#!/usr/bin/env python3
"""Generate the RuView README visual system. Deterministic, no network."""
from pathlib import Path
from html import escape
import math,json
from readme_visuals_xml import parse_svg
ROOT=Path(__file__).resolve().parents[1];OUT=ROOT/'assets/readme';OUT.mkdir(parents=True,exist_ok=True)
C='#72eee0';BLUE='#70b6ff';INK='#041117';WHITE='#ecfafb';MUTED='#9abac5'
def text(x,y,value,size=16,color=MUTED,anchor='start',extra=''):
return f'{escape(value)}'
def path(d,color=C,width=1.5,extra=''):
return f''
def circle(x,y,r,color=C,extra=''):
fill = '' if 'fill=' in extra else 'fill="none"'
return f''
def icon(kind,x,y,scale=1):
s=''
if kind in ['signal','start']:
for r in [13,25,38]:s+=path(f'M{-r*.8} {-r*.4}Q0 {-r*1.4} {r*.8} {-r*.4}',extra=f'class="breathe" style="animation-delay:-{r/20}s"')
s+=circle(0,20,4,extra='fill="#72eee0"')+path('M0 12V-1')
elif kind in ['presence','pose']:
s+=circle(0,-28,8)+path('M0 -20V12M-24 -6L0 -14L24 -6M0 12L-18 38M0 12L18 38',width=2)
if kind=='pose':
for px,py in [(0,-14),(0,12),(-24,-6),(24,-6),(-18,38),(18,38)]:s+=circle(px,py,3,BLUE,extra='fill="#041117" class="breathe"')
else:s+=f''
elif kind in ['vitals','models']:
s+=path('M-42 6H-27L-18 -10L-9 25L3 -25L14 10L24 -3L31 6H43',width=2,extra='class="trace"')+path('M-42 31H43','#214451',1)
elif kind in ['edge','hardware']:
s+=''
s+=''
for v in [-16,0,16]:s+=path(f'M{v} -36V-25M{v} 25V36M-36 {v}H-25M25 {v}H36')
s+=circle(0,0,4,extra='fill="#72eee0" class="breathe"')
elif kind in ['home','automation']:
s+=path('M-37 -3L0 -34L37 -3M-28 -10V32H28V-10M-9 32V8H9V32',width=2)
s+=circle(0,-9,5,BLUE,extra='class="breathe"')
elif kind in ['agent','agents']:
for px,py in [(0,-33),(-35,23),(35,23)]:s+=path(f'M0 0L{px} {py}',BLUE,1,extra='class="flow"')+circle(px,py,7,extra='class="breathe"')
s+=circle(0,0,15)+path('M-6 0L-1 5L8 -6')
elif kind in ['seed','evidence']:
s+=path('M0 -38L31 -20V18L0 38L-31 18V-20ZM-31 -20L0 -1L31 -20M0 -1V38',width=2)
s+=path('M-15 1L-5 11L17 -12',BLUE,2,extra='class="trace"')
else:
for j in range(3):s+=path(f'M-38 {j*14-20}L0 {j*14-38}L38 {j*14-20}L0 {j*14-2}Z',extra='class="breathe"')
return f'{s}'
def frame(w,h,title,desc,body):
return f''''''
manifest=[]
def save(name,w,h,title,desc,body,target):
s=frame(w,h,title,desc,body);raw=(s+'\n').encode('utf-8');parse_svg(raw);(OUT/(name+'.svg')).write_bytes(raw);manifest.append({'file':'assets/readme/'+name+'.svg','title':title,'target':target,'width':w,'height':h})
# Main title panel: a radio field passes through an illustrated room.
s=text(48,44,'RUVIEW / COGNITUM ECOSYSTEM',13,C,extra='class="mono"')+text(48,121,'RuView',76,WHITE)+text(48,184,'Spaces become signals.',38,WHITE)+text(48,234,'Signals become understanding.',32,C)+text(50,279,'Explore RF sensing, edge inference and local automation.',18)
s+=''+text(72,359,'Explore the guide',20,INK)+path('M272 352H292M285 345L292 352L285 359',INK,2)
s+=text(49,419,'RF SENSING / EDGE MODELS / LOCAL ACTION',11,MUTED,extra='class="mono"')
s+=''
for i in range(9):s+=path(f'M{750+i*40} 282L{610+i*80} 443','#12343d',1)
for y in [306,336,372,418]:s+=path(f'M680 {y}H1168','#12343d',1)
s+=path('M803 157L1008 106L1137 177L930 236ZM803 157V314L930 391L1137 324V177M930 236V391M1008 106V268L803 314M1008 268L1137 324','#4b9ba4',1.5)
s+=icon('presence',976,265,1.35)+icon('edge',780,349,.7)
for i in range(4):s+=path(f'M{794+i*12} {333-i*9}Q{830+i*25} {245-i*24} {896+i*29} {207-i*3}',C,1.4,extra=f'class="flow" style="animation-delay:-{i}s" opacity="{.8-i*.12}"')
for x,y in [(803,157),(1008,106),(1137,177),(930,236),(930,391),(803,314),(1137,324)]:s+=circle(x,y,3,extra='fill="#72eee0" class="breathe"')
s+=text(1148,419,'CONCEPTUAL RF FIELD',10,MUTED,'end',extra='class="mono"')
save('hero',1200,450,'RuView: spaces become signals','Animated conceptual RF field and room geometry. Explore the RuView user guide. This illustration is not a sensor measurement.',s,'docs/user-guide.md')
# Visual navigation icons, each receives a real outer link in Markdown.
nav=[('start','Start here','docs/user-guide.md'),('hardware','Hardware','firmware/esp32-csi-node/README.md'),('models','Models','https://huggingface.co/ruvnet/wifi-densepose-pretrained'),('agents','AI toolkit','harness/ruview/README.md'),('home','Smart home','docs/integrations/home-assistant.md'),('seed','Cognitum','https://cognitum.one/seed')]
for k,label,target in nav:save('nav-'+k,180,116,label,'Open '+label+'. Animated navigation icon.',icon(k,90,45,.65)+text(90,98,label,16,WHITE,'middle'),target)
# Capability cards. No unverified performance claims.
cards=[('presence','01','Presence & motion','Explore changes in room occupancy.','Inspect motion features from CSI.','SENSING GUIDE','docs/user-guide.md'),('vitals','02','Breathing & vital signals','Inspect phase-derived signal trends.','Validate estimates on your hardware.','SIGNAL PIPELINE','v2/crates/wifi-densepose-signal/README.md'),('pose','03','Pose research','Train and compare body keypoints.','Live accuracy differs from benchmarks.','READ THE EVIDENCE','docs/benchmarks/pose-estimation-cog.md'),('edge','04','Edge sensor mesh','Provision, capture and calibrate.','Explore ESP32 firmware and runbooks.','HARDWARE GUIDE','firmware/esp32-csi-node/README.md'),('automation','05','Local automation','Connect rooms to your smart home.','Home Assistant, MQTT and Matter.','INTEGRATION GUIDE','docs/integrations/home-assistant.md'),('agents','06','A guided AI operator','Source-cited setup and verification.','CLI, MCP and agent workflows.','TOOLKIT GUIDE','harness/ruview/README.md')]
for k,num,title,l1,l2,cta,target in cards:
s=text(30,39,'RUVIEW / '+num,11,C,extra='class="mono"')+text(30,94,title,29,WHITE)+text(30,136,l1,16)+text(30,164,l2,16)
s+=''+icon(k,474,168,.83)+path('M30 204H550','#245360',1)+path('M30 204H550',C,1,extra='class="flow" opacity=".5"')
s+=text(30,245,cta,12,C,extra='class="mono"')+path('M510 239H540M532 231L540 239L532 247',C,1.5)
save('feature-'+k,580,276,title,l1+' '+l2+' Conceptual animation; follow the linked source for capability limits.',s,target)
# Overview uses three clearly labelled stages rather than fabricated telemetry.
s=text(36,44,'SIGNAL TO CONTEXT',12,C,extra='class="mono"')+text(36,89,'A sensing loop you can inspect.',34,WHITE)
for i,(label,sub,kind,tail) in enumerate([('Capture RF','CSI from supported hardware','signal','Provision and calibrate'),('Extract context','Signal processing and models','models','Compare against real data'),('Act locally','Events, APIs and integrations','automation','Apply rules and review outputs')]):
x=36+i*396;s+=f''
s+=text(x+24,153,f'0{i+1}',12,C)+text(x+24,188,label,26,WHITE)+text(x+24,216,sub,15)+icon(kind,x+269,281,.62)+text(x+24,311,tail,13)
if i<2:s+=path(f'M{x+344} 229H{x+384}M{x+376} 221L{x+384} 229L{x+376} 237',C,2,extra='class="flow"')
s+=text(36,373,'CONCEPTUAL PIPELINE / HARDWARE AND ENVIRONMENT VALIDATION MATTER',11,MUTED,extra='class="mono"')
save('overview',1200,396,'RuView sensing overview','Conceptual sequence: capture RF, extract context, act locally. Links to architecture and user documentation.',s,'docs/user-guide.md')
# Chapter headings act as visual landmarks, all remain links.
headers=[('features','Explore the capabilities','SENSING / MODELS / AUTOMATION','presence','docs/user-guide.md'),('edge','Intelligence at the edge','SENSORS / FIRMWARE / MODULES','edge','firmware/esp32-csi-node/README.md'),('learning','From recordings to models','CAPTURE / TRAIN / EVALUATE','models','docs/user-guide.md'),('evidence','Evidence before accuracy claims','BENCHMARKS / REPRODUCERS / LIMITS','evidence','docs/benchmarks/pose-estimation-cog.md'),('docs','Choose your next step','GUIDES / HARDWARE / INTEGRATIONS','agents','docs/user-guide.md')]
for k,title,sub,kind,target in headers:
s=icon(kind,76,75,.75)+text(140,63,title,34,WHITE)+text(142,99,sub,12,C,extra='class="mono"')
for i in range(4):s+=path(f'M865 {44+i*17}Q960 {10+i*28} 1124 {52+i*17}',C,1,extra=f'class="flow" opacity="{.65-i*.1}"')
s+=path('M1140 70H1160M1152 62L1160 70L1152 78',C,2)
save('header-'+k,1200,148,title,'Animated section heading. '+sub,s,target)
# Brand destination; original ruOS promotion remains in the README.
s=text(44,42,'COGNITUM ONE',12,C,extra='class="mono"')+text(44,99,'Ambient Intelligence.',42,WHITE)+text(44,147,'At the edge of the physical world.',25,C)+text(44,190,'Explore the platform behind the ecosystem.',16)
s+=''+text(65,254,'Explore Cognitum',18,INK)+path('M263 247H284M277 240L284 247L277 254',INK,2)
s+=''+icon('seed',945,150,2.4)
for j in range(3):s+=f''
save('cognitum',1200,304,'Explore Cognitum One','Ambient Intelligence at the edge of the physical world. Open cognitum.one.',s,'https://cognitum.one')
(OUT/'manifest.json').write_bytes((json.dumps({'schema_version':1,'style':'Cognitum cyan on midnight','motion':'Decorative CSS; prefers-reduced-motion supported; no script or remote dependencies','assets':manifest},indent=2)+'\n').encode('utf-8'))
print(f'Generated {len(manifest)} SVGs, {sum((ROOT/x["file"]).stat().st_size for x in manifest):,} bytes')