How to override font family for all the text in html in PdfRendererBuilder?

[asu2]

Hi,

The problem we have is that the html source files are from different vendors/countries so it seems impossible to load all fonts they are using or will use. I'd like to know if there's a way to override font for all the text, regardless which font family is declared for them in CSS (but still keep font style and size if possible), so I can simply specify an unicode-friendly font for all html sources.

Thanks!

[draco1023]

1. Register FONT_NAME to PdfRendererBuilder.

PdfRendererBuilder builder = new PdfRendererBuilder();
builder.useFont(YOUR_FONT_SUPPILER, FONT_NAME);
builder.useCacheStore(PdfRendererBuilder.CacheStore.PDF_FONT_METRICS, new FSDefaultCacheStore());

2. Parse html content with Jsoup.
3. Append a style tag to the document head with content * { font-family: FONT_NAME; }.
4. Find all elements with inline style and append FONT_NAME at the end of font-family declaration if it exists.

[asu2]

Thanks for the reply.

I found another way to do this:

1. Create a subclass from PdfBoxFontResolver then override resolveFont() method to return a 'default' font;
2. Now comes the ugly: I have to use a lot of reflections to private fields to make my subclass work in PdfBoxRenderer.

I will definitely try your suggestion as I don't like reflection-to-private at all.

[syjer]

hi @asu2 , I think it would be a nice feature to have a customizable Font resolver (edit: for fallback). I'll have a look at what it would mean during the week end.

If you can post your subclass here, it would be a nice start to see what we would need to change/modify.

[asu2]

@syjer here you go ;-)

import com.openhtmltopdf.css.constants.IdentValue;
import com.openhtmltopdf.css.value.FontSpecification;
import com.openhtmltopdf.extend.FSCacheEx;
import com.openhtmltopdf.extend.FSCacheValue;
import com.openhtmltopdf.layout.SharedContext;
import com.openhtmltopdf.pdfboxout.PdfBoxFSFont;
import com.openhtmltopdf.pdfboxout.PdfBoxFontResolver;
import com.openhtmltopdf.pdfboxout.PdfBoxRenderer;
import com.openhtmltopdf.pdfboxout.PdfRendererBuilder;
import com.openhtmltopdf.render.FSFont;
import org.apache.fontbox.ttf.OTFParser;
import org.apache.fontbox.ttf.OpenTypeFont;
import org.apache.pdfbox.pdmodel.PDDocument;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.apache.pdfbox.pdmodel.font.PDType0Font;
import org.apache.pdfbox.pdmodel.font.PDType1Font;

import java.io.Closeable;
import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.lang.reflect.Constructor;
import java.lang.reflect.Field;
import java.util.ArrayList;
import java.util.List;

public class OverriddenFontResolver extends PdfBoxFontResolver {

	private static String[] FONT_FILES = {
		"fonts/NotoSansNaskhArabic-Regular.ttf",
		"fonts/NotoSansHebrew-Regular.ttf",
		"fonts/NotoSansCJKjp-Regular.otf",
		"fonts/NotoSansCJKkr-Regular.otf",
		"fonts/NotoSansCJKsc-Regular.otf",
		"fonts/NotoSans-Regular.ttf",
	};
	private List<FontDescription> fonts;

	// The way I did this is because I have no idea where I can get these parameters correctly!
	public OverriddenFontResolver(PdfBoxFontResolver resolver) {
		super((SharedContext) readField(resolver, "_sharedContext"), (PDDocument) readField(resolver, "_doc"),
			(FSCacheEx<String, FSCacheValue>) readField(resolver, "_fontMetricsCache"), (PdfRendererBuilder.PdfAConformance) readField(resolver, "_pdfAConformance"),
			(Boolean) readField(resolver, "_pdfUaConform"));
	}

	@Override
	public FSFont resolveFont(SharedContext renderingContext, FontSpecification spec) {
		return new PdfBoxFSFont(getFonts(), spec.size);
	}

	private static Object readField(PdfBoxFontResolver resolver, String name) {
		try {
			Field f = PdfBoxFontResolver.class.getDeclaredField(name);
			f.setAccessible(true);
			return f.get(resolver);
		}
		catch (Exception ex) {
			throw new RuntimeException("Failed to read field: " + name, ex);
		}
	}

	private static FontDescription newFontDescription(PDFont font) {
		try {
			Constructor<FontDescription> c = FontDescription.class.getDeclaredConstructor(PDFont.class, IdentValue.class, Integer.TYPE);
			c.setAccessible(true);
			return c.newInstance(font, IdentValue.NORMAL, 400);
		}
		catch (Exception ex) {
			throw new RuntimeException("Failed to create new instance: FontDescription", ex);
		}
	}

	private List<FontDescription> getFonts() {
		if (fonts == null) {
			List<FontDescription> temp = new ArrayList<>();
			for (String f : FONT_FILES) {
				try (InputStream in = new FileInputStream(new File(f))) {
					OTFParser parser = new OTFParser();
					OpenTypeFont otf = parser.parse(in);
					temp.add(newFontDescription(PDType0Font.load((PDDocument) readField(this, "_doc"), otf, false)));
				}
				catch (IOException ex) {
					//todo
					System.err.println("Error in loading font file: " + f + ", " + ex);
				}
			}
			temp.add(newFontDescription(PDType1Font.HELVETICA));
			fonts = temp;
		}
		return fonts;
	}

	public static void main(String[] args) {
		// Replace default resolver.
		PdfRendererBuilder builder = new PdfRendererBuilder() {
			@Override
			public PdfBoxRenderer buildPdfRenderer(Closeable diagnosticConsumer) {
				PdfBoxRenderer renderer = super.buildPdfRenderer(diagnosticConsumer);
				PdfBoxFontResolver resolver = renderer.getFontResolver();
				renderer.getSharedContext().setFontResolver(new OverriddenFontResolver(resolver));
				return renderer;
			}
		};
	}
}

[draco1023]

openhtmltopdf/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxFontResolver.java

Lines 421 to 430 in d20825b

	if (_pdfAConformance == PdfAConformance.NONE &&
	!_pdfUaConform) {
	// We don't have a final fallback font for PDF/A documents as serif may not be available
	// unless the user has explicitly embedded it.
	// For now, we end up with "Serif" built-in font.
	// Q: Should this change?
	// Q: Should we have a final automatically added font?
	fonts.add(resolveFont(ctx, "Serif", size, weight, style, variant));
	}

Serif is not working for unicode characters. Maybe we should expose a fallback font property in PdfRendererBuilder and SharedContext, then we can use the fallback font in PdfBoxFontResolver.

[asu2]

yeah, an easily-accessible fallback font is exactly what I'm looking for. Currently I'm using your 'css-injection' solution which works very well. Thanks!

[syjer]

looks like exposing the possibility to provide a programmatic way to let the user select the fallback font is the most elegant way.

You register all the fonts that you need as normal and in the builder you pass something like:

builder.defineFallbackFontResolver( (x,y,z,??) -> {your code here; return font name;});

[danfickle]

I've opened PR #669. Feel free to have a look and comment. Trying to do a long overdue release tomorrow though. Will merge first.

[IStrangers]

Add !important after CSS font