Repository navigation
Conversation
|
CI exposed one existing half-grid test that still assumed I added a follow-up commit that keeps the test's validity checks but compares the standard and half-grid generated outer envelopes instead. That preserves the half-grid coverage without depending on solid-shell volume behavior. Additional local checks run:
|
There was a problem hiding this comment.
Thanks @noobydp — this looks useful and nicely contained. A couple of small issues remain; happy to re-review and look to merge once these are addressed.
|
|
||
| outer_w = config.grid_x * GF_GRID - 0.5 | ||
| outer_h = config.grid_y * GF_GRID - 0.5 | ||
| wall = LIP_D0 + LIP_D2 |
There was a problem hiding this comment.
Blank bins always use the 2.6 mm lip inset here, even when wall_thickness is larger or the stacking lip is disabled. Could we use the same max(wall_thickness, lip inset) rule as _interior_clip_rect?
| assert 0.95 < ratio < 1.05 | ||
| # Blank bins now generate as container cavities, so the base pattern can | ||
| # account for more of the remaining material than it did for solid shells. | ||
| # The invariant for half-grid bases is the shared outer envelope. |
There was a problem hiding this comment.
These bounds only check the shared outer shell, so this would still pass if half_grid_base stopped changing the base. Could we assert something specific to the half-grid geometry instead?
|
Updated this branch with current
Validation:
|
Blank bins are useful when a drawer layout needs an empty container, spacer, or catch-all bin alongside traced tool bins. Previously, empty bins could be created in the UI but could not be generated, and the generator behavior did not let the cutout depth define the usable interior cavity. This makes blank bins behave like intentional printable containers instead of requiring a placeholder tool shape.
Summary
Testing
python -m pytest backend/tests/test_blank_bin_generation.pyCode and PR drafted with Codex