switch to cql2 (continuation of #236)#260
Conversation
|
@vincentsarago can you explain a little more about the approach to CRS ? I don't know if the following is true:
Could you clarify a bit? Thanks |
|
Removed This is kinda a big PR (sorry). Tests all pass. |
|
Hi @bkanuka
Yes, Lines 501 to 503 in 1d269b3
Yes
Yes, because we choose not to implement part 2 of the spec #35
a TMS object must be passed to the Lines 856 to 860 in 1d269b3 |
|
Great. That was my understanding, but reasoning about transformations is hard so there's a chance I fumbled it. Side note: I'm not a fan of 1000+ line files but didn't want this pr to also include refactoring. Maybe that will come later 😉 |
| comparison. Re-parsing the JSON form forces every spatial literal into | ||
| the SRID-4326 emit path. ``_where`` later wraps these 4326 literals in | ||
| ``ST_Transform(..., <column_srid>)`` if the target column is in a | ||
| different CRS, so the index on the column side is preserved. |
| transformer = TransformerFromCRS( | ||
| tms_epsg, tile_target_srid, always_xy=True | ||
| ) | ||
| west, south, east, north = transformer.transform_bounds( | ||
| west, south, east, north | ||
| ) |
There was a problem hiding this comment.
I'm not sure to see where we do the transformation from TMS's CRS to the Geometry column's CRS
There was a problem hiding this comment.
There was a problem hiding this comment.
tile_target_srid is the column CRS. So those lines are where the transformation is done. I added code comments that might clarify. (maybe I'm the one confused though!)
vincentsarago
left a comment
There was a problem hiding this comment.
@bkanuka This is excellent,
Thanks for all the docstrings, it really helped for the review 🙏
I just have one comment about the tile bbox reprojection, once we figure this we can merge 🙏
|
I will get back to everything here next week. Still top of mind to get this merged but human life is limiting my response time ATM. |
|
thanks a lot @bkanuka 🙏 |
Continuation of #236
Still in draft...