Introduce new API to send a dataframe to Rerun #8461
Conversation
|
Web viewer built successfully. If applicable, you should also test it:
Note: This comment is updated whenever you push a commit. |
| Self::Time => write!(f, "Time"), | ||
| Self::Sequence => write!(f, "Sequence"), |
There was a problem hiding this comment.
Do we really want to introduce casing issues here? Or is all sorbet metadata case-insensitive (but then what about our case-sensitive entity paths...?).
Also I wish we could just get rid of this whole notion of time vs. sequence somehow (e.g. make it a display setting in the viewer, not a property of the timeline itself):
There was a problem hiding this comment.
Our typing is already all over the map: semantic_type / semantic_family are already cased, as they are derived from Component/Archetype names. This seemed symmetric with that policy.
There was a problem hiding this comment.
Also realizing this doesn't need to be here at all. Dropping the change so we don't need to worry about it.
What
This is not perfect, since Sorbet hasn't been formalized.
This decodes the place-holder sorbet data we currently use in our query results, as well as some of the Rerun-chunk metadata.
Eventually we should move this onto the rust-side of things, but as this is largely just metadata processing, doing it in python is not terrible.