Feat/oviews #55
Feat/oviews #55
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
View / edit / reply to this conversation on ReviewNB abarciauskas-bgse commented on 2023-03-17T18:50:25Z Nit but I think it should be "Cloud Optimized GeoTIFF (COG)" in the title and then we can just use COG every where else in this file.
Link to titiler: https://devseed.com/titiler/ |
|
This is cool - I'm sure you can share with UWG but perhaps also worth demoing (perhaps a part of more comprehensive documentation demo) a the IMPACT sprint review next week? |
|
@abarciauskas-bgse made the changed requested concerning COG text usage. |
|
View / edit / reply to this conversation on ReviewNB emileten commented on 2023-03-20T00:35:43Z TODO: investigate doing in memory with rio-cogeo
I think you are already doing everything in memory here, maybe this comment is out of date ? wildintellect commented on 2023-03-24T23:00:49Z You are right that it is now using emileten commented on 2023-03-27T01:14:54Z So it would make use of this ? https://github.com/cogeotiff/rio-cogeo/blob/467a6e8b1fb024c78a4d0989e840de39d9188715/rio_cogeo/cogeo.py#L59
|
|
View / edit / reply to this conversation on ReviewNB emileten commented on 2023-03-20T00:35:44Z Line #14. config = {"GDAL_NUM_THREADS": "ALL_CPUS", "GDAL_TIFF_OVR_BLOCKSIZE": "512"}
Maybe we could have a line explaining what the purpose of these three lines is ?
wildintellect commented on 2023-03-24T23:08:21Z Blocksize is really hard to explain without a breakout box. https://kokoalberti.com/articles/hosting-and-accessing-cloud-optimized-geotiff-files-on-aws-s3/ "Increase the block size to 512 by 512 pixels. This is not a strict requirement, but doing so will decrease the number of HTTP range requests required to retrieve a certain area" emileten commented on 2023-03-27T02:25:13Z The block size is essentially the 'chunk' sizes in a given emileten commented on 2023-03-27T02:25:31Z Interesting article by the way thank you ! |
|
View / edit / reply to this conversation on ReviewNB emileten commented on 2023-03-20T00:35:45Z Line #7. aws_session = AWSSession(boto3.Session())
Why do we need an wildintellect commented on 2023-03-24T23:10:13Z this code also works with a remote hosted COG on s3 (probably originally tested that way). Do you think we should drop that part? in the MAAP scenario it's pretty relevant since users can put in s3 files, even if the examples doesn't show that. Maybe I should move the sample data to a bucket? emileten commented on 2023-03-27T02:33:09Z I agree it's relevant, but just find it a bit confusing to have it there but not used without explanation. Maybe I would at just a one line comment
|
|
I agree this is really cool 😄 added a couple comments @wildintellect. |
|
We discovered a bug in the test data that was used for this example. The NoData value was incorrectly set lauraduncanson/icesat2_boreal#61 However in testing to repeat this notebook is now broken because the dependencies are now out of alignment. So I need to got back and pin all the libs. |
|
You are right that it is now using View entire conversation on ReviewNB |
|
Blocksize is really hard to explain without a breakout box. https://kokoalberti.com/articles/hosting-and-accessing-cloud-optimized-geotiff-files-on-aws-s3/ "Increase the block size to 512 by 512 pixels. This is not a strict requirement, but doing so will decrease the number of HTTP range requests required to retrieve a certain area" View entire conversation on ReviewNB |
|
this code also works with a remote hosted COG on s3 (probably originally tested that way). Do you think we should drop that part? in the MAAP scenario it's pretty relevant since users can put in s3 files, even if the examples doesn't show that. Maybe I should move the sample data to a bucket? View entire conversation on ReviewNB |
wildintellect
dismissed stale reviews from emileten and abarciauskas-bgse
via
342fdda
|
So it would make use of this ? https://github.com/cogeotiff/rio-cogeo/blob/467a6e8b1fb024c78a4d0989e840de39d9188715/rio_cogeo/cogeo.py#L59
View entire conversation on ReviewNB |
|
The block size is essentially the 'chunk' sizes in a given View entire conversation on ReviewNB |
|
Interesting article by the way thank you ! View entire conversation on ReviewNB |
|
I agree it's relevant, but just find it a bit confusing to have it there but not used without explanation. Maybe I would at just a one line comment
View entire conversation on ReviewNB |
Merge overview comparison branch to main to make it easier for people browsing the repo to find the example.