background image embedding works now

README.md

@@ -35,11 +35,11 @@ Dependency
 
 HOW TO COMPILE
 ----------------------------
-    cmake . && make
+    cmake . && make && sudo make install
 
 HOW TO USE
 ----------------------------
-    bin/pdf2htmlEX /path/to/sample.pdf
+    pdf2htmlEX /path/to/sample.pdf
 
 
 LICENSE

src/HTMLRenderer.h

@@ -30,6 +30,31 @@
 #include "Param.h"
 #include "util.h"
 
+
+/*
+ * Naming Convention
+ *
+ * ID
+ *
+ * p<hex> - Page
+ *
+ * CSS classes
+ *
+ * p - Page
+ * l - Line
+ * w - White space
+ * i - Image
+ *
+ * Reusable CSS classes
+ *
+ * f<hex> - Font (also for font names)
+ * s<hex> - font Size
+ * w<hex> - White space
+ * t<hex> - Transform matrix
+ * c<hex> - Color
+ *
+ */
+
 class HTMLRenderer : public OutputDev
 {
     public:

src/HTMLRenderer/export.cc

@@ -15,25 +15,6 @@
 
 using boost::algorithm::ifind_first;
 
-/*
- * CSS classes
- *
- * p - Page
- * l - Line
- * w - White space
- * i - Image
- *
- * 
- * Reusable CSS classes
- *
- * f<hex> - Font (also for font names)
- * s<hex> - font Size
- * w<hex> - White space
- * t<hex> - Transform matrix
- * c<hex> - Color
- *
- */
-
 void HTMLRenderer::export_remote_font(long long fn_id, const string & suffix, const string & fontfileformat, GfxFont * font)
 {
 

src/HTMLRenderer/general.cc

@@ -3,8 +3,6 @@
  *
  * Hanlding general stuffs
  *
- * TODO: better name for this file?
- *
  * by WangLu
  * 2012.08.14
  */
@@ -20,8 +18,6 @@
 
 using std::flush;
 
-using boost::filesystem::file_size;
-
 HTMLRenderer::HTMLRenderer(const Param * param)
     :line_opened(false)
     ,image_count(0)
@@ -73,7 +69,6 @@ void HTMLRenderer::process(PDFDoc *doc)
             bg_renderer->getBitmap()->writeImgFile(splashFormatPng, (char*)(working_dir() / (format("p%|1$x|.png")%i).str()).c_str(), param->h_dpi2, param->v_dpi2);
         }
 
-
         doc->displayPage(this, i, param->h_dpi, param->v_dpi,
                 0, true, false, false,
                 nullptr, nullptr, nullptr, nullptr);
@@ -90,7 +85,8 @@ void HTMLRenderer::process(PDFDoc *doc)
 
 void HTMLRenderer::pre_process()
 {
-    html_fout.open(working_dir() / param->output_filename, ofstream::binary); // we may output utf8 characters, so use binary
+    // we may output utf8 characters, so use binary
+    html_fout.open(working_dir() / param->output_filename, ofstream::binary); 
     allcss_fout.open(working_dir() / "all.css", ofstream::binary);
     fontscript_fout.open(tmp_dir / "pdf2htmlEX.pe", ofstream::binary);
 
@@ -129,27 +125,25 @@ void HTMLRenderer::startPage(int pageNum, GfxState *state)
 
     assert(!line_opened);
 
-    html_fout << format("<div id=\"page-%3%\" class=\"p\" style=\"width:%1%px;height:%2%px;") % pageWidth % pageHeight % pageNum;
+    html_fout << format("<div id=\"p%|1$x|\" class=\"p\">") % pageNum << endl;
 
-    html_fout << "background-image:url(";
+    allcss_fout << format("#p%|1$x|{width:%2%px;height:%3%px;") % pageNum % pageWidth % pageHeight;
+
+    allcss_fout << "background-image:url(";
 
     const std::string fn = (format("p%|1$x|.png") % pageNum).str();
     if(param->single_html)
     {
         auto path = tmp_dir / fn;
-        html_fout << "'data:image/png;base64,";
-        copy_base64(html_fout, ifstream(path), file_size(path));
-        html_fout << "'";
+        allcss_fout << "'data:image/png;base64," << base64_filter(ifstream(path, ifstream::binary)) << "'";
     }
     else
     {
-        html_fout << fn;
+        allcss_fout << fn;
     }
     
-    html_fout << format(");background-position:0 0;background-size:%1%px %2%px;background-repeat:no-repeat;") % pageWidth % pageHeight;
+    allcss_fout << format(");background-position:0 0;background-size:%1%px %2%px;background-repeat:no-repeat;}") % pageWidth % pageHeight;
             
-    html_fout << "\">" << endl;
-
     cur_fn_id = cur_fs_id = cur_tm_id = cur_color_id = 0;
     cur_tx = cur_ty = 0;
     cur_font_size = 0;

src/util.h

@@ -14,10 +14,10 @@
 #include <istream>
 #include <ostream>
 #include <iterator>
+#include <iomanip>
 
-#include <boost/archive/iterators/transform_width.hpp>
-#include <boost/archive/iterators/base64_from_binary.hpp>
-
+#include <GfxState.h>
+#include <CharTypes.h>
 #include <UTF8.h>
 
 #include "Consts.h"
@@ -25,11 +25,8 @@
 using std::istream;
 using std::ostream;
 using std::istream_iterator;
-using std::ostream_iterator;
-using std::copy;
-
-using boost::archive::iterators::base64_from_binary;
-using boost::archive::iterators::transform_width;
+using std::endl;
+using std::noskipws;
 
 // mute gcc warning of unused function
 namespace
@@ -123,40 +120,61 @@ public:
     double _[6];
 };
 
-static inline void copy_base64 (ostream & out, istream & in, size_t length)
+class base64_filter
 {
-    typedef base64_from_binary < transform_width < istream_iterator<char>, 6, 8 > > base64_iter;
-    copy(base64_iter(istream_iterator<char>(in)), base64_iter(istream_iterator<char>()), ostream_iterator<char>(out));
-    switch(length % 3)
-    {
-        case 1:
-            out << '=';
-            // fall through
-        case 2:
-            out << '=';
-            // fall through
-        case 0:
-        default:
-            break;
-    }
-}
+public:
 
-static inline void copy_base64 (ostream & out, istream && in, size_t length)
-{
-    typedef base64_from_binary < transform_width < istream_iterator<char>, 6, 8 > > base64_iter;
-    copy(base64_iter(istream_iterator<char>(in)), base64_iter(istream_iterator<char>()), ostream_iterator<char>(out));
-    switch(length % 3)
+    base64_filter(istream & in)
+        : in_iter(istream_iterator<char>(in >> noskipws))
+    { }
+
+    base64_filter(istream && in)
+        : in_iter(istream_iterator<char>(in >> noskipws))
+    { }
+
+    ostream & dumpto(ostream & out)
     {
-        case 1:
+        unsigned char buf[3];
+        istream_iterator<char> end_iter;
+        int cnt = 0;
+        while(in_iter != end_iter)
+        {
+            buf[cnt++] = *(in_iter++);
+            if(cnt == 3)
+            {
+                out << base64_encoding[(buf[0] & 0xfc)>>2];
+                out << base64_encoding[((buf[0] & 0x03)<<4) | ((buf[1] & 0xf0)>>4)];
+                out << base64_encoding[((buf[1] & 0x0f)<<2) | ((buf[2] & 0xc0)>>6)];
+                out << base64_encoding[(buf[2] & 0x3f)];
+                cnt = 0;
+            }
+        } 
+        if(cnt > 0)
+        {
+            for(int i = cnt; i < 3; ++i)
+                buf[i] = 0;
+            out << base64_encoding[(buf[0] & 0xfc)>>2];
+            out << base64_encoding[((buf[0] & 0x03)<<4) | ((buf[1] & 0xf0)>>4)];
+            if(cnt > 1)
+            {
+                out << base64_encoding[(buf[1] & 0x0f)<<2];
+            }
+            else
+            {
+                out <<  '=';
+            }
             out << '=';
-            // fall through
-        case 2:
-            out << '=';
-            // fall through
-        case 0:
-        default:
-            break;
+        }
+
+        return out;
     }
-}
+
+private:
+    static constexpr const char * base64_encoding = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
+    istream_iterator<char> in_iter;
+};
+
+static inline ostream & operator << (ostream & out, base64_filter & bf) { return bf.dumpto(out); }
+static inline ostream & operator << (ostream & out, base64_filter && bf) { return bf.dumpto(out); }
 
 #endif //UTIL_H__