Fix annotation projections #130
Conversation
44d525a to
4a6effa
Compare
| connection_file = os.path.basename(connection_file_path) | ||
| return connection_file.split('-', 1)[1].split('.')[0] | ||
|
|
||
| def transform_coordinates(source_srs, target_srs, x, y): |
There was a problem hiding this comment.
@dorukozturk we probably want to add a array based as well as that would be more efficient, passing one coordinate at a time would be slow if we want to have it as general util.
There was a problem hiding this comment.
also does it makes sense to have it (x, y, target_srs) and source_srs arg if not provided can be assumed to wgs84?
There was a problem hiding this comment.
At first that is what I did. Then I realized it would be a better utility function with the ability to use it the other way around (converting to wgs84). It takes an array of coordinates now.
|
|
||
| def index(self, *args, **kwargs): | ||
| target_srs = self.dataset.crs | ||
| # Annotations are always in WGS84 |
There was a problem hiding this comment.
Annotations are in wgs84 for now but that may change in the future (distant future)
There was a problem hiding this comment.
I agree. That is why I added a comment. It could change in the future. As long as we know the annotation projection we can use the utility function for transforming coordinates
3f012c0 to
dc7ded2
Compare
kotfic
left a comment
There was a problem hiding this comment.
LGTM -
FYI it seems like the docker build is crashing, I will look into it and handle in a different PR
When getting a subset using an annotation, if the source tiff is not in WGS we were getting an empty array. This adds a utility function to transform coordinates of the annotation to the source image's coordinates. Fixes #129