Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Commit 2413d13

Browse files
story645story645
authored
Merge pull request #61 from aitikgupta/aitikgupta/gsoc-quarter
GSoC'21 Quarter Progress: Aitik Gupta
2 parents 3055fef + 379326f commit 2413d13

File tree

2 files changed

+144
-0
lines changed

2 files changed

+144
-0
lines changed

‎content/posts/GSoC_2021_Quarter/AitikGupta_GSoC.png‎

288 KB
Loading

‎content/posts/GSoC_2021_Quarter/index.md‎

Lines changed: 144 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,144 @@
1+
---
2+
title: "GSoC'21: Quarter Progress"
3+
date: 2021-08-03T18:48:00+05:30
4+
draft: false
5+
categories: ["News", "GSoC"]
6+
description: "Quarter Progress with Google Summer of Code 2021 project under NumFOCUS: Aitik Gupta"
7+
displayInList: true
8+
author: Aitik Gupta
9+
10+
resources:
11+
- name: featuredImage
12+
src: "AitikGupta_GSoC.png"
13+
params:
14+
showOnTop: true
15+
---
16+
17+
**“<ins>Matplotlib, I want 多个汉字 in between my text.</ins>”**
18+
19+
Let's say you asked Matplotlib to render a plot with some label containing 多个汉字 (multiple Chinese characters) in between your English text.
20+
21+
Or conversely, let's say you use a Chinese font with Matplotlib, but you had English text in between (which is quite common).
22+
23+
> Assumption: the Chinese font doesn't have those English glyphs, and vice versa
24+
25+
With this short writeup, I'll talk about how does a migration from a font-first to a text-first approach in Matplotlib looks like, which ideally solves the above problem.
26+
### Have the fonts?
27+
Logically, the very first step to solving this would be to ask whether you _have_ multiple fonts, right?
28+
29+
Matplotlib doesn't ship [CJK](https://en.wikipedia.org/wiki/List_of_CJK_fonts) (Chinese Japanese Korean) fonts, which ideally contains these Chinese glyphs. It does try to cover most grounds with the [default font](https://matplotlib.org/stable/users/dflt_style_changes.html#normal-text) it ships with, however.
30+
31+
So if you don't have a font to render your Chinese characters, go ahead and install one! Matplotlib will find your installed fonts (after rebuilding the cache, that is).
32+
### Parse the fonts
33+
This is where things get interesting, and what my [previous writeup](https://matplotlib.org/matplotblog/posts/gsoc_2021_prequarter/) was all about..
34+
35+
> Parsing the whole family to get multiple fonts for given font properties
36+
37+
## FT2Font Magic!
38+
To give you an idea about how things used to work for Matplotlib:
39+
1. A single font was chosen _at draw time_
40+
(fixed: re [previous writeup]((https://matplotlib.org/matplotblog/posts/gsoc_2021_prequarter/)))
41+
2. Every character displayed in your document was rendered by only that font
42+
(partially fixed: re <ins>_this writeup_</ins>)
43+
44+
> FT2Font is a matplotlib-to-font module, which provides high-level Python API to interact with a _single font's operations_ like read/draw/extract/etc.
45+
46+
Being written in C++, the module needs wrappers around it to be converted into a [Python extension](https://docs.python.org/3/extending/extending.html) using Python's C-API.
47+
48+
> It allows us to use C++ functions directly from Python!
49+
50+
So wherever you see a use of font within the library (by library I mean the readable Python codebase XD), you could have derived that:
51+
```
52+
FT2Font === SingleFont
53+
```
54+
55+
Things are be a bit different now however..
56+
## Designing a multi-font system
57+
FT2Font is basically itself a wrapper around a library called [FreeType](https://www.freetype.org/), which is a freely available software library to render fonts.
58+
59+
<p align="center">
60+
<figure>
61+
<img src="https://user-images.githubusercontent.com/43996118/128352387-76a3f52a-20fc-4853-b624-0c91844fc785.png" alt="FT2Font Naming" />
62+
<figcaption style="text-align: center; font-style: italic;">How FT2Font was named</figcaption>
63+
</figure>
64+
</p>
65+
66+
In my initial proposal.. while looking around how FT2Font is structured, I figured:
67+
```
68+
Oh, looks like all we need are Faces!
69+
```
70+
> If you don't know what faces/glyphs/ligatures are, head over to why [Text Hates You](https://gankra.github.io/blah/text-hates-you/). I can guarantee you'll definitely enjoy some real life examples of why text rendering is hard. 🥲
71+
72+
Anyway, if you already know what Faces are, it might strike you:
73+
74+
If we already have all the faces we need from multiple fonts (let's say we created a child of FT2Font.. which only <ins>tracks the faces</ins> for its families), we should be able to render everything from that parent FT2Font right?
75+
76+
As I later figured out while finding segfaults in implementing this design:
77+
```
78+
Each FT2Font is linked to a single FT_Library object!
79+
```
80+
81+
If you tried to load the face/glyph/character (basically anything) from a different FT2Font object.. you'll run into serious segfaults. (because one object linked to an `FT_Library` can't really access another object which has it's own `FT_Library`)
82+
```cpp
83+
// face is linked to FT2Font; which is
84+
// linked to a single FT_Library object
85+
FT_Face face = this->get_face();
86+
FT_Get_Glyph(face->glyph, &placeholder); // works like a charm
87+
88+
// somehow get another FT2Font's face
89+
FT_Face family_face = this->get_family_member()->get_face();
90+
FT_Get_Glyph(family_face->glyph, &placeholder); // segfaults!
91+
```
92+
93+
Realizing this took a good amount of time! After this I quickly came up with a recursive approach, wherein we:
94+
1. Create a list of FT2Font objects within Python, and pass it down to FT2Font
95+
2. FT2Font will hold pointers to its families via a \
96+
`std::vector<FT2Font *> fallback_list`
97+
3. Find if the character we want is available in the current font
98+
1. If the character is available, use that FT2Font to render that character
99+
2. If the character isn't found, go to step 3 again, but now iterate through the `fallback_list`
100+
4. That's it!
101+
102+
A quick overhaul of the above piece of code^
103+
```cpp
104+
bool ft_get_glyph(FT_Glyph &placeholder) {
105+
FT_Error not_found = FT_Get_Glyph(this->get_face(), &placeholder);
106+
if (not_found) return False;
107+
else return True;
108+
}
109+
110+
// within driver code
111+
for (uint i=0; i<fallback_list.size(); i++) {
112+
// iterate through all FT2Font objects
113+
bool was_found = fallback_list[i]->ft_get_glyph(placeholder);
114+
if (was_found) break;
115+
}
116+
```
117+
118+
With the idea surrounding this implementation, the [Agg backend](https://matplotlib.org/stable/api/backend_agg_api.html) is able to render a document (either through GUI, or a PNG) with multiple fonts!
119+
120+
<p align="center">
121+
<figure>
122+
<img src="https://user-images.githubusercontent.com/43996118/128347495-1f4f858d-33d3-4119-8732-5b26c4e9ca2a.png" alt="ChineseInBetween" />
123+
<figcaption style="text-align: center; font-style: italic;">PNG straight outta Matplotlib!</figcaption>
124+
</figure>
125+
</p>
126+
127+
## Python C-API is hard, at first!
128+
I've spent days at Python C-API's [argument doc](https://docs.python.org/3/c-api/arg.html), and it's hard to get what you need at first, ngl.
129+
130+
But, with the help of some amazing people in the GSoC community ([@srijan-paul](https://srijan-paul.github.io/), [@atharvaraykar](https://atharvaraykar.me/)) and amazing mentors, blockers begone!
131+
132+
## So are we done?
133+
Oh no. XD
134+
135+
Things work just fine for the Agg backend, but to generate a PDF/PS/SVG with multiple fonts is another story altogether! I think I'll save that for later.
136+
137+
<p align="center">
138+
<figure>
139+
<img src="https://user-images.githubusercontent.com/43996118/128350093-13695b91-5ad2-4f96-91f5-8373ee7a189e.gif" alt="ThankYouDwight" />
140+
<figcaption style="text-align: center; font-style: italic;">If you've been following the progress so far, mayn you're awesome!</figcaption>
141+
</figure>
142+
</p>
143+
144+
#### NOTE: This blog post is also available at my [personal website](https://aitikgupta.github.io/gsoc-quarter/).

0 commit comments

Comments
 (0)