insert_textbox vs insert_htmlbox

[tenghian]

Hi, when i use insert_textbox get correct font, but insert_htmlbox get different result, may i know how to change args in insert_htmlbox? thanks
jp_1.pdf
jp_2.pdf

#pymupdf 1.24.8

`import pymupdf

fname = "r:/jp_1.pdf"
new_file = "r:/jp_2.pdf"

doc = pymupdf.open(fname) # open a document
page = doc[0]

pfont = page.get_fonts(full=True)
print(pfont)
#[(11, 'ttf', 'TrueType', 'ABKYFM+HelveticaNeue-Bold', 'TT0', 'WinAnsiEncoding', 0),

(14, 'ttf', 'TrueType', 'ABFMUM+HelveticaNeue', 'TT1', 'WinAnsiEncoding', 0)]

fontbuffer = doc.extract_font(11)[-1]
page.insert_font(fontname="TT0", fontbuffer=fontbuffer)\

#-> font:HelveticaNeue-Bold (display correct font)
rect1 = pymupdf.Rect((100.0, 550.0, 300.0, 600.0))
page.insert_textbox(rect1, buffer="Embracing", fontsize=24, fontname="TT0")

#-> font:CharisSIL (insert_htmlbox cannot detect HelveticaNeue-Bold)
rect2 = pymupdf.Rect((100.0, 600.0, 300.0, 650.0))
page.insert_htmlbox(rect2, 'Embracing', css="* {font-family: HelveticaNeue-Bold;font-size:24px;}")

doc.save(new_file)
doc.close()

`

[JorjMcKie]

All you have to do is using CSS source (= styling information).
If you let MuPDF (the ultimate font selection logic behind everything Story-based like htmlbox), it will decide character-by-character, and the default is the "serif" font Charis.
If the character set of that is exhausted, Noto Serif is used - but previous decisions for Charis are of course not reverted in a retrograde manner.
So you can use css="* [font-family: sans-serif;}", which will lead to using the Noto Sans fonts (regular, bold, italic, ...) and no Charis, Noto Serif, etc. Or you specify one or more fonts from package pymupdf-fonts or some arbitrary choice of your own favorite fonts. Look at the recipes for insert_htmlbox` to see some more examples, among which are also file-based ones.

In any Case, MuPDF never looks at the fonts intalled on your machine. Either you accept the default behavior or you must actively provide fonts via file spec or memory equivalents.

[tenghian]

insert_textbox can read fonts within the page, how can I make insert_htmlbox do the same?

Is there a sample code to read fonts from memory? thanks

[JorjMcKie]

Method page.insert_textbox() does not "read" fonts from the PDF!
It is you who must first extract a font from the PDF and then make it available again via page.insert_font() giving it a new fontname to make it referenceable for insert_textbox.

A font existing in a PDF can also be made available to page.insert_htmlbox().
Because this method depends on a Story object and text styling is supported via HTML and CSS syntax, you must provide font support in that syntax too.

1. Extract existing font as before and create a Font object.
2. Create a PyMuPDF Archive object that references this Font via font.buffer.
3. Define CSS source that references the extracted font.
4. Do page.insert_htmlbox(rect, text, archive=arch, css=css)

Define Archive and CSS:

arch = pymupdf.Archive(font.buffer, "myfont")  # make font accessible via a name
css = """@font-face {font-family: myfont; src: url(myfont);}
      * {font-family: myfont;}"""

As always: a font in a PDF normally is not complete, but a subset of some full font: several / many characters are missing.
If characters are missing in supplied fonts, Story will behave differently than insert_textbox ... so you need to experiment a little here.

[tenghian]

fname = "r:/jp_1.pdf"
new_file = "r:/jp_2.pdf"

doc = pymupdf.open(fname) # open a document
page = doc[0]

pfont = page.get_fonts(full=True)
print(pfont)

fontbuffer = doc.extract_font(11)[-1]
font = pymupdf.Font(fontname="TT0", fontbuffer=fontbuffer)
page.insert_font(fontname="TT0", fontbuffer=fontbuffer)

arch = pymupdf.Archive(font.buffer, "myfont") # make font accessible via a name
css = """@font-face {font-family: myfont; src: url(myfont);}
* {font-family: myfont;}"""

rect2 = pymupdf.Rect((100.0, 600.0, 300.0, 650.0))
page.insert_htmlbox(rect2, 'Embracing', css=css)

rect1 = pymupdf.Rect((100.0, 550.0, 300.0, 600.0))
page.insert_textbox(rect1, buffer="Embracing", fontsize=24, fontname="TT0")

-> MuPDF error: format error: cannot locate font 'myfont' specified by css
I am using the text and font that already exist on the page, and insert_textbox can read it normally.

[JorjMcKie]

You did not specify the archive parameter in insert_htmlblox!

[JorjMcKie]

This demo script works as desired:

import pymupdf
from pprint import pprint

doc = pymupdf.open("test-doc-2.pdf")
buff = doc.extract_font(124)[-1]
doc.close()  # original PDF no longer needed

font = pymupdf.Font(fontbuffer=buff)
arch = pymupdf.Archive(font.buffer, "myfont")
css = """ @font-face {font-family: myfont; src: url(myfont);}
* {font-family: myfont;}"""
doc = pymupdf.open()
page = doc.new_page()
page.insert_htmlbox((100, 100, 200, 200), "Text in custom font", css=css, archive=arch)
pprint(page.get_fonts())

[tenghian]

Thank you. ^_^