ikat 0.11.0

Weave Markdown into camera-ready LaTeX: mermaid-to-TikZ, pgfplots figures, publisher templates
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269

"""Probe TeX for exact math-glyph widths and emit a calibrated Rust table.

Measures at the three math styles (display/text = 10pt, script = 7pt,
scriptscript = 5pt) every glyph, Greek letter, relation, binary, delimiter,
big operator, and operator name that appears in the corpus, plus construct
overheads (radical, fraction delimiters, matrix column separation, script
placement, thin spaces). Emits `glyphs_widths.json` and a Rust table.
"""
from __future__ import annotations

import json
import re
import subprocess
from pathlib import Path

HERE = Path(__file__).resolve().parent
WORK = HERE / "probe-w"

LOWERS = "abcdefghijklmnopqrstuvwxyz"
UPPERS = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
DIGITS = "0123456789"

GREEK = ["alpha", "beta", "gamma", "delta", "epsilon", "varepsilon", "zeta",
         "eta", "theta", "vartheta", "mu", "nu", "xi", "pi", "phi", "varphi",
         "rho", "sigma", "tau", "Gamma", "Delta", "Pi", "wp", "hbar", "infty"]
RELS = ["sim", "to", "leq", "geq", "in", "notin", "approx", "equiv", "neq"]
BINS = ["otimes", "times", "ast", "cdot"]
DELIMS = ["langle", "rangle"]
BIGOPS = ["sum", "int", "oint", "prod"]
OPNAMES = ["max", "min", "sup", "inf", "lim", "liminf", "limsup", "Pr",
           "cos", "sin", "tan", "log", "exp", "ker"]

STYLES = {"D": "displaystyle", "S": "scriptstyle", "SS": "scriptscriptstyle"}

lines: list[str] = [
    r"\documentclass{article}",
    r"\usepackage{amsmath,amssymb}",
    r"\begin{document}",
]


def probe(name: str, body: str) -> None:
    lines.append(rf"\sbox0{{$\displaystyle {body}$}}")
    lines.append(rf"\typeout{{{name} \the\wd0}}")


def probe_style(key: str, token: str, copies: int) -> None:
    """Width of `copies` consecutive tokens, per style."""
    body = token * copies
    for sk, sv in STYLES.items():
        probe(f"G|{sk}|{key}|{copies}", "\\" + sv + " " + body)


# --- single glyphs (8 copies for precision) ---
for c in LOWERS + UPPERS + DIGITS:
    probe_style(c, c, 8)
for g in GREEK:
    probe_style(g, "\\" + g, 8)
for g in RELS:
    probe_style(g, "\\" + g, 8)
probe_style("=", "=", 8)
probe_style("<", "<", 8)
probe_style(">", ">", 8)
probe_style("+", "+", 8)
probe_style("-", "-", 8)
probe_style("(", "(", 8)
probe_style(")", ")", 8)
probe_style("[", "[", 8)
probe_style("]", "]", 8)
for g in BINS:
    probe_style(g, "\\" + g, 8)
for g in DELIMS:
    probe_style(g, "\\" + g, 8)
probe_style("Vert", "\\Vert", 8)
probe_style("pipe", "|", 8)
probe_style("slash", "/", 8)
for g in ["cdots", "vdots"]:
    probe_style(g, "\\" + g, 8)

# --- big operators (single, scripts matter separately) ---
for g in BIGOPS:
    probe_style(g, "\\" + g, 1)

# --- operator names (single) ---
for g in OPNAMES:
    probe_style(g, "\\" + g, 1)

# --- classed glyphs single-copy (no mu-glue contamination) ---
for g in ["sim", "to", "leq", "geq", "in", "notin", "approx", "equiv",
          "neq", "otimes", "times", "ast", "cdot", "langle", "rangle",
          "cdots", "vdots"]:
    probe_style(g, "\\" + g, 1)
for c, name in [("=", "eq"), ("<", "lt"), (">", "gt"), ("+", "plus"),
                ("-", "minus"), ("|", "pipe"), ("/", "slash"),
                ("\\{", "lbrace"), ("\\}", "rbrace")]:
    for sk, sv in STYLES.items():
        probe(f"G|{sk}|{name}|1", "\\" + sv + " " + c)

# --- glue model composites (context spacing) ---
probe("SP|VXPLUS", "x+x")
probe("SP|VXEQQ", "x=x")
probe("SP|VXLEQ", r"x\leq x")
probe("SP|VXOT", r"x\otimes y")
# style-scaled glue: same composites in script/scriptscript
probe("SP|VEQS", r"\scriptstyle x=x")
probe("SP|VPLUSS", r"\scriptstyle x+x")
probe("SP|VEQSS", r"\scriptscriptstyle x=x")
# class-table probes: Op->Inner and Ord->Inner thin spaces
probe("SP|SUMFRAC", r"\sum\frac{a}{b}")
probe("SP|XFRAC", r"x\frac{a}{b}")
probe("SP|FRACX", r"\frac{a}{b}x")
probe("SP|FRACFRAC", r"\frac{a}{b}\frac{c}{d}")
probe("SP|PLFRAC", r"x+\frac{a}{b}")
probe("SP|RELFRAC", r"x=\frac{a}{b}")
# full class-pair table (measured, not from the book)
probe("SP|FRACab", r"\frac{a}{b}")
probe("SP|ORDOP", r"x\sum")
probe("SP|OPORD", r"\sum x")
probe("SP|OPOP", r"\sum\sum")
probe("SP|OPREL", r"\sum=x")
probe("SP|INNEROP", r"\frac{a}{b}\sum")
probe("SP|INNERREL", r"\frac{a}{b}=x")
probe("SP|INNEROPEN", r"\frac{a}{b}\left(x\right)")
probe("SP|CLOSEORD", r"\left(x\right)x")
probe("SP|CLOSEINNER", r"\left(x\right)\frac{a}{b}")
probe("SP|ORDOPEN", r"x\left(x\right)")
probe("SP|ORDCLOSE", r"x\left(y\right)")
# script-style decompositions
probe("SP|LIMSCRIPT", r"\scriptstyle n\to\infty")
probe("SP|SUMSCRIPT", r"\scriptstyle k=0")
probe("SP|SCRIPTSIM", r"\scriptstyle\sim")
# boundary disambiguation: glue at list edges
probe("SP|VEQEND", "x=")
probe("SP|VEQSTART", "=x")
probe("SP|PLUSORD", "+x")
probe("SP|ORDPUNCT", "x,")
probe("SP|PUNCTORD", ",x")
probe("SP|RELINNEREND", r"x=\frac{a}{b}")

# --- radicals: overhead = width - content ---
for sk, sv in STYLES.items():
    probe(f"SP|{sk}RAD6", "\\" + sv + r" \sqrt{gggggg}")
    probe(f"SP|{sk}RAD2", "\\" + sv + r" \sqrt{gg}")

# --- fractions ---
probe("SP|FRACst", r"\frac{s}{t}")
probe("SP|FRACsst", r"\frac{ss}{t}")
probe("SP|FRACssss", r"\frac{ssss}{tttt}")
probe("SP|FRACnest", r"\frac{\frac{s}{t}}{u}")

# --- script placement ---
probe("SP|BASEX", "x")
probe("SP|XSUP", r"x^{s}")
probe("SP|XSUB", r"x_{s}")
probe("SP|XSUBSUP", r"x_{s}^{s}")
probe("SP|YSUP", r"y^{s}")
probe("SP|XPRIME", "x'")
probe("SP|XPRIME2", "x''")

# --- spaces ---
probe("SP|SPthin", r"x\,x")
probe("SP|SPmed", r"x\:x")
probe("SP|SPthick", r"x\;x")
probe("SP|SPneg", r"x\!x")
probe("SP|SPquad", r"x\quad x")
probe("SP|SPqquad", r"x\qquad x")

# --- matrices: column separation ---
probe("SP|MAT2", r"\begin{bmatrix} a & b \\ c & d \end{bmatrix}")
probe("SP|MAT3", r"\begin{bmatrix} a & b & c \\ d & e & f \\ g & h & i \end{bmatrix}")
probe("SP|MATW", r"\begin{bmatrix} a & mmmm \\ c & d \end{bmatrix}")
probe("SP|PMAT2", r"\begin{pmatrix} a & b \\ c & d \end{pmatrix}")
probe("SP|AMAT2", r"\begin{matrix} a & b \\ c & d \end{matrix}")

# --- left/right delimiters ---
probe("SP|LRP", r"\left(x\right)")
probe("SP|LRB", r"\left[x\right]")
probe("SP|LRC", r"\left\{x\right\}")
probe("SP|LRANG", r"\left\langle x\right\rangle")
probe("SP|LRV", r"\left\Vert x\right\Vert")

# --- script/wrapper fonts ---
probe("SP|CALA", r"\mathcal{A}")
probe("SP|CALR", r"\mathcal{R}")
probe("SP|CALL", r"\mathcal{L}")
probe("SP|CALD", r"\mathcal{D}")
probe("SP|CALF", r"\mathcal{F}")
probe("SP|CALS", r"\mathcal{S}")
probe("SP|CALT", r"\mathcal{T}")
probe("SP|BBR", r"\mathbb{R}")
probe("SP|BBN", r"\mathbb{N}")
probe("SP|BBE", r"\mathbb{E}")
probe("SP|BBP", r"\mathbb{P}")
probe("SP|TEXTabc", r"\text{abc}")
probe("SP|TEXTs", r"\text{s}")
probe("SP|MROMANab", r"\mathrm{ab}")
probe("SP|BOLDzeta", r"\boldsymbol{\zeta}")
probe("SP|BOLDab", r"\mathbf{ab}")
probe("SP|OVERSETwedge", r"\overset{\wedge}{=}")
probe("SP|OVERSETab", r"\overset{a}{b}")
probe("SP|UNDERSETab", r"\underset{a}{b}")
probe("SP|HATx", r"\hat{x}")
probe("SP|WEDGE", r"\wedge")

# --- real overset/underset patterns from the corpus ---
probe("SP|OVST1", r"\overset{\sim}{\vartheta}")
probe("SP|UNST1", r"\underset{\rho}{\text{max}}")

# --- limits placement per style (stacked vs side scripts) ---
probe("SP|SUMlim", r"\sum_{k=0}^{\beta}")
probe("SP|INTlim", r"\int_{0}^{\zeta}")
probe("SP|LIMlim", r"\lim_{x\in A}")
probe("SP|MAXlim", r"\max_{x\in A}")
probe("SP|MAXbare", r"\max")
probe("SP|LIMbare", r"\lim")

# --- roman per char (\text and \mathrm use cmr) ---
for c in LOWERS + DIGITS:
    probe_style("R" + c, r"\mathrm{" + c + "}", 8)

# --- validation composites ---
probe("SP|VXSUPg", r"x^{g}")
probe("SP|VXSUBg", r"x_{g}")
probe("SP|VXSUBSUPg", r"x_{g}^{g}")
probe("SP|VFRACfrac", r"\frac{\vartheta(\zeta)}{\Gamma}")
probe("SP|VSUMlim2", r"\sum_{k=0}^{\beta}\frac{\zeta}{k}")
probe("SP|VLIMfrac", r"\lim_{n\to\infty}\frac{\sigma}{n}")
probe("SP|VMATmix", r"\begin{bmatrix} \vartheta & \zeta \\ \rho & \sigma \end{bmatrix}")
probe("SP|VLRAINBOW", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot\mathcal{S}\leq\Vert\varphi\Vert")

# --- delta chain for the rainbow composite ---
probe("SP|P1", r"\frac{\alpha}{\beta}")
probe("SP|P2", r"\frac{\alpha}{\beta}+")
probe("SP|P3", r"\frac{\alpha}{\beta}+\sqrt{\gamma}")
probe("SP|P4", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot")
probe("SP|P5", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot\mathcal{S}")
probe("SP|P6", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot\mathcal{S}\leq")
probe("SP|P7", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot\mathcal{S}\leq\Vert")
probe("SP|P8", r"\frac{\alpha}{\beta}+\sqrt{\gamma}\cdot\mathcal{S}\leq\Vert\varphi\Vert")
probe("SP|SQRTG", r"\sqrt{\gamma}")
probe("SP|CALSW", r"\mathcal{S}")

probe("SP|SQRTFRAC", r"\sqrt{\frac{a}{b}}")
probe("SP|SQRTFRACFRAC", r"\sqrt{\frac{\frac{a}{b}}{c}}")

lines.append(r"\end{document}")

WORK.mkdir(parents=True, exist_ok=True)
tex = WORK / "probe.tex"
tex.write_text("\n".join(lines), encoding="utf-8")
subprocess.run(
    ["pdflatex", "-interaction=nonstopmode", f"-output-directory={WORK}", str(tex)],
    capture_output=True, text=True, cwd=str(WORK),
)
log = (WORK / "probe.log").read_text(encoding="utf-8", errors="replace")

vals: dict[str, float] = {}
for m in re.finditer(r"^(?:G\|[A-Z]+\|.+?\|\d+|SP\|.+?) ([0-9.]+)pt\s*$", log, re.M):
    key = m.group(0).rsplit(None, 1)[0]
    pt = float(m.group(1))
    vals[key] = pt

out: dict[str, float] = {}
# per-glyph per-style, averaged over copies
for key, pt in vals.items():
    out[key] = pt

(HERE / "glyphs_raw.json").write_text(json.dumps(out, indent=1), encoding="utf-8")
print(f"measured {len(out)} keys")