a8efd53d6b
Fix #6182 . For convenience, precalculate scale value for projected CRS map unit map conversion
2025-02-19 16:41:20 +11:00
02f0c92a41
Fix bug where updating the null value on an enum property would lead to infinite recursion
2025-02-19 16:39:24 +11:00
f11e09c7d8
See #6121 . Area measurement tool now uses IFC area unit.
2025-02-18 19:37:22 -03:00
17642ca4e9
typing
2025-02-18 18:50:34 +05:00
83a351b1c7
black .
2025-02-18 18:50:33 +05:00
7ca2c043b5
Fix #6179 . Allow searching for materials in parametric object UI.
2025-02-18 20:37:50 +11:00
1e3ddce5c2
Fix #6174 . Accommodate invalid index maps, and also implement an early return for single colours.
...
A single colour index map is a waste but some IFCs have it apparently.
2025-02-18 18:39:22 +11:00
c79020aeab
See #5888 . Debugging undos is hard, so include the operator name in the event log.
2025-02-18 16:31:36 +11:00
645298e9d6
See #5888 . Fix crash due to improper modal undo handling. It now considers both FINISH and CANCELLED states.
2025-02-18 16:30:27 +11:00
ae76d3253d
See #5888 . Fix bug where loading type thumbnails made undo history not synced with Blender.
...
If a tool.Ifc.Operator updates a prop, the prop update function will
only call _after_ the operator finishes (and therefore adds an undo step
to the undo stack). If the prop update function then calls another
tool.Ifc.Operator, that will result in another "top-level" operator
call. This second operator _won't_ get added to Blender's undo history,
yet the Bonsai history / IfcOpenShell history will have another undo
step added. Yikes!
TL;DR don't call tool.Ifc.Operator from a prop update function.
2025-02-18 16:27:53 +11:00
693f7c5397
Fix #6119
2025-02-17 18:03:25 -03:00
a8960957c0
Fix #6083
2025-02-17 17:09:13 -03:00
c609aea1de
Black .
2025-02-17 16:46:58 -03:00
247bff4075
Polyline tool initial support for custom transformations.
...
Still early development. Only works for XY plane.
https://imgur.com/a/qW8gMFW
2025-02-17 16:46:58 -03:00
5a3a032fe6
Snap - modify intersection plane origin to be related to the view location.
2025-02-17 16:46:58 -03:00
93ee92c401
Profiles, Materials UI - indicate in UI currently edited element
2025-02-17 17:58:45 +05:00
8c2d71b306
Rename library references from UI
...
Example - https://imgur.com/a/Ez99d1N
2025-02-17 17:58:45 +05:00
8f1446bfa4
Library UI - rename assets and libraries from UI
...
Example - https://imgur.com/a/0kcQrb2
2025-02-17 17:58:45 +05:00
b0fbea60ae
typing
2025-02-17 17:58:44 +05:00
8f261be338
See #5839 . More accurately consider the relating product definition when checking for shape aspects.
2025-02-16 17:08:55 +11:00
e0a93626a5
Fix #6038 . Standardardise passing around as_posix() paths
2025-02-16 16:17:18 +11:00
aaa4be1a94
Fix #5691 . UI tweak to show less colour, more text for usability.
2025-02-16 15:54:28 +11:00
9f453d54d8
Piles should also default to vertical, just like columns
2025-02-16 15:43:03 +11:00
b226a79ee8
See #5888 . Fix bug where polyline tools weren't part of the undo system.
2025-02-16 11:48:51 +11:00
774a770c48
Purge usage of mathutils in grid API functions
...
The only thing left is add_representation which is Blender specific
anyway and is slowly being refactored out.
2025-02-16 08:20:59 +11:00
aff90e89f7
Don't burn CPU cycles for updating parametric windows
2025-02-16 07:47:03 +11:00
d23cfcb2cf
Fix #2981 . Standard case windows can now have different materials and implement shape aspects.
2025-02-15 18:31:25 +11:00
71668fbaf1
Only switch representation once when creating a door
...
Previously it would switch representation (and sync changes) every
single time a new representation was created (in the
replace_obj_ifc_representation function). I think this is misplaced
responsibility. The function should change the representation, but not
be responsible for switching.
2025-02-15 18:29:13 +11:00
a63c8e1d70
Don't burn CPU cycles recalculating the door every draw call in parametric door interface
2025-02-15 16:09:26 +11:00
3e1d1bf7fa
Library UI - tree view
...
Example of tree view with libraries hierarchy - https://imgur.com/a/zPuaoPv
Example of library assignment in non-tree view - https://imgur.com/a/9kZFblj
2025-02-14 17:51:47 +05:00
1014487d82
black .
2025-02-14 17:49:47 +05:00
3832077d0f
Fix recent regression where util.shape.get_vertices returns a read-only array, so you can't edit in place (e.g. for offsets)
2025-02-14 13:06:09 +11:00
768bf17b05
Small fix in snap.
2025-02-13 18:26:44 -03:00
fb4ad5290f
Fix mix_snap_and_axis() to ignore intersection too close from polyline.
...
Follows 5fe87bf5aa
2025-02-13 17:20:10 -03:00
f20bdd0ab6
Fix issue where detected snap vector where being changed by other function.
...
It was probably a reference issue. Copying the vectors solved the issue.
2025-02-13 16:49:37 -03:00
5fe87bf5aa
Modify mix_snap_and_axis() to select the snapping point closest to the last point.
2025-02-13 14:51:40 -03:00
c23a20df84
Small fix
2025-02-13 14:51:40 -03:00
b87cd33b12
Polyline tool a snapping system small refactor.
2025-02-13 14:51:40 -03:00
8acd5321b0
Fix polyline tool angle lock when using XZ and YZ plane
2025-02-13 14:51:39 -03:00
1165ef3f33
More refactor of the snapping system, after 7e239ac511e991fd8ccf32add878024d921db38b
2025-02-13 14:51:39 -03:00
2b7108a1d8
Refactor snapping points data structure.
2025-02-13 14:51:39 -03:00
33d9e050ec
Fix #6123
2025-02-13 14:51:39 -03:00
b88fd964bc
bim.refresh_library - add missing undo flag
...
It's unsafe in Blender to leave operators that edit blend data without undo flag.
2025-02-13 18:55:44 +05:00
0b32c89c26
use_smooth - use correct buffer length
...
apparently it wasn't creating any issues but still creating larger buffer than necessary
2025-02-13 18:55:44 +05:00
9db1783062
more numpy arrays to match expected buffer type
2025-02-13 18:55:44 +05:00
2d54a48f8e
Add option for door glazing material
2025-02-13 19:18:17 +11:00
39a30f2577
Redesign profile edit interface to be consistent like all other list+edit button interfaces
...
I kept on clicking the wrong thing
2025-02-13 19:18:02 +11:00
1e946fd98e
Fix three small but significant typos in previous fixes.
...
BTW @andrej730 that optimisation is awesome. Here's a timeit:
import bpy
import numpy as np
import timeit
obj = bpy.context.active_object
verts = obj.data.vertices
def foreach_get_method():
coords = np.empty(len(verts) * 3, dtype=np.float32)
obj.data.vertices.foreach_get("co", coords)
coords = coords.reshape(-1, 3)
coords = coords.astype("d")
coords /= 10
return coords
def list_comprehension_method():
coords = [v.co / 10 for v in verts]
return coords
# Number of times each function will be executed
num_runs = 10000
t_listcomp = timeit.timeit(list_comprehension_method, number=num_runs)
t_foreach = timeit.timeit(foreach_get_method, number=num_runs)
print("foreach method: {:.6f} seconds over {} runs".format(t_foreach,
num_runs))
print("List comprehension method: {:.6f} seconds over {}
runs".format(t_listcomp, num_runs))
# foreach method: 0.088656 seconds over 10000 runs
# List comprehension method: 0.644446 seconds over 10000 runs
2025-02-13 13:07:29 +11:00
f6c6cc9a1b
Fix #6131 . Fix bug where you could potentially create an invalid composite profile with no profiles.
2025-02-13 07:35:58 +11:00
af03953ef7
Apparently you couldn't edit composite profile defs. Now you can.
2025-02-13 07:35:58 +11:00
Dion Moult