moi15moi / fontcollector Goto Github PK

View Code? Open in Web Editor NEW

32.0 2.0 2.0 23.82 MB

FontCollector for Advanced SubStation Alpha file.

Home Page: https://pypi.org/project/FontCollector/

License: MIT License

Python 100.00%

advanced-substation-alpha aegisub ass python subtitle fonts variable-fonts

fontcollector's Introduction

fontcollector's People

Contributors

Stargazers

Watchers

Forkers

lordflashmeow naveen-98

fontcollector's Issues

\r<style_name> don't work

Currently, FontCollector will always reset the style to the line style.

It should see if there is a style specify after the \r.

If yes, parse it and apply the style, if no, take the line style.

Warning: Need to strip the style only at the end and match exactly the case with the style name. Also, it can not contain "(" or ")".

Fonts with same name same weight and same italic

So I have the following issue. I have a font named SF Pro Display, that has different weight (and italics) options in the same name (see attached screenshots)

In the ass file, I'm using the first one and Aegisub font collector does collect the first one (shown below). So for some reason, this tool collects the italic variant.

Fonts activated by FontBase are not found by FontCollector

(Windows 11)

The current font enumeration method does not find fonts activated by FontBase; even though they should be detected as system fonts considering every app with a font picker can find them.

Potentially could also be causing issues with other apps that utilize the same method as FontBase to activate their fonts.

I've confirmed with other users that they also experience the same issue. One user have testified that FontCollector successfully finds their FontBase fonts though but we couldn't identify the cause—so there might be a configuration change necessary for it to work.

Variable font without AxisValue cause crash

FontCollector is not able to read variable font without AxisValue.
Here is an font that it cannot parse: Inter-VariableFont_slntwght.zip

Here is the STAT table of the font.

  <STAT>
    <Version value="0x00010001"/>
    <DesignAxisRecordSize value="8"/>
    <!-- DesignAxisCount=2 -->
    <DesignAxisRecord>
      <Axis index="0">
        <AxisTag value="wght"/>
        <AxisNameID value="270"/>  <!-- Weight -->
        <AxisOrdering value="0"/>
      </Axis>
      <Axis index="1">
        <AxisTag value="slnt"/>
        <AxisNameID value="271"/>  <!-- Slant -->
        <AxisOrdering value="1"/>
      </Axis>
    </DesignAxisRecord>
    <!-- AxisValueCount=0 -->
    <ElidedFallbackNameID value="2"/>  <!-- Regular -->
  </STAT>

GDI name this font with 2 value:
1- typo/family name = "Inter"
2- typo/family name + ElidedFallbackNameID = "Inter Regular"

Add mkvfontvalidator

mkvfontvalidator would be a new command to verify if all the font files are muxed to the mkv.

It will automatically detect the muxed fonts and ass file by using the codec id.
To be compatible with mpv, we will use the same codec id as them. Codec id for ass file and codec id for font file.

Finally, mkvfontvalidator should have at least these options

  -mkv MKV
                        The video file to be verified. Must be a Matroska file.
  -mkvpropedit MKVPROPEDIT
                        Path to mkvpropedit.exe if not in variable environments. If -mkv is not specified, it will do
                        nothing.
  --need-draw-fonts
                        If specified, FontCollector will report a error if a font used in a draw isn't muxed to the mkv. For more detail when this
                        is usefull, see: https://github.com/libass/libass/issues/617

[Feature] Add Logging

Please add a logging parameter --logging -log .

Example:

--logging OUTPUT, -log OUTPUT
                        Destination path of log. If OUTPUT isn't specified, it will be the current path.

Helpful when doing large batches to easily search for any errors warnings etc.

You could just have its default log name be something like:

<SYSTEMStartTime>_log.txt

Or something like that it really doesn't matter too much, Or just require the user to declare filename in OUTPUT with DIR.

Thanks!

Win95 Trad Chinese

Windows 95 OSR2.5 (Trad. Chinese)
https://archive.org/details/Win95CCHT
Extract ISO.
Find 3 .cab files - MINI.CAB, MINI1.CAB, MINI2.CAB (3-parts)
Use 7-Zip, or WinRAR, etc. to extract from MINI.CAB.
You will find what you are looking for.

Add mypy and pytest verification workflow

When someones open a PR, we should run a workflow and run pylint, mypy and pytest to ensure the quality of the code

ASS style fonts are not collected if they're unused

AssDocument::get_used_style(true) does not return Arial in this case. __set_used_styles gets called for Arial, but it's not being set as a used style.

[Script Info]
ScriptType: v4.00+
WrapStyle: 0
PlayResX: 1920
PlayResY: 1080
ScaledBorderAndShadow: yes
YCbCr Matrix: TV.709
Video Aspect Ratio: c1.777780

[V4+ Styles]
Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
Style: Sign,Arial,30,&H00FFFFFF,&H000000FF,&H00000000,&H00000000,0,0,0,0,100,100,0,0,1,0,0,2,15,15,15,1

[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:01.10,0:00:10.10,Sign,,0,0,0,,{\pos(1118.334,292.5)\fay-0.6\fnAdobe Gothic Std B\fs15\b1\c&H574942&\blur0.75}Oozuka

The only font collected here is Adobe Gothic Std B; so Arial does not end up being muxed into the final output when used in a muxing script.

Playing with mpv and without a font provider, using only the fonts collected by fontcollector, this appears:

mpv --no-config --sub-font-provider=none <path to mkv>
[sub/ass] fontselect: failed to find any fallback with glyph 0x0 for font: (Arial, 400, 0)

Reproduce lfPitchAndFamily ET lfCharSet

FontCollector tries to reproduce what VSFilter does. VSFilter uses GDI to select a font and display it, so we need to emulate it.
We already properly emulate the score for the lfWeight and lfItalic attributes, but we don't emulate the score for lfPitchAndFamily and lfCharSet.

99% of the time, these properties don't affect the font that will be picked up. In general, they can affect the font picked up when the user has two fonts that are the same, but one is TrueType and the other is OpenType.

lfPitchAndFamily score algorithm

font_family = FF_DECORATIVE, FF_DONTCARE, FF_MODERN, FF_ROMAN, FF_SCRIPT or FF_SWISS
requested_family = is it always FF_DONTCARE
requested_charset = Depend on \fe tag or the ass style
if (jMapCharSet == SYMBOL_CHARSET) {
	requested_family = font_family
} else if (font_family != FF_DONTCARE) {
	requested_family = FF_SWISS // Technically, if the requested font is "Tms Rmn", it will be FF_ROMAN, but I am not sure if we want this behavior
}

if (requested_family != font_family) {
	if (font_family == FF_DONTCARE) {
		score += 8000;
	} else {
		if (font_family  > FF_MODERN) {
			score += 50;
		}
		score += 9000;
	}
}

lfCharSet score algorithm

requested_charset = Depend on \fe tag or the ass style
font_charsets = list of charset
if requested_charset != DEFAULT_CHARSET and requested_charset not in font_charsets:
	REJECT the font

I still don't fully understand how GDI sets lfCharSet and lfPitchAndFamily, so I will need to dig deeper before I can implement it.

[Feature] Delete glyph that are NOT been used in the .ass file

fontTools allow to create a subset of the font like this:

from fontTools import subset
from fontTools.ttLib.ttFont import TTFont

font = TTFont("font.TTF")

subsetter = subset.Subsetter()
subsetter.populate(text="ABÉé")
subsetter.subset(font)

font.save("font strip.TTF")

Option --additional-fonts doesn't work

In the latest version 2.0.2, the --additional-fonts function does not work. I get an error every time.
I have the font collections on another disk, I get this error:

C:\Users\xyz\Desktop\Subs>fontcollector -i --additional-fonts D:\Junk\Fonts\
Traceback (most recent call last):
  File "C:\Users\xyz\scoop\apps\Python\current\Scripts\fontcollector-script.py", line 33, in <module>
    sys.exit(load_entry_point('FontCollector==2.0.2', 'console_scripts', 'fontcollector')())
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\__main__.py", line 23, in main
    ) = parse_arguments()
        ^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\parse_arguments.py", line 121, in parse_arguments
    additional_fonts = FontLoader.load_additional_fonts(args.additional_fonts)
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\font_loader.py", line 124, in load_additional_fonts
    additional_fonts.update(Font.from_font_path(path))
                            ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\font.py", line 72, in from_font_path
    axis_value_tables = ParseFont.get_axis_value_from_coordinates(
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\parse_font.py", line 132, in get_axis_value_from_coordinates
    raise ValueError("Cannot get axis_value since there is no STAT table.")
ValueError: Cannot get axis_value since there is no STAT table.

Also I tested what if the fonts were on the same drive as the subtitles, an error appears interestingly it is different from the first case.

C:\Users\xyz\Desktop\Subs>fontcollector -i --additional-fonts ..\Fonts\
Traceback (most recent call last):
  File "C:\Users\xyz\scoop\apps\Python\current\Scripts\fontcollector-script.py", line 33, in <module>
    sys.exit(load_entry_point('FontCollector==2.0.2', 'console_scripts', 'fontcollector')())
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\__main__.py", line 25, in main
    font_collection = FontLoader(additional_fonts, use_system_font).fonts
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\font_loader.py", line 29, in __init__
    self.additional_fonts = FontLoader.load_additional_fonts(additional_fonts_path)
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xyz\scoop\apps\Python\current\Lib\site-packages\font_collector\font_loader.py", line 120, in load_additional_fonts
    if os.path.isfile(font_path):
       ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<frozen genericpath>", line 30, in isfile
TypeError: stat: path should be string, bytes, os.PathLike or integer, not Font

\r and \fn regex are wrong

\r and \fn can match to string when it should not.

Example:
The \r will match with the r of the tag \frz
{\frz0\alpha&H00&\fn123marker\blur0.5\fscx37\fscy34\pos(963.28,788.5)\c&H071B1E&\3c&H071B1E&}tooth{\clip(m 962 761 l 961 773 961 782 962 793 997 790 997 765)}

AttributeError: 'AxisLimits' object has no attribute 'populateDefaults'

fontcollector -i "01.ass"
ERROR - An unknown error occurred while reading the font "C:\Windows\Fonts\bahnschrift.ttf"
Please open an issue on github, share the font and the following error message:
Traceback (most recent call last):
  File "C:\Users\leah\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "C:\Users\leah\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "c:\users\leah\.local\bin\fontcollector.exe\__main__.py", line 7, in <module>
  File "C:\Users\leah\.local\pipx\venvs\fontcollector\lib\site-packages\font_collector\__main__.py", line 26, in main
    font_collection = FontLoader(additional_fonts, use_system_font).fonts
  File "C:\Users\leah\.local\pipx\venvs\fontcollector\lib\site-packages\font_collector\font_loader.py", line 25, in __init__
    self.system_fonts = FontLoader.load_system_fonts()
  File "C:\Users\leah\.local\pipx\venvs\fontcollector\lib\site-packages\font_collector\font_loader.py", line 97, in load_system_fonts
    system_fonts.update(Font.from_font_path(font_path))
  File "C:\Users\leah\.local\pipx\venvs\fontcollector\lib\site-packages\font_collector\font.py", line 76, in from_font_path
    axis_value_tables = FontParser.get_axis_value_from_coordinates(
  File "C:\Users\leah\.local\pipx\venvs\fontcollector\lib\site-packages\font_collector\font_parser.py", line 157, in get_axis_value_from_coordinates
    axisLimits = AxisLimits(coordinates).populateDefaults(font)
AttributeError: 'AxisLimits' object has no attribute 'populateDefaults'

Traceback error

When attempting to execute fontCollector, I encounter a traceback error. I have the most recent version of fontCollector (1.2.1), so not sure what's going wrong.

D:\Downloads\test>fontCollector -i "animorphs_01.ass" -o
Loaded successfully animorphs_01.ass
Matplotlib is building the font cache; this may take a moment.
Traceback (most recent call last):
  File "C:\Users\Animorphs\AppData\Local\Programs\Python\Python39\Scripts\fontCollector-script.py", line 33, in <module>
    sys.exit(load_entry_point('fontCollector==1.2.1', 'console_scripts', 'fontCollector')())
  File "c:\users\animorphs\appdata\local\programs\python\python39\lib\site-packages\fontCollector.py", line 630, in main
    fontCollection = initializeFontCollection(additionalFonts)
  File "c:\users\animorphs\appdata\local\programs\python\python39\lib\site-packages\fontCollector.py", line 480, in initializeFontCollection
    font = createFont(fontPath)
  File "c:\users\animorphs\appdata\local\programs\python\python39\lib\site-packages\fontCollector.py", line 438, in createFont
    isItalic = bool(fontTtLib["OS/2"].fsSelection & 1)
  File "C:\Users\animorphs\AppData\Roaming\Python\Python39\site-packages\fontTools\ttLib\ttFont.py", line 362, in __getitem__
    table = self._readTable(tag)
  File "C:\Users\animorphs\AppData\Roaming\Python\Python39\site-packages\fontTools\ttLib\ttFont.py", line 369, in _readTable
    data = self.reader[tag]
  File "C:\Users\animorphs\AppData\Roaming\Python\Python39\site-packages\fontTools\ttLib\sfnt.py", line 105, in __getitem__
    entry = self.tables[Tag(tag)]
KeyError: 'OS/2'

I've also tried it with a very simple ass file just in case, which solely consists of the following, but the same error occurs:

[Script Info]
; Script generated by Aegisub 9213, Daydream Cafe Edition [Shinon]
; http://www.aegisub.org/
ScriptType: v4.00+
PlayResX: 1920
PlayResY: 1080
WrapStyle: 0
ScaledBorderAndShadow: yes
YCbCr Matrix: TV.709

[Aegisub Project Garbage]

[V4+ Styles]
Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
Style: Default,Gandhi Sans,74,&H00FFFFFF,&H000000FF,&H00000000,&HA0000000,-1,0,0,0,100,100,0,0,1,3.6,1.5,2,180,180,56,1

[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:05.00,0:00:10.00,Default,,0,0,0,,Test.

Any help is appreciated, thanks

VSFilter/GDI truncate font name to 31 characters

Problem

VSFilter/GDI truncate font name to 31 characters.

Currently, Libass doesn't emulated this behaviour. See: libass/libass#459
Since it doesn't emulated it, this won't be implemente until libass does.

Which test I did to know how GDI works.

All the test I did was with this font: A-Farzian-2-Regular.zip
I rename the font with different name to see what name GDI would return me:

from fontTools.ttLib.ttFont import TTFont

font = TTFont("A-Farzian-2-Regular.ttf")

names = [
    "a"*30 + "😋"
    "a"*31 + "😋",
    "a"*32,
    "😋" * 20,
]

for name in names:
    font["name"].setName(name, 1, 3, 1, 0x409)
    font.save(f"{name}.TTF")

Specific font causing error when collecting fonts on Windows

When running the script with a command:
fontCollector -i "test.ass", I get the following error:
ERROR - An unknown error occurred while reading the font "C:\USERS\BEGIJ\FONTBASE\FONTS\MY COLLECTION\BOXED_IN1.TTF".
If I remove this specific font, script finishes successfully. I'm on Windows 11 and using 3.0.0c1 of Fontcollector.
I'm attaching this font and test file as a google drive link
Ass and Font

[Feature] Pass font root directory instead of individual font folders

I have a root font folder that contains folders for each font family.

The --additional-fonts option isn't recursive, so I have to pass a list of all the font family folders on the command line.

It would be very helpful if there was a --font-root argument that would recursively search one (or more) folders for font files.

There is a potential problem that can cause an error when running workflows, especially on Windows.

In the Install platform-specific requirements (Windows) step, after installing mkvtoolnix, you are attempting to run mkvpropedit -h. The issue here is that the environment variables set or modified in one step are not automatically carried over to the next step.

To address this, you may need to persist the changes to the environment variables

Some fontName are not correctly decoded

Font example: https://mega.nz/file/S9ERDRpQ#bcPhS06kv-D5jt64aTNDbZVd6gZr6ZfJDYT91yYsoWk

When fontCollector.py decoded the fontName, it returns: f j000 3 1 0

Here is what Aegisub return: Fjイーマ310

Currently, I have no idea how to solve this problem. Here is the documentation about the naming table of fonttools: https://fonttools.readthedocs.io/en/latest/ttLib/tables/_n_a_m_e.html#fontTools.ttLib.tables._n_a_m_e.NameRecord

FontCollector is too slow

@matheousse contacted me.

He sent me this .ass and the font that are needed for the .ass file: Font + ass.zip

It take him 1 minute to collec the font which is way to slow.

FontCollector can't create folders

Seems like FontCollector can't create folders on Windows. When I'm setting an output path to an not existing folder, I get an error. Would be nice if the program could create the named folder instead.

Support duplicate style name

VSFilter can handle duplicate style name by adding an "id" AND if the style name is empty, it rename it to Default.
FontCollector doesn't replicate this behavior.

Actually, it will even raise an exception: ValueError: Error: Unknown style "Default3" on line 1. You need to correct the .ass file.
Here is a .ass sample that cause problems

[Script Info]
Title: Default Aegisub file
ScriptType: v4.00+
WrapStyle: 0
ScaledBorderAndShadow: yes
YCbCr Matrix: None

[V4+ Styles]
Format: Name,   Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
Style: Default,Raleway,35,&H00FFFFFF,&H000000FF,&H002C2C2C,&H00000000,-1,0,0,0,100,100,0,0,1,2.5,0,2,0,0,72,1
Style: Default,Jester,35,&H00FFFFFF,&H000000FF,&H002C2C2C,&H00000000,-1,0,0,0,100,100,0,0,1,2.5,0,2,0,0,72,1
Style: ,Arial,35,&H00FFFFFF,&H000000FF,&H002C2C2C,&H00000000,-1,1,0,0,100,100,0,0,1,2.5,0,2,0,0,72,1

[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:00.00,0:00:05.00,Default3,,0,0,0,,Arial

Clear the font cache on uninstall

When the user uninstall the program with pip uninstall FontCollector, it won't clear the font cache.
To actually clear the cache, the user needs to call FontLoader.discard_system_font_cache() and FontLoader.discard_generated_font_cache().

It seems that it isn't possible to run custom code when the user do a pip uninstall FontCollector. See Custom code on pip uninstall.
Because of that, I don't really know how I could clear the font cache on uninstall.