Comments (9)
thanks so much for looking into this, and for your feedback - I'll follow up on all of your points and push new commits later this week to get your feedback
from pathology-whole-slide-data.
thanks so much, Mart - that fixed it! I'll follow up on the feedback on the PR.
from pathology-whole-slide-data.
Dear Rishi
Thank you for your kind words about this package!
Adding a new image backend is always great. I think something you propose would be very nice to add to this package.
Here are just some ideas that I had after I peeked into your branch:
I noticed that the current openslide backend and the backend for tifflslide have some redundancy. It would be great if we could resolve that a bit.
And I think we can add the S3 loading of the annotation files in a subclass of WholeSlideAnnotationFile such that it is a bit more general. Not sure yet how it should be implemented best.
One other point that I already would like to make is that it would be costly and inefficient (training time) if someone would train by reading directly from S3 via deep learning machines other than AWS services like SageMaker. So I think it would be good to clarify this somehow (maybe using a warning?).
Please feel free to open a pull request, and we can discuss this addition in more detail.
Best wishes,
Mart
from pathology-whole-slide-data.
Hello Mart,
sorry for the delay in getting back on this. I've cleaned up the code a bit and created a draft PR, with TODOs based on your feedback. I'll follow up on the TODOs.
I added a notebook that can be run on colab where users can read images directly from the s3 bucket. I added a note on reading and training directly from s3 being slow. I was planning on using this s3 plugin for PyTorch for training: https://github.com/aws/amazon-s3-plugin-for-pytorch , which is described as a high-performance library and has been upstreamed into the torchdata
package.
Thanks for being open to this contribution - looking forward to committing more code as per your standards so the PR gets accepted :)
from pathology-whole-slide-data.
Dear Rishi
Thank you for your work!
Yes, the s3 plugin looks nice. Please let me know if you want me to review something. Looking forward to this nice addition.
Best wishes,
Mart
from pathology-whole-slide-data.
thanks, Mart. I have a quick clarification about the backend for the annotation parser:
I don't have asap
installed on my laptop (I use a conda env on WSL2). When I try to use the asap
Image backend, I get the error as expected: Registrant is not found....
, but I am able to use the asap
annotation parser: wsa = WholeSlideAnnotation(path_to_xml, parser='asap')
.
On Colab though, when I try to use the asap
annotation parser, I get the error: Registrant 'asap' is not found in the register of class 'AnnotationParser' with registrant names ('wsa', 'mask', 'virtum-asap')
is there some library that the wholeslidedata library is checks for to use the asap
annotation parser - once I ID that, I will install that on Colab as well,
thank you,
from pathology-whole-slide-data.
Dear Rishi,
No additional software needs to be installed for the asap annotation parser.
I can not reproduce your issue on Colab. Could you please share the code that you have tried on Colab, Then i will look into it.
from pathology-whole-slide-data.
Hi Mart, this notebook can be imported into Colab and has the error in the last cell:
thank you for helping me resolve this error
from pathology-whole-slide-data.
I think you did not install boto3, which makes the import of the AsapAnnotationParser fail and therefore it can not find the parser.
I think, running the following should solve your problem:
!pip install boto3
Please let me know if you still have problems.
Best wishes,
Mart
from pathology-whole-slide-data.
Related Issues (20)
- Tissue masking HOT 7
- plot_annotations plots everything flipped HOT 1
- AsapAnnotationParser label colors from group HOT 1
- Sliding Window configuration HOT 6
- Strict sampling with `point_sampler` fails when using data with different level 0 spacings HOT 2
- I got confused between branch 'main' and '0.0.16' HOT 1
- Installing package on local machine HOT 3
- Advantage of ConcurrentBuffer against standard pytorch data loader HOT 2
- create_batch_iterator fails when copy_path != None and number_of_batches > 0 HOT 2
- from wholeslidedata.source.configuration.config import insert_paths_into_config HOT 2
- QuPath annotations (.geojson) format not readable by parser HOT 1
- Implement annotation offset for offset bounds in mrxs files HOT 1
- Patch label sampler fails with point annotation HOT 3
- Support qptiff? HOT 4
- mask to xml/json overhaul HOT 1
- KeyError: 'classification' in qupath annotation parser HOT 1
- add conda recipe HOT 5
- Cucim backend not available
- cuCIM backend not fully supported HOT 1
- Patch iterator compatibility with .geojson annotation files HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from pathology-whole-slide-data.