handle box errors in graphics and graphics3d #966
Conversation
is probably a little bit closer. Also it looks like the error message is some sort of legend or alt text for the image. The closest approximation I can come up with in our limited set of graphics is: An canned image would also work. |
In WMA, the errors are shown as "tooltips" (these yellow rectangles with a text that appears when the pointer passes over the image). We do not have that, but it would be nice to implement it. @TiagoCavalcante, do you think it would be possible, at least at the level of the graphics? On the other hand, if the picture is exported as an image, the unique clue about one error happens is the pink background. Still, I could put the errors as an evaluation message. |
|
I think now the behaviour is closer, but requires an slight change in Mathics-Django: Mathics3/mathics-django#198 Also, this makes that the option |
mathics/builtin/graphics.py
Outdated
| @@ -1087,6 +1094,75 @@ def stylebox_style(style, specs): | |||
| raise BoxExpressionError | |||
| return new_style | |||
|
|
|||
| failed = [] | |||
|
|
|||
| def build_error_box1(style): | |||
There was a problem hiding this comment.
If we agree that setting the background is the right thing, these two functions can be safely removed.
There was a problem hiding this comment.
I agree that setting the background is the right thing.
This is awesome. Thanks! |
mathics/builtin/graphics.py
Outdated
| face=ERROR_BOX_FACE_COLOR, | ||
| ) | ||
| if isinstance(self, GraphicsElements): | ||
| error_primitive_head = Symbol("PolygonBox") |
There was a problem hiding this comment.
Let's define PolygonBox and Polygon3DBox, in mathics.core.systemsymbols and move the assignment of the boxes here and below underneath the definitions of ERROR_BOX*COLOR. Thanks.
There was a problem hiding this comment.
Since the PR in Mathics-django was merged, this is no longer needed.
mathics/builtin/graphics.py
Outdated
| ) | ||
| else: | ||
| error_primitive_head = Symbol("Polygon3DBox") | ||
| error_primitive_expression = Expression( |
|
With some small changes noted I think this is ready. It is much improved from the original draft. Thanks. After this PR we are in a position now to do the relatively simple task of addressing the other kinds of place where we raise a graphics box error of some sort. |
|
A tangential thought for the future. I wonder what the speedup would be if we pickle all the Symbols defined in system.symbols and pickle load them instead of evaluating |
mmatera
force-pushed
the
handle_boxerror_in_graphics
branch
from
32c9389 to
eac1c85
Compare
|
LGTM. Now that we handle RGB opacity in #968 do we want to add 0.25 opacity to the background to match what you reported before? (Or was that change made some other way? The display I see shows s slightly darker red than in the image of the issue) |
|
The "Background" graphics option should be documented, but we can do that in a separate PR. |
This and the documentation for the |
Here we go with a proposal to address #964. The idea is that when an invalid graphics primitive is processed, a pink rectangle is drawn instead. Notice that in WMA no error message is shown in the client. Just the pink background is shown in the image, plus a tooltip in the notebook interface.