wrong result from get_string_width

[matipos2]

It seems that function get_string_width() gives wrong result. It's always 2 points less than I'd expect.

Minimal code
Minimal Python code reproducing the issue (I couldn't find any information about that behavior in documentation):

import fpdf

print(f"fpdf version: {fpdf.FPDF_VERSION}")
pdf = fpdf.FPDF(orientation='portrait', format='A4')
pdf.set_font("helvetica", "I", 8)
pdf.add_page()
text = "THIS is Text to test width"
text_length = pdf.get_string_width(text)

first_x = pdf.get_x()
print(f"x={pdf.get_x()}, y={pdf.get_y()}")
print(f"get_string_width={text_length}")
pdf.cell(text=text)
print(f"new x={pdf.get_x()}, new y={pdf.get_y()}")
print(f"expected x={first_x + text_length}")

print("----- now with custom font -----")
pdf.add_font("OpenSans", style="", fname="OpenSans-Regular.ttf")
pdf.ln(10)
pdf.set_font("OpenSans", "", 8)
text_length = pdf.get_string_width(text)

first_x = pdf.get_x()
print(f"x={pdf.get_x()}, y={pdf.get_y()}")
print(f"get_string_width={text_length}")
pdf.cell(text=text)
print(f"new x={pdf.get_x()}, new y={pdf.get_y()}")
print(f"expected x={first_x + text_length}")

Here is the result:

fpdf version: 2.7.6
x=9.999999999999998, y=9.999999999999998
get_string_width=31.368999999999996
new x=43.36899999999999, new y=9.999999999999998
expected x=41.36899999999999
----- now with custom font -----
x=9.999999999999998, y=20.0
get_string_width=32.322911111111104
new x=44.322911111111104, new y=20.0
expected x=42.322911111111104

[Lucas-C]

Hi @matipos2 and welcome

What you are observing is the impact of the cell margin setting (.c_margin).

You get get rid of this difference simply by setting pdf.c_margin = 0

We should probably better document this.
Where in our documentation would expect this to be explained clearly?
Maybe would you like to contribute a PR to improve our documentation with a mention of this?

[matipos2]

Hi @Lucas-C thank you for quick response.
I would add this description to cell and multi_cell. I'm not sure if it applies to others:

Final width of the cell is extended by the interior cell margin (c_margin property). In the result it will get the size: c_margin + string width + c_margin.

And additional note to the description of get_string_width:

Please note that it calculates only the string width. The width of the cell (pdf.cell()) will be wider by interior cell margin (pdf.c_margin).

If you accept, I may prepare a PR later on.

[Lucas-C]

Looks great !
A PR would be welcome 🙂

I would suggest those slights rewordings, to keep those docstrings short:

• in the docstring for FPDF.cell() & FPDF.multi_cell():

A cell has an horizontal padding, on the left & right sides, defined by the.c_margin property.

• in the docstring for get_string_width()

Note that the width of a cell has some extra padding added to this width, on the left & right sides, equal to the .c_margin property.

What do you think?

[matipos2]

I like your version more. The PR is ready: #1012

[Lucas-C]

I merged your PR, thank you for your contribution!

Closing this now