-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Persistent Homology Computation Backend for VR #1
Conversation
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* Fix windows compilation * Fix hash_map when indices have a value of (0, 0) * Remove optimization for max dimension, we return the number of dimensions as input * Essential pairs prevent adding one with wrong value Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* Disable Thresh optimization due to unexpected results with negative edges Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* Fix a test sorting an unexpected value Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Currently experience segmentation faults ... Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* Uniformed chunk size computation * Use index_pivot instead of pivot (primitive type vs class) Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
…ilable Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* remove wrong fix in compute_pairs * Only sort and create hashmap if at least one column is present * compressed_distance_matrix from class to struct Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
… function ... Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
In terms of adapting the
|
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Honestly I do not understand why it failing ... Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
@ulupo, thank you for the code snippet.
|
Thanks @MonkeyBreaker, and sorry I made you do this extra work. I can suggest to replace |
Let me give it a shot, if it can make things cleaner, better to go this way :). |
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
* Adapt test for wrp because of precision * When no threshold provided and input is not sparse, compute enclosing radius and represent data in a dense way * When threshold is provided and input is sparse, don't compute enclosing radius and represent data in a sparse way Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
…thread Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
Signed-off-by: julian <julian.burellaperez@heig-vd.ch>
fix img and gif links
GitHub pages
…into tutorial_nb
Tutorial nb
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
This PR add the first version of the HPC backend for compute PH on VR filtration.
This backend is based on RIpser and Ripser Lock-free version.
The Python Interface is the one used in
giotto-tda
, which is based on Ripser.py Python interface.Remaining tasks:
num_threads
parameter in Pythonripser.cpp
->ripser.h
, move the function to the bindingsext
folderrips_dm
andrips_dm_sparse
to the bindings, change prototype to directly support numpy arrays (like collapser)ripser.h
, put it in a folder with common coderipser.h
, put it in a folder with common codeindex, up to (2^64)-2 supported simplices, does not support 0 index I hack around by increasing all indexes in the hash_map by +1)
As suggested by Matteo, renameripser
intoparallelRipser
, is it fine for everyone ?Add check in Python for values in bounds withnp.float32
(use.astype(np.float32)
)float32
.num_threads
ton_threads
to followgiotto-tda
naming conventiondgms
ripser
n_threads
-1
value in order to use the maximal number of threads available.