c++mupdf(1.17.0)读取文本

需要读取pdx的内容，主要是文本部分，本来以为比较简单，还是折腾了一下，特作记录。

关于vs2017集成mupdf，这个直接百度关键词就ok了，网上介绍的是0.17.0版本的mupdf，所以就用了这个。

然后搜索读取文本内容，很多都是python的，相关函数找不到，c++很少资料，还得是外网的示例代码里翻到，特此记录供大家取用：

std::string readPDF(std::string filePath,int pageNum)
{
	std::string ret;
	
	//使用默认值
	int pageCount = 0;
	int pdfPageIdx = 0;
	float zoom = 100;
	float rotate = 0;

	fz_context *ctx;
	fz_document *doc;
	fz_pixmap *pix;
	fz_text * ptxt;

	fz_page *page;
	fz_device *dev;
	
	int x, y;

	/* Create a context to hold the exception stack and various caches. */
	ctx = fz_new_context(NULL, NULL, FZ_STORE_UNLIMITED);
	if (!ctx)
	{
		fprintf(stderr, "cannot create mupdf context\n");
		return false;
	}

	/* Register the default file types to handle. */
	fz_try(ctx)
	{
		fz_register_document_handlers(ctx);
	}
	fz_catch(ctx)
	{
		fprintf(stderr, "cannot register document handlers: %s\n", fz_caught_message(ctx));
		fz_drop_context(ctx);
		return false;
	}

	/* Open the document. */
	fz_try(ctx)
	{
		doc = fz_open_document(ctx, filePath.c_str());
	}
	fz_catch(ctx)
	{
		fprintf(stderr, "cannot open document: %s\n", fz_caught_message(ctx));
		fz_drop_context(ctx);
		return false;
	}

	/* Count the number of pages. */
	fz_try(ctx)
	{
		pageCount = fz_count_pages(ctx, doc);
	}
	fz_catch(ctx)
	{
		fprintf(stderr, "cannot count number of pages: %s\n", fz_caught_message(ctx));
		fz_drop_document(ctx, doc);
		fz_drop_context(ctx);
		return ret;
	}


	int target_page = pageNum;

	fz_stext_page *text_page = fz_new_stext_page_from_page_number(ctx, doc, target_page, 0);
	std::string r;
	for (fz_stext_block* text_block = text_page->first_block;
		text_block != nullptr; text_block = text_block->next) {
		if (text_block->type != FZ_STEXT_BLOCK_TEXT) {
			continue;
		}
		for (fz_stext_line* text_line = text_block->u.t.first_line;
			text_line != nullptr; text_line = text_line->next) {
			for (fz_stext_char* text_char = text_line->first_char;
				text_char != nullptr; text_char = text_char->next) {
					{
						const int c = text_char->c;

					// A single UTF-8 character cannot take more than 4 bytes, but let's
					// go for 8.
					char buffer[8];
					const int num_bytes = fz_runetochar(buffer, c);
					if (num_bytes > 1)
					{
						continue;
					}
					assert(num_bytes <= static_cast<int>(sizeof(buffer)));
					buffer[num_bytes] = '\0';
					r += buffer;
				}
			}
			if (!isspace(r.back())) {
				//r += int(0);
			}
		}
	}
	
	ret = r;
	/* Clean up. */
	fz_drop_stext_page(ctx, text_page);
	fz_drop_document(ctx, doc);
	fz_drop_context(ctx);

	return ret;
}