);
virtual void processLink(AnnotLink * al);
@ -208,11 +215,12 @@ class HTMLRenderer : public OutputDev
* w,h should be the metrics WITHOUT border
*
* line_color & fill_color may be specified as nullptr to indicate none
* style_function & style_function_data may be provided to provide more styles
*/
void css_draw_rectangle(double x, double y, double w, double h,
void css_draw_rectangle(double x, double y, double w, double h, const double * tm,
double * line_width_array, int line_width_count,
const GfxRGB * line_color, const GfxRGB * fill_color,
GfxState * state);
const GfxRGB * line_color, const GfxRGB * fill_color,
void (*style_function)(void *, std::ostream &) = nullptr, void * style_function_data = nullptr );
////////////////////////////////////////////////////
@ -396,7 +404,7 @@ class HTMLRenderer : public OutputDev
std::unordered_map<long long, FontInfo> font_name_map;
std::map<double, long long> font_size_map;
std::map<TM, long long> transform_matrix_map;
std::map<Matrix, long long, Matrix_less> transform_matrix_map;
std::map<double, long long> letter_space_map;
std::map<double, long long> word_space_map;
std::unordered_map<GfxRGB, long long, GfxRGB_hash, GfxRGB_equal> color_map;

19
src/include/util.h
View File

@ -104,28 +104,21 @@ public:
bool has_space; // whether space is included in the font
};
// wrapper of the transform matrix double[6]
// Transform Matrix
class TM
class Matrix_less
{
public:
TM() {}
TM(const double * m) {memcpy(_, m, sizeof(_));}
bool operator < (const TM & m) const {
bool operator () (const Matrix & m1, const Matrix & m2) const
{
// Note that we only care about the first 4 elements
for(int i = 0; i < 4; ++i)
{
if(_[i] < m._[i] - EPS)
if(m1.m[i] < m2.m[i] - EPS)
return true;
if(_[i] > m._[i] + EPS)
if(m1.m[i] > m2.m[i] + EPS)
return false;
}
return false;
}
bool operator == (const TM & m) const {
return _tm_equal(_, m._, 4);
}
double _[6];
};
class base64stream
@ -203,7 +196,7 @@ public:
va_end(vlist);
if(l >= (int)buf.capacity())
{
buf.reserve(std::max((long)(l+1), (long)buf.capacity() * 2));
buf.reserve(std::max<long>((long)(l+1), (long)buf.capacity() * 2));
va_start(vlist, format);
l = vsnprintf(&buf.front(), buf.capacity(), format, vlist);
va_end(vlist);

4
src/pdf2htmlEX.cc
View File

@ -183,8 +183,8 @@ int main(int argc, char **argv)
throw "Copying of text from this document is not allowed.";
}
param.first_page = min(max(param.first_page, 1), doc->getNumPages());
param.last_page = min(max(param.last_page, param.first_page), doc->getNumPages());
param.first_page = min<int>(max<int>(param.first_page, 1), doc->getNumPages());
param.last_page = min<int>(max<int>(param.last_page, param.first_page), doc->getNumPages());
if(param.output_filename == "")
{