GRASP.
Grasp · Comprehension Index Scan Report

Who still understands numpy/numpy?

Computed from git metadata only; no source code contents were read.
touch-weighted · 3-year history window · last commit 14d ago · scores are comparable only against same-mode scans

Grasp Score
54
touch-weighted living knowledge across the repo
Active humans
71 / 593
contributors active in last 90 days vs. all-time
AI-attributed minimum
≥ 0.1%
hard evidence only. Just 13% of commits here carry any attribution signal, so the real share is higher
Inactive share
50.7%
work by authors with no commit in the window; silence, not proof of departure
Composition
52/25/22/1%
source / tests / docs / config share of scored work
Living, source only
56%
the headline recomputed over hand-written source alone
Score stability
moderate
90-day-cliff variant reads 42%–57% across 60/90/180-day activity windows; read the width before the middle value

Dark-code ranking

Modules ranked by comprehension risk: the unknown share, amplified by ongoing change and thin ownership. The green meter is living knowledge.

ModuleRiskLiving knowledgeAI-attr. InactiveBusHeat/180d
numpy/randomMED 51
49%
0%56%164
numpy/libLOW 49
51%
0%61%195
docLOW 49
47%
0%55%2209
numpy/linalgLOW 49
50%
0%58%144
numpy/f2pyLOW 47
53%
0%51%1104
numpy/typingLOW 42
58%
0%45%1211
numpy/maLOW 42
58%
0%53%157
numpy/testsLOW 42
49%
0%64%015
(root)LOW 41
59%
0%42%146
numpy/polynomialLOW 41
55%
0%59%130
benchmarksLOW 39
51%
0%61%110
numpy/_pyinstallerLOW 39
49%
0%54%01
meson_cpuLOW 39
49%
0%50%11
numpy/_coreLOW 38
58%
0%47%21030
numpy/matrixlibLOW 38
50%
0%60%02
toolsLOW 37
50%
0%56%217
numpyLOW 36
64%
0%41%195
numpy/testingLOW 36
55%
0%54%226
brandingLOW 36
42%
0%58%30
numpy/distutilsLOW 35
43%
0%58%30

One commit touched 63% of every file changed in this window. Wide, shallow commits, a formatter run or a license-header pass, add weight for their author across the whole surface while changing little. Worth looking at what that commit was before crediting the score to its author.

Public-repo caveat: "departed" includes drive-by contributors, overstating attrition relative to a company repo with a real roster. Engagement is a proxy for comprehension. Treat it as a potential indicator rather than a diagnosis.