Files
crosspoint-reader-mod/lib/EpdFont/scripts/fontconvert.py

424 lines
15 KiB
Python
Raw Normal View History

2025-12-03 22:00:29 +11:00
#!python3
import freetype
import zlib
import sys
import re
import math
import argparse
from collections import namedtuple
# Originally from https://github.com/vroland/epdiy
2025-12-03 22:00:29 +11:00
parser = argparse.ArgumentParser(description="Generate a header file from a font to be used with epdiy.")
parser.add_argument("name", action="store", help="name of the font.")
parser.add_argument("size", type=int, help="font size to use.")
parser.add_argument("fontstack", action="store", nargs='+', help="list of font files, ordered by descending priority.")
parser.add_argument("--2bit", dest="is2Bit", action="store_true", help="generate 2-bit greyscale bitmap instead of 1-bit black and white.")
2025-12-03 22:00:29 +11:00
parser.add_argument("--additional-intervals", dest="additional_intervals", action="append", help="Additional code point intervals to export as min,max. This argument can be repeated.")
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
parser.add_argument("--compress", dest="compress", action="store_true", help="Compress glyph bitmaps using DEFLATE with group-based compression.")
fix: force auto-hinting for Bookerly to fix inconsistent stem widths (#1098) ## Summary Bookerly's native TrueType hinting is effectively a no-op at the sizes used here, causing FreeType to place stems at inconsistent sub-pixel positions. This results in the 'k' stem (8-bit fringe: 0x38=56) falling just below the 2-bit quantization threshold while 'l' and 'h' stems (fringes: 0x4C=76, 0x40=64) land above it --- making 'k' visibly narrower (2.00px vs 2.33px effective width). FreeType's auto-hinter snaps all stems to consistent grid positions, normalizing effective stem width to 2.67px across all glyphs. Adds --force-autohint flag to fontconvert.py and applies it to Bookerly only. NotoSans, OpenDyslexic, and Ubuntu fonts are unaffected. Here is an example of before/after. Take notice of the vertical stems on characters like `l`, `k`, `n`, `i`, etc. The font is Bookerly 12pt regular: **BEFORE**: ![before](https://github.com/user-attachments/assets/65b2acab-ad95-489e-885e-e3a0163cc252) **AFTER**: ![after](https://github.com/user-attachments/assets/d09a8b5d-40af-4a7d-b622-e1b2cabcce85) Claude generated this script to quantitatively determine that this change makes the vertical stems on a variety of characters more consistent for Bookerly _only_. <details> <summary>Python script</summary> ```python #!/usr/bin/env python3 """Compare stem consistency across all font families with and without auto-hinting. Run from repo root: python3 compare_all_fonts.py """ import freetype DPI = 150 CHARS = ["k", "l", "h", "i", "b", "d"] SIZES = [12, 14, 16, 18] FONTS = { "Bookerly": "lib/EpdFont/builtinFonts/source/Bookerly/Bookerly-Regular.ttf", "NotoSans": "lib/EpdFont/builtinFonts/source/NotoSans/NotoSans-Regular.ttf", "OpenDyslexic": "lib/EpdFont/builtinFonts/source/OpenDyslexic/OpenDyslexic-Regular.otf", "Ubuntu": "lib/EpdFont/builtinFonts/source/Ubuntu/Ubuntu-Regular.ttf", } MODES = { "default": freetype.FT_LOAD_RENDER, "autohint": freetype.FT_LOAD_RENDER | freetype.FT_LOAD_FORCE_AUTOHINT, } def q4to2(v): if v >= 12: return 3 elif v >= 8: return 2 elif v >= 4: return 1 else: return 0 def get_stem_eff(face, char, flags): gi = face.get_char_index(ord(char)) if gi == 0: return None face.load_glyph(gi, flags) bm = face.glyph.bitmap w, h = bm.width, bm.rows if w == 0 or h == 0: return None p2 = [] for y in range(h): row = [] for x in range(w): row.append(q4to2(bm.buffer[y * bm.pitch + x] >> 4)) p2.append(row) # Measure leftmost stem in stable middle rows mid_start, mid_end = h // 4, h - h // 4 widths = [] for y in range(mid_start, mid_end): first = next((x for x in range(w) if p2[y][x] > 0), -1) if first < 0: continue last = first for x in range(first, w): if p2[y][x] > 0: last = x else: break eff = sum(p2[y][x] for x in range(first, last + 1)) / 3.0 widths.append(eff) return round(sum(widths) / len(widths), 2) if widths else None def main(): for font_name, font_path in FONTS.items(): try: freetype.Face(font_path) except Exception: print(f"\n {font_name}: SKIPPED (file not found)") continue print(f"\n{'=' * 80}") print(f" {font_name}") print(f"{'=' * 80}") for size in SIZES: print(f"\n {size}pt:") print(f" {'':6s}", end="") for c in CHARS: print(f" '{c}' ", end="") print(" | spread") for mode_name, flags in MODES.items(): face = freetype.Face(font_path) face.set_char_size(size << 6, size << 6, DPI, DPI) vals = [] print(f" {mode_name:6s}", end="") for c in CHARS: v = get_stem_eff(face, c, flags) vals.append(v) print(f" {v:5.2f}" if v else " N/A", end="") valid = [v for v in vals if v is not None] spread = max(valid) - min(valid) if len(valid) >= 2 else 0 marker = " <-- inconsistent" if spread > 0.5 else "" print(f" | {spread:.2f}{marker}") if __name__ == "__main__": main() ``` </details> Here are the results. The table compares how the font-generation `autohint` flag affects the range of widths of various characters. Lower `spread` mean that glyph stroke widths should appear more consistent. ``` Spread = max stem width - min stem width across glyphs (lower = more consistent): ┌──────────────┬──────┬─────────┬──────────┬──────────┐ │ Font │ Size │ Default │ Autohint │ Winner │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Bookerly │ 12pt │ 1.49 │ 1.12 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.39 │ 1.13 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.38 │ 1.16 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.90 │ 1.58 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ NotoSans │ 12pt │ 1.16 │ 0.94 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 0.83 │ 1.14 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.41 │ 1.51 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.74 │ 1.63 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ OpenDyslexic │ 12pt │ 2.22 │ 1.44 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 2.57 │ 3.29 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 3.13 │ 2.60 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 3.21 │ 3.23 │ ~tied │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Ubuntu │ 12pt │ 1.25 │ 1.31 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.41 │ 1.64 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 2.21 │ 1.71 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.80 │ 1.71 │ autohint │ └──────────────┴──────┴─────────┴──────────┴──────────┘ ``` --- ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? I used AI to make sure I'm not doing something stupid, since I'm not a typography expert. I made the changes though.
2026-02-24 06:13:08 +11:00
parser.add_argument("--force-autohint", dest="force_autohint", action="store_true", help="Force FreeType auto-hinter instead of native font hinting. Improves stem width consistency for fonts with weak or no native TrueType hints.")
2025-12-03 22:00:29 +11:00
args = parser.parse_args()
GlyphProps = namedtuple("GlyphProps", ["width", "height", "advance_x", "left", "top", "data_length", "data_offset", "code_point"])
2025-12-03 22:00:29 +11:00
font_stack = [freetype.Face(f) for f in args.fontstack]
is2Bit = args.is2Bit
2025-12-03 22:00:29 +11:00
size = args.size
font_name = args.name
fix: force auto-hinting for Bookerly to fix inconsistent stem widths (#1098) ## Summary Bookerly's native TrueType hinting is effectively a no-op at the sizes used here, causing FreeType to place stems at inconsistent sub-pixel positions. This results in the 'k' stem (8-bit fringe: 0x38=56) falling just below the 2-bit quantization threshold while 'l' and 'h' stems (fringes: 0x4C=76, 0x40=64) land above it --- making 'k' visibly narrower (2.00px vs 2.33px effective width). FreeType's auto-hinter snaps all stems to consistent grid positions, normalizing effective stem width to 2.67px across all glyphs. Adds --force-autohint flag to fontconvert.py and applies it to Bookerly only. NotoSans, OpenDyslexic, and Ubuntu fonts are unaffected. Here is an example of before/after. Take notice of the vertical stems on characters like `l`, `k`, `n`, `i`, etc. The font is Bookerly 12pt regular: **BEFORE**: ![before](https://github.com/user-attachments/assets/65b2acab-ad95-489e-885e-e3a0163cc252) **AFTER**: ![after](https://github.com/user-attachments/assets/d09a8b5d-40af-4a7d-b622-e1b2cabcce85) Claude generated this script to quantitatively determine that this change makes the vertical stems on a variety of characters more consistent for Bookerly _only_. <details> <summary>Python script</summary> ```python #!/usr/bin/env python3 """Compare stem consistency across all font families with and without auto-hinting. Run from repo root: python3 compare_all_fonts.py """ import freetype DPI = 150 CHARS = ["k", "l", "h", "i", "b", "d"] SIZES = [12, 14, 16, 18] FONTS = { "Bookerly": "lib/EpdFont/builtinFonts/source/Bookerly/Bookerly-Regular.ttf", "NotoSans": "lib/EpdFont/builtinFonts/source/NotoSans/NotoSans-Regular.ttf", "OpenDyslexic": "lib/EpdFont/builtinFonts/source/OpenDyslexic/OpenDyslexic-Regular.otf", "Ubuntu": "lib/EpdFont/builtinFonts/source/Ubuntu/Ubuntu-Regular.ttf", } MODES = { "default": freetype.FT_LOAD_RENDER, "autohint": freetype.FT_LOAD_RENDER | freetype.FT_LOAD_FORCE_AUTOHINT, } def q4to2(v): if v >= 12: return 3 elif v >= 8: return 2 elif v >= 4: return 1 else: return 0 def get_stem_eff(face, char, flags): gi = face.get_char_index(ord(char)) if gi == 0: return None face.load_glyph(gi, flags) bm = face.glyph.bitmap w, h = bm.width, bm.rows if w == 0 or h == 0: return None p2 = [] for y in range(h): row = [] for x in range(w): row.append(q4to2(bm.buffer[y * bm.pitch + x] >> 4)) p2.append(row) # Measure leftmost stem in stable middle rows mid_start, mid_end = h // 4, h - h // 4 widths = [] for y in range(mid_start, mid_end): first = next((x for x in range(w) if p2[y][x] > 0), -1) if first < 0: continue last = first for x in range(first, w): if p2[y][x] > 0: last = x else: break eff = sum(p2[y][x] for x in range(first, last + 1)) / 3.0 widths.append(eff) return round(sum(widths) / len(widths), 2) if widths else None def main(): for font_name, font_path in FONTS.items(): try: freetype.Face(font_path) except Exception: print(f"\n {font_name}: SKIPPED (file not found)") continue print(f"\n{'=' * 80}") print(f" {font_name}") print(f"{'=' * 80}") for size in SIZES: print(f"\n {size}pt:") print(f" {'':6s}", end="") for c in CHARS: print(f" '{c}' ", end="") print(" | spread") for mode_name, flags in MODES.items(): face = freetype.Face(font_path) face.set_char_size(size << 6, size << 6, DPI, DPI) vals = [] print(f" {mode_name:6s}", end="") for c in CHARS: v = get_stem_eff(face, c, flags) vals.append(v) print(f" {v:5.2f}" if v else " N/A", end="") valid = [v for v in vals if v is not None] spread = max(valid) - min(valid) if len(valid) >= 2 else 0 marker = " <-- inconsistent" if spread > 0.5 else "" print(f" | {spread:.2f}{marker}") if __name__ == "__main__": main() ``` </details> Here are the results. The table compares how the font-generation `autohint` flag affects the range of widths of various characters. Lower `spread` mean that glyph stroke widths should appear more consistent. ``` Spread = max stem width - min stem width across glyphs (lower = more consistent): ┌──────────────┬──────┬─────────┬──────────┬──────────┐ │ Font │ Size │ Default │ Autohint │ Winner │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Bookerly │ 12pt │ 1.49 │ 1.12 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.39 │ 1.13 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.38 │ 1.16 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.90 │ 1.58 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ NotoSans │ 12pt │ 1.16 │ 0.94 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 0.83 │ 1.14 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.41 │ 1.51 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.74 │ 1.63 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ OpenDyslexic │ 12pt │ 2.22 │ 1.44 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 2.57 │ 3.29 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 3.13 │ 2.60 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 3.21 │ 3.23 │ ~tied │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Ubuntu │ 12pt │ 1.25 │ 1.31 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.41 │ 1.64 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 2.21 │ 1.71 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.80 │ 1.71 │ autohint │ └──────────────┴──────┴─────────┴──────────┴──────────┘ ``` --- ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? I used AI to make sure I'm not doing something stupid, since I'm not a typography expert. I made the changes though.
2026-02-24 06:13:08 +11:00
load_flags = freetype.FT_LOAD_RENDER
if args.force_autohint:
load_flags |= freetype.FT_LOAD_FORCE_AUTOHINT
2025-12-03 22:00:29 +11:00
# inclusive unicode code point intervals
# must not overlap and be in ascending order
intervals = [
### Basic Latin ###
# ASCII letters, digits, punctuation, control characters
(0x0000, 0x007F),
### Latin-1 Supplement ###
# Accented characters for Western European languages
(0x0080, 0x00FF),
### Latin Extended-A ###
# Eastern European and Baltic languages
(0x0100, 0x017F),
### General Punctuation (core subset) ###
# Smart quotes, en dash, em dash, ellipsis, NO-BREAK SPACE
(0x2000, 0x206F),
### Basic Symbols From "Latin-1 + Misc" ###
# dashes, quotes, prime marks
(0x2010, 0x203A),
# misc punctuation
(0x2040, 0x205F),
# common currency symbols
(0x20A0, 0x20CF),
### Combining Diacritical Marks (minimal subset) ###
# Needed for proper rendering of many extended Latin languages
(0x0300, 0x036F),
### Greek & Coptic ###
# Used in science, maths, philosophy, some academic texts
# (0x0370, 0x03FF),
### Cyrillic ###
# Russian, Ukrainian, Bulgarian, etc.
(0x0400, 0x04FF),
2025-12-03 22:00:29 +11:00
### Math Symbols (common subset) ###
# Superscripts and Subscripts
(0x2070, 0x209F),
2025-12-03 22:00:29 +11:00
# General math operators
(0x2200, 0x22FF),
# Arrows
(0x2190, 0x21FF),
### CJK ###
# Core Unified Ideographs
# (0x4E00, 0x9FFF),
# # Extension A
# (0x3400, 0x4DBF),
# # Extension B
# (0x20000, 0x2A6DF),
# # Extension CF
# (0x2A700, 0x2EBEF),
# # Extension G
# (0x30000, 0x3134F),
# # Hiragana
# (0x3040, 0x309F),
# # Katakana
# (0x30A0, 0x30FF),
# # Katakana Phonetic Extensions
# (0x31F0, 0x31FF),
# # Halfwidth Katakana
# (0xFF60, 0xFF9F),
# # Hangul Syllables
# (0xAC00, 0xD7AF),
# # Hangul Jamo
# (0x1100, 0x11FF),
# # Hangul Compatibility Jamo
# (0x3130, 0x318F),
# # Hangul Jamo Extended-A
# (0xA960, 0xA97F),
# # Hangul Jamo Extended-B
# (0xD7B0, 0xD7FF),
# # CJK Radicals Supplement
# (0x2E80, 0x2EFF),
# # Kangxi Radicals
# (0x2F00, 0x2FDF),
# # CJK Symbols and Punctuation
# (0x3000, 0x303F),
# # CJK Compatibility Forms
# (0xFE30, 0xFE4F),
# # CJK Compatibility Ideographs
# (0xF900, 0xFAFF),
### Specials
# Replacement Character
(0xFFFD, 0xFFFD),
2025-12-03 22:00:29 +11:00
]
add_ints = []
if args.additional_intervals:
add_ints = [tuple([int(n, base=0) for n in i.split(",")]) for i in args.additional_intervals]
def norm_floor(val):
return int(math.floor(val / (1 << 6)))
def norm_ceil(val):
return int(math.ceil(val / (1 << 6)))
def chunks(l, n):
for i in range(0, len(l), n):
yield l[i:i + n]
def load_glyph(code_point):
face_index = 0
while face_index < len(font_stack):
face = font_stack[face_index]
glyph_index = face.get_char_index(code_point)
if glyph_index > 0:
fix: force auto-hinting for Bookerly to fix inconsistent stem widths (#1098) ## Summary Bookerly's native TrueType hinting is effectively a no-op at the sizes used here, causing FreeType to place stems at inconsistent sub-pixel positions. This results in the 'k' stem (8-bit fringe: 0x38=56) falling just below the 2-bit quantization threshold while 'l' and 'h' stems (fringes: 0x4C=76, 0x40=64) land above it --- making 'k' visibly narrower (2.00px vs 2.33px effective width). FreeType's auto-hinter snaps all stems to consistent grid positions, normalizing effective stem width to 2.67px across all glyphs. Adds --force-autohint flag to fontconvert.py and applies it to Bookerly only. NotoSans, OpenDyslexic, and Ubuntu fonts are unaffected. Here is an example of before/after. Take notice of the vertical stems on characters like `l`, `k`, `n`, `i`, etc. The font is Bookerly 12pt regular: **BEFORE**: ![before](https://github.com/user-attachments/assets/65b2acab-ad95-489e-885e-e3a0163cc252) **AFTER**: ![after](https://github.com/user-attachments/assets/d09a8b5d-40af-4a7d-b622-e1b2cabcce85) Claude generated this script to quantitatively determine that this change makes the vertical stems on a variety of characters more consistent for Bookerly _only_. <details> <summary>Python script</summary> ```python #!/usr/bin/env python3 """Compare stem consistency across all font families with and without auto-hinting. Run from repo root: python3 compare_all_fonts.py """ import freetype DPI = 150 CHARS = ["k", "l", "h", "i", "b", "d"] SIZES = [12, 14, 16, 18] FONTS = { "Bookerly": "lib/EpdFont/builtinFonts/source/Bookerly/Bookerly-Regular.ttf", "NotoSans": "lib/EpdFont/builtinFonts/source/NotoSans/NotoSans-Regular.ttf", "OpenDyslexic": "lib/EpdFont/builtinFonts/source/OpenDyslexic/OpenDyslexic-Regular.otf", "Ubuntu": "lib/EpdFont/builtinFonts/source/Ubuntu/Ubuntu-Regular.ttf", } MODES = { "default": freetype.FT_LOAD_RENDER, "autohint": freetype.FT_LOAD_RENDER | freetype.FT_LOAD_FORCE_AUTOHINT, } def q4to2(v): if v >= 12: return 3 elif v >= 8: return 2 elif v >= 4: return 1 else: return 0 def get_stem_eff(face, char, flags): gi = face.get_char_index(ord(char)) if gi == 0: return None face.load_glyph(gi, flags) bm = face.glyph.bitmap w, h = bm.width, bm.rows if w == 0 or h == 0: return None p2 = [] for y in range(h): row = [] for x in range(w): row.append(q4to2(bm.buffer[y * bm.pitch + x] >> 4)) p2.append(row) # Measure leftmost stem in stable middle rows mid_start, mid_end = h // 4, h - h // 4 widths = [] for y in range(mid_start, mid_end): first = next((x for x in range(w) if p2[y][x] > 0), -1) if first < 0: continue last = first for x in range(first, w): if p2[y][x] > 0: last = x else: break eff = sum(p2[y][x] for x in range(first, last + 1)) / 3.0 widths.append(eff) return round(sum(widths) / len(widths), 2) if widths else None def main(): for font_name, font_path in FONTS.items(): try: freetype.Face(font_path) except Exception: print(f"\n {font_name}: SKIPPED (file not found)") continue print(f"\n{'=' * 80}") print(f" {font_name}") print(f"{'=' * 80}") for size in SIZES: print(f"\n {size}pt:") print(f" {'':6s}", end="") for c in CHARS: print(f" '{c}' ", end="") print(" | spread") for mode_name, flags in MODES.items(): face = freetype.Face(font_path) face.set_char_size(size << 6, size << 6, DPI, DPI) vals = [] print(f" {mode_name:6s}", end="") for c in CHARS: v = get_stem_eff(face, c, flags) vals.append(v) print(f" {v:5.2f}" if v else " N/A", end="") valid = [v for v in vals if v is not None] spread = max(valid) - min(valid) if len(valid) >= 2 else 0 marker = " <-- inconsistent" if spread > 0.5 else "" print(f" | {spread:.2f}{marker}") if __name__ == "__main__": main() ``` </details> Here are the results. The table compares how the font-generation `autohint` flag affects the range of widths of various characters. Lower `spread` mean that glyph stroke widths should appear more consistent. ``` Spread = max stem width - min stem width across glyphs (lower = more consistent): ┌──────────────┬──────┬─────────┬──────────┬──────────┐ │ Font │ Size │ Default │ Autohint │ Winner │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Bookerly │ 12pt │ 1.49 │ 1.12 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.39 │ 1.13 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.38 │ 1.16 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.90 │ 1.58 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ NotoSans │ 12pt │ 1.16 │ 0.94 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 0.83 │ 1.14 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 1.41 │ 1.51 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.74 │ 1.63 │ mixed │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ OpenDyslexic │ 12pt │ 2.22 │ 1.44 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 2.57 │ 3.29 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 3.13 │ 2.60 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 3.21 │ 3.23 │ ~tied │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ Ubuntu │ 12pt │ 1.25 │ 1.31 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 14pt │ 1.41 │ 1.64 │ default │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 16pt │ 2.21 │ 1.71 │ autohint │ ├──────────────┼──────┼─────────┼──────────┼──────────┤ │ │ 18pt │ 1.80 │ 1.71 │ autohint │ └──────────────┴──────┴─────────┴──────────┴──────────┘ ``` --- ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? I used AI to make sure I'm not doing something stupid, since I'm not a typography expert. I made the changes though.
2026-02-24 06:13:08 +11:00
face.load_glyph(glyph_index, load_flags)
2025-12-03 22:00:29 +11:00
return face
face_index += 1
print(f"code point {code_point} ({hex(code_point)}) not found in font stack!", file=sys.stderr)
return None
unmerged_intervals = sorted(intervals + add_ints)
intervals = []
unvalidated_intervals = []
for i_start, i_end in unmerged_intervals:
if len(unvalidated_intervals) > 0 and i_start + 1 <= unvalidated_intervals[-1][1]:
unvalidated_intervals[-1] = (unvalidated_intervals[-1][0], max(unvalidated_intervals[-1][1], i_end))
continue
unvalidated_intervals.append((i_start, i_end))
for i_start, i_end in unvalidated_intervals:
start = i_start
for code_point in range(i_start, i_end + 1):
face = load_glyph(code_point)
if face is None:
if start < code_point:
intervals.append((start, code_point - 1))
start = code_point + 1
if start != i_end + 1:
intervals.append((start, i_end))
for face in font_stack:
face.set_char_size(size << 6, size << 6, 150, 150)
total_size = 0
all_glyphs = []
for i_start, i_end in intervals:
for code_point in range(i_start, i_end + 1):
face = load_glyph(code_point)
bitmap = face.glyph.bitmap
# Build out 4-bit greyscale bitmap
pixels4g = []
2025-12-03 22:00:29 +11:00
px = 0
for i, v in enumerate(bitmap.buffer):
y = i / bitmap.width
x = i % bitmap.width
if x % 2 == 0:
px = (v >> 4)
else:
px = px | (v & 0xF0)
pixels4g.append(px);
2025-12-03 22:00:29 +11:00
px = 0
# eol
if x == bitmap.width - 1 and bitmap.width % 2 > 0:
pixels4g.append(px)
2025-12-03 22:00:29 +11:00
px = 0
if is2Bit:
# 0-3 white, 4-7 light grey, 8-11 dark grey, 12-15 black
# Downsample to 2-bit bitmap
pixels2b = []
px = 0
pitch = (bitmap.width // 2) + (bitmap.width % 2)
for y in range(bitmap.rows):
for x in range(bitmap.width):
px = px << 2
bm = pixels4g[y * pitch + (x // 2)]
bm = (bm >> ((x % 2) * 4)) & 0xF
if bm >= 12:
px += 3
elif bm >= 8:
px += 2
elif bm >= 4:
px += 1
if (y * bitmap.width + x) % 4 == 3:
pixels2b.append(px)
px = 0
if (bitmap.width * bitmap.rows) % 4 != 0:
px = px << (4 - (bitmap.width * bitmap.rows) % 4) * 2
pixels2b.append(px)
# for y in range(bitmap.rows):
# line = ''
# for x in range(bitmap.width):
# pixelPosition = y * bitmap.width + x
# byte = pixels2b[pixelPosition // 4]
# bit_index = (3 - (pixelPosition % 4)) * 2
# line += '#' if ((byte >> bit_index) & 3) > 0 else '.'
# print(line)
# print('')
else:
# Downsample to 1-bit bitmap - treat any 2+ as black
pixelsbw = []
px = 0
pitch = (bitmap.width // 2) + (bitmap.width % 2)
for y in range(bitmap.rows):
for x in range(bitmap.width):
px = px << 1
bm = pixels4g[y * pitch + (x // 2)]
px += 1 if ((x & 1) == 0 and bm & 0xE > 0) or ((x & 1) == 1 and bm & 0xE0 > 0) else 0
if (y * bitmap.width + x) % 8 == 7:
pixelsbw.append(px)
px = 0
if (bitmap.width * bitmap.rows) % 8 != 0:
px = px << (8 - (bitmap.width * bitmap.rows) % 8)
pixelsbw.append(px)
# for y in range(bitmap.rows):
# line = ''
# for x in range(bitmap.width):
# pixelPosition = y * bitmap.width + x
# byte = pixelsbw[pixelPosition // 8]
# bit_index = 7 - (pixelPosition % 8)
# line += '#' if (byte >> bit_index) & 1 else '.'
# print(line)
# print('')
pixels = pixels2b if is2Bit else pixelsbw
# Build output data
packed = bytes(pixels)
2025-12-03 22:00:29 +11:00
glyph = GlyphProps(
width = bitmap.width,
height = bitmap.rows,
advance_x = norm_floor(face.glyph.advance.x),
left = face.glyph.bitmap_left,
top = face.glyph.bitmap_top,
data_length = len(packed),
2025-12-03 22:00:29 +11:00
data_offset = total_size,
code_point = code_point,
)
total_size += len(packed)
all_glyphs.append((glyph, packed))
2025-12-03 22:00:29 +11:00
# pipe seems to be a good heuristic for the "real" descender
face = load_glyph(ord('|'))
glyph_data = []
glyph_props = []
for index, glyph in enumerate(all_glyphs):
props, packed = glyph
glyph_data.extend([b for b in packed])
2025-12-03 22:00:29 +11:00
glyph_props.append(props)
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
compress = args.compress
# Build groups for compression
if compress:
# Script-based grouping: glyphs that co-occur in typical text rendering
# are grouped together for efficient LRU caching on the embedded target.
# Since glyphs are in codepoint order, glyphs in the same Unicode block
# are contiguous in the array and form natural groups.
SCRIPT_GROUP_RANGES = [
(0x0000, 0x007F), # ASCII
(0x0080, 0x00FF), # Latin-1 Supplement
(0x0100, 0x017F), # Latin Extended-A
(0x0300, 0x036F), # Combining Diacritical Marks
(0x0400, 0x04FF), # Cyrillic
(0x2000, 0x206F), # General Punctuation
(0x2070, 0x209F), # Superscripts & Subscripts
(0x20A0, 0x20CF), # Currency Symbols
(0x2190, 0x21FF), # Arrows
(0x2200, 0x22FF), # Math Operators
(0xFFFD, 0xFFFD), # Replacement Character
]
def get_script_group(code_point):
for i, (start, end) in enumerate(SCRIPT_GROUP_RANGES):
if start <= code_point <= end:
return i
return -1
groups = [] # list of (first_glyph_index, glyph_count)
current_group_id = None
group_start = 0
group_count = 0
for i, (props, packed) in enumerate(all_glyphs):
sg = get_script_group(props.code_point)
if sg != current_group_id:
if group_count > 0:
groups.append((group_start, group_count))
current_group_id = sg
group_start = i
group_count = 1
else:
group_count += 1
if group_count > 0:
groups.append((group_start, group_count))
# Compress each group
compressed_groups = [] # list of (compressed_bytes, uncompressed_size, glyph_count, first_glyph_index)
compressed_bitmap_data = []
compressed_offset = 0
# Also build modified glyph props with within-group offsets
modified_glyph_props = list(glyph_props)
for first_idx, count in groups:
# Concatenate bitmap data for this group
group_data = b''
for gi in range(first_idx, first_idx + count):
props, packed = all_glyphs[gi]
# Update glyph's dataOffset to be within-group offset
within_group_offset = len(group_data)
old_props = modified_glyph_props[gi]
modified_glyph_props[gi] = GlyphProps(
width=old_props.width,
height=old_props.height,
advance_x=old_props.advance_x,
left=old_props.left,
top=old_props.top,
data_length=old_props.data_length,
data_offset=within_group_offset,
code_point=old_props.code_point,
)
group_data += packed
# Compress with raw DEFLATE (no zlib/gzip header)
compressor = zlib.compressobj(level=9, wbits=-15)
compressed = compressor.compress(group_data) + compressor.flush()
compressed_groups.append((compressed, len(group_data), count, first_idx))
compressed_bitmap_data.extend(compressed)
compressed_offset += len(compressed)
glyph_props = modified_glyph_props
total_compressed = len(compressed_bitmap_data)
total_uncompressed = len(glyph_data)
print(f"// Compression: {total_uncompressed} -> {total_compressed} bytes ({100*total_compressed/total_uncompressed:.1f}%), {len(groups)} groups", file=sys.stderr)
print(f"""/**
* generated by fontconvert.py
* name: {font_name}
* size: {size}
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
* mode: {'2-bit' if is2Bit else '1-bit'}{' compressed: true' if compress else ''}
* Command used: {' '.join(sys.argv)}
*/
#pragma once
#include "EpdFontData.h"
""")
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
if compress:
print(f"static const uint8_t {font_name}Bitmaps[{len(compressed_bitmap_data)}] = {{")
for c in chunks(compressed_bitmap_data, 16):
print (" " + " ".join(f"0x{b:02X}," for b in c))
print ("};\n");
else:
print(f"static const uint8_t {font_name}Bitmaps[{len(glyph_data)}] = {{")
for c in chunks(glyph_data, 16):
print (" " + " ".join(f"0x{b:02X}," for b in c))
print ("};\n");
2025-12-03 22:00:29 +11:00
print(f"static const EpdGlyph {font_name}Glyphs[] = {{")
for i, g in enumerate(glyph_props):
print (" { " + ", ".join([f"{a}" for a in list(g[:-1])]),"},", f"// {chr(g.code_point) if g.code_point != 92 else '<backslash>'}")
print ("};\n");
print(f"static const EpdUnicodeInterval {font_name}Intervals[] = {{")
offset = 0
for i_start, i_end in intervals:
print (f" {{ 0x{i_start:X}, 0x{i_end:X}, 0x{offset:X} }},")
offset += i_end - i_start + 1
print ("};\n");
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
if compress:
print(f"static const EpdFontGroup {font_name}Groups[] = {{")
compressed_offset = 0
for compressed, uncompressed_size, count, first_idx in compressed_groups:
print(f" {{ {compressed_offset}, {len(compressed)}, {uncompressed_size}, {count}, {first_idx} }},")
compressed_offset += len(compressed)
print("};\n")
2025-12-03 22:00:29 +11:00
print(f"static const EpdFontData {font_name} = {{")
print(f" {font_name}Bitmaps,")
print(f" {font_name}Glyphs,")
print(f" {font_name}Intervals,")
print(f" {len(intervals)},")
print(f" {norm_ceil(face.size.height)},")
print(f" {norm_ceil(face.size.ascender)},")
print(f" {norm_floor(face.size.descender)},")
print(f" {'true' if is2Bit else 'false'},")
perf: Reduce overall flash usage by 30.7% by compressing built-in fonts (#831) ## Summary **What is the goal of this PR?** Compress reader font bitmaps to reduce flash usage by 30.7%. **What changes are included?** - New `EpdFontGroup` struct and extended `EpdFontData` with `groups`/`groupCount` fields - `--compress` flag in `fontconvert.py`: groups glyphs (ASCII base group + groups of 8) and compresses each with raw DEFLATE - `FontDecompressor` class with 4-slot LRU cache for on-demand decompression during rendering - `GfxRenderer` transparently routes bitmap access through `getGlyphBitmap()` (compressed or direct flash) - Uses `uzlib` for decompression with minimal heap overhead. - 48 reader fonts (Bookerly, NotoSans 12-18pt, OpenDyslexic) regenerated with compression; 5 UI fonts unchanged - Round-trip verification script (`verify_compression.py`) runs as part of font generation ## Additional Context ## Flash & RAM | | baseline | font-compression | Difference | |--|--------|-----------------|------------| | Flash (ELF) | 6,302,476 B (96.2%) | 4,365,022 B (66.6%) | -1,937,454 B (-30.7%) | | firmware.bin | 6,468,192 B | 4,531,008 B | -1,937,184 B (-29.9%) | | RAM | 101,700 B (31.0%) | 103,076 B (31.5%) | +1,376 B (+0.5%) | ## Script-Based Grouping (Cold Cache) Comparison of uncompressed baseline vs script-based group compression (4-slot LRU cache, cleared each page). Glyphs are grouped by Unicode block (ASCII, Latin-1, Latin Extended-A, Combining Marks, Cyrillic, General Punctuation, etc.) instead of sequential groups of 8. ### Render Time | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Median** | 414.9 ms | 431.6 ms | +16.7 ms (+4.0%) | | **Pages** | 37 | 37 | | ### Memory Usage | | Baseline | Compressed (cold cache) | Difference | |---|---|---|---| | **Heap free (median)** | 187.0 KB | 176.3 KB | -10.7 KB | | **Heap free (min)** | 186.0 KB | 166.5 KB | -19.5 KB | | **Largest block (median)** | 148.0 KB | 128.0 KB | -20.0 KB | | **Largest block (min)** | 148.0 KB | 120.0 KB | -28.0 KB | ### Cache Effectiveness | | Misses/page | Hit rate | |---|---|---| | **Compressed (cold cache)** | 2.1 | 99.85% | ------ ### AI Usage While CrossPoint doesn't have restrictions on AI tools in contributing, please be transparent about their usage as it helps set the right context for reviewers. Did you use AI tools to help write this code? _**YES**_ Implementation was done by Claude Code (Opus 4.6) based on a plan developed collaboratively. All generated font headers were verified with an automated round-trip decompression test. The firmware was compiled successfully but has not yet been tested on-device. --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-19 20:30:15 +11:00
if compress:
print(f" {font_name}Groups,")
print(f" {len(compressed_groups)},")
else:
print(f" nullptr,")
print(f" 0,")
2025-12-03 22:00:29 +11:00
print("};")