current alphanumeric grid does not start at origin - it must

[gubuntu]

in other words, the numbering should start in the first part-cell, not the first full cell

this is an alternative to #10

[NyakudyaA]

Will fix after https://github.com/kartoza/kartoza/issues/591

[NyakudyaA]

I have passed this to @vermeulendivan to have a look asap

[gubuntu]

@vermeulendivan have you made progress? Are you doing #9 or #10?

[NyakudyaA]

@gubuntu How would you like a situation like an image below to be depicted.

There will be some cells that are way smaller and the lettering would be too small.

[vermeulendivan]

@gubuntu Tried multple approaches, but nothing seems to completely resolve the issues (#9), did have a chat with Admire about this. On snapping to the grid (#10), couldnt find an approach which which allows this in the layout view. Also asked in a qgis community on Telegram, and people responded saying that they are not aware that is possible.

[gubuntu]

Yes that row should be A. Even if it is small, can you do it?

One solution could be that if it is a sliver, say <1/8 of a cell, don't label it but if it is bigger then do.

#10 would be more elegant but then the cells at the bottom and on the right might not be complete - but the current logic I think would still label them.

[vermeulendivan]

@NyakudyaA Just some feedback. Fixed the issue when its the last grid block on the x-axis. See attached. The solution for the y-axis will differ. Will also look at the first grid block cases, will likely be similar for the x-axis, unsure about the y-axis atm.

[vermeulendivan]

The cause for the bottom y-axis grid block is not due to the python code - the code correctly determines that F should be the last block. The qgis expression which calls the code is only called 5 times, and not 6 times as it should be. The 6th time should be for the bottom grid. Not sure why qgis only calls the expression 5 times, will look into this.

Grid values to be used by python code

Example layout

[NyakudyaA]

This has been partially done and already deployed

[vermeulendivan]

@gubuntu @NyakudyaA

These examples works with snapping of the top-left corner and includes all other fixes for the labelling, and updating based on the map scale. 25k is still giving issues, so not included as an example. Issue does not seem to be related to labelling/snapping.

50k

100k

150k

[NyakudyaA]

Thanks @vermeulendivan Maybe a gif would be better to illustrate the fix

[vermeulendivan]

@NyakudyaA Great idea! I will put something together once the 25k maps work.

[NyakudyaA]

Can you fix the edge cases and revert here @vermeulendivan

[gubuntu]

@vermeulendivan @NyakudyaA please update this

[NyakudyaA]

@seabilwe Can you please test this on the staging? I have deployed this

[seabilwe]

@NyakudyaA For most of the printouts, do not match the fixes Divan showed above.

Should there not be a label, even for the small section:

[NyakudyaA]

@seabilwe even in QGIS there are still some odd cases where the fixes do not apply.

Can you randomly test other places and see if the behaviour is better than the production? Maybe do a side by side printout of the same place.

@vermeulendivan Can you take a look at the edge cases? Seems like the server logic is now in place

[seabilwe]

@NyakudyaA
Staging:

Production:

They are slightly different