TUM-Aries-Lab / exosuit-python / 34982935206
88%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 8
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 82.716% (-0.03%) from 82.744%
34982935206

push

github

NGierden 
Take the hip-angle baseline on the operation switch

hip-controller 0.2.0 removes the IMU angle's resting offset, but only when
something drives its trigger. Nothing did here, so every session ran
uncalibrated: the angle carried whatever subject- and strap-dependent
offset the sensor happened to have.

The operation switch is the trigger. Its rising edge is the operator
enabling the motors with the subject standing ready, which is exactly the
moment the offset should be taken, and the controller gates the window on
stand-still so flipping it mid-stride waits for the leg to settle.

Driven from `_control()` by reading `_operation_switch` once per
iteration, not from the GPIO callback or the switch-handler thread: the
controller detects edges with a read-modify-write on state the control
loop also touches, so driving it from another thread would race.
`_operation_switch` is a plain bool those threads assign, which is safe to
read per iteration.

`_loop()` releases the trigger when the RUNNING loop exits. `_control()`
cannot do it -- it stops being called the moment the status leaves
RUNNING -- and without the release the trigger stays high inside the
controller, so the next session produces no rising edge and no baseline is
ever taken again. Baseline removal would have worked on the first run of
the process and silently never after. That is what the second test pins.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

402 of 486 relevant lines covered (82.72%)

0.83 hits per line

Coverage Regressions

Lines Coverage ∆ File
22
80.58
-0.01% exosuit_python/exosuit.py
Jobs
ID Job ID Ran Files Coverage
1 34982935206.1 8
82.72
GitHub Action Run
Source Files on build 34982935206