关于c++:PDFium-渲染

PDFium 是 Chromium 的 PDF 渲染引擎，许可协定为 BSD 3-Clause。不同于 Mozilla 基于 HTML5 的 PDF.js，PDFium 是基于 Foxit Software（福昕软件）的渲染代码，Google 与其单干开源出的。

此外，Qt PDF 模块也选用了 PDFium，可见 QtWebEngine / QtPdf。

本文将介绍如何用 PDFium 实现一个简略的 PDF 阅读器，代码见：https://github.com/ikuokuo/pd…。

编译 PDFium

应用预编译库：https://github.com/bblanchon/…

不然，参考 PDFium / README 本人编译，实际步骤如下：

# get depot_tools, contains: gclient, ninja, gn, ...
git clone --depth 1 https://chromium.googlesource.com/chromium/tools/depot_tools.git
export PATH="$PATH:$HOME/Codes/Star/depot_tools"

# get pdfium
cd pdfium-reader/
mkdir -p third_party/chromium
cd third_party/chromium
gclient config --unmanaged https://pdfium.googlesource.com/pdfium.git
gclient sync
cd pdfium

# get deps
#  on linux, install additional build dependencies
./build/install-build-deps.sh

# gn config
#  args see the following `out/Release/args.gn`
gn args out/Release

# ninja build
#  pdfium
ninja -C out/Release pdfium
#  pdfium_test
ninja -C out/Release pdfium_test

# run sample: pdf > ppm
./out/Release/pdfium_test --ppm path/to/myfile.pdf

期间 out/Release/args.gn 内容如下：

use_goma = false  # Googlers only. Make sure goma is installed and running first.
is_debug = false  # Enable debugging features.

# Set true to enable experimental Skia backend.
pdf_use_skia = false
# Set true to enable experimental Skia backend (paths only).
pdf_use_skia_paths = false

pdf_enable_xfa = false  # Set false to remove XFA support (implies JS support).
pdf_enable_v8 = false  # Set false to remove Javascript support.
pdf_is_standalone = true  # Set for a non-embedded build.
pdf_is_complete_lib = true  # Set for a static library build.
is_component_build = false  # Disable component build (Though it should work)

应用 PDFium

浏览 PDFium / Getting Started，理解如何初始化 PDFium 及载入文档。步骤如下，或见 pdfium_start.c:

#include <fpdfview.h>
#include <stdio.h>

int main(int argc, char const *argv[]) {
  FPDF_STRING test_doc = "test_doc.pdf";
  if (argc >= 2) {test_doc = argv[1];
  }
  printf("test_doc: %s\n", test_doc);

  FPDF_InitLibrary();

  FPDF_DOCUMENT doc = FPDF_LoadDocument(test_doc, NULL);
  if (!doc) {unsigned long err = FPDF_GetLastError();
    // Load pdf docs unsuccessful: ...
    goto EXIT;
  }

  FPDF_CloseDocument(doc);
EXIT:
  FPDF_DestroyLibrary();
  return 0;
}

获取信息

样例见 pdf_info.cc，可打印 PDF 元数据、页面信息等。

如 FPDF_GetMetaText 获取元数据（UTF-16LE 编码）：

void PrintPdfMetaData(FPDF_DOCUMENT doc) {static constexpr const char *kMetaTags[] = {
      "Title",   "Author",   "Subject",      "Keywords",
      "Creator", "Producer", "CreationDate", "ModDate"};
  for (const char *meta_tag : kMetaTags) {const unsigned long len = FPDF_GetMetaText(doc, meta_tag, nullptr, 0);
    if (!len)
      continue;

    std::vector<char16_t> buf(len);
    FPDF_GetMetaText(doc, meta_tag, buf.data(), buf.size());
    auto text = strings::FromUtf16(std::u16string(buf.data()));
    if (strcmp(meta_tag, "CreationDate") == 0 ||
        strcmp(meta_tag, "ModDate") == 0) {text = fpdf::DateToRFC3399(text);
    }
    std::cout << "" << meta_tag <<": " << text << std::endl;
  }
}

渲染页面

样例见 pdf_render.cc，可渲染 PDF 页面并保留为 PNG。

FPDF_RenderPageBitmap 渲染某一页：

void PdfRenderPage(const std::string &pdf_name, FPDF_DOCUMENT doc, int index) {
  Timer t;

  FPDF_PAGE page = FPDF_LoadPage(doc, index);

  double scale = 1.0;
  // double scale = 2.0;
  int width = static_cast<int>(FPDF_GetPageWidth(page) * scale);
  int height = static_cast<int>(FPDF_GetPageHeight(page) * scale);
  int alpha = FPDFPage_HasTransparency(page) ? 1 : 0;
  ScopedFPDFBitmap bitmap(FPDFBitmap_Create(width, height, alpha));  // BGRx

  if (bitmap) {
    FPDF_DWORD fill_color = alpha ? 0x00000000 : 0xFFFFFFFF;
    FPDFBitmap_FillRect(bitmap.get(), 0, 0, width, height, fill_color);

    int rotation = 0;
    int flags = FPDF_ANNOT;
    FPDF_RenderPageBitmap(bitmap.get(), page, 0, 0, width, height,
        rotation, flags);
    auto t_render = t.Elapsed();

    int stride = FPDFBitmap_GetStride(bitmap.get());
    void *buffer = FPDFBitmap_GetBuffer(bitmap.get());

    char img_name[256];
    int chars_formatted = snprintf(img_name, sizeof(img_name), "%s.%d.png", pdf_name.c_str(), index);
    if (chars_formatted < 0 ||
        static_cast<size_t>(chars_formatted) >= sizeof(img_name)) {fprintf(stderr, "Filename is too long: %s\n", img_name);
      exit(EXIT_FAILURE);
    }

    auto ok = PdfWritePng(img_name, buffer, width, height, stride);
    if (!ok) {fprintf(stderr, "Write png failed: %s\n", img_name);
      exit(EXIT_FAILURE);
    }
    auto t_write = t.Elapsed();

    fprintf(stdout, "%s\n", img_name);
    fprintf(stdout, "%02d: %dx%d, render=%lldms, write=%lldms\n",
        index, width, height, t_render, t_write);
  } else {fprintf(stderr, "Page was too large to be rendered.\n");
    exit(EXIT_FAILURE);
  }

  FPDF_ClosePage(page);
}

stb_image_write.h 存为 PNG：

bool PdfWritePng(const std::string &img_name, void *buffer,
                 int width, int height, int stride) {
  // BGRA > RGBA
  auto buf = reinterpret_cast<uint8_t *>(buffer);
  for (int r = 0; r < height; ++r) {for (int c = 0; c < width; ++c) {auto pixel = buf + (r*stride) + (c*4);
      auto b = pixel[0];
      pixel[0] = pixel[2];  // b = r
      pixel[2] = b;         // r = b
    }
  }
  return stbi_write_png(img_name.c_str(), width, height, 4, buf, stride) != 0;
}

实现 UI

本文给出的 PDFium Reader 代码，用的 ImGui+GLFW+OpenGL3 实现的 UI，可跨三大桌面零碎。

想进一步理解的，能够间接看代码，编译运行按照 README。

GoCoding 集体实际的教训分享，可关注公众号！