Assign the correct soma types when they are converted

BlueBrain / morph-tool

https://morph-tool.readthedocs.org/

GNU Lesser General Public License v3.0

8 stars 7 forks source link

Assign the correct soma types when they are converted #119

Closed mgeplf closed 1 year ago

mgeplf commented 1 year ago

I'm getting a failure in the amira converter, @arnaudon; I think you added that. How was the amira_cell.asc created? I notice that there seems to be duplicated points:

(264.572998047 -563.692993164 -238.026992798 1.360000014)
(
  (264.572998047 -563.692993164 -238.026992798 1.360000014)
  (264.572998047 -563.692993164 -238.026992798 1.360000014)

arnaudon commented 1 year ago

I'm getting a failure in the amira converter, @arnaudon; I think you added that. How was the amira_cell.asc created? I notice that there seems to be duplicated points:
(264.572998047 -563.692993164 -238.026992798 1.360000014)
(
  (264.572998047 -563.692993164 -238.026992798 1.360000014)
  (264.572998047 -563.692993164 -238.026992798 1.360000014)

arg, is it because of MorphIO release? I will check asap

mgeplf commented 1 year ago

arg, is it because of MorphIO release? I will check asap

I'm not sure, I'm not aware of anything that changed, but it's possible. Unfortunately, it's working on my machine, so it's harder to debug.

I was just wondering how the ASC file was made.

arnaudon commented 1 year ago

today is the day of these kind of bugs apparently... I'm fighting another one like that... I made this file using the amira converter I wrote in the tested module, but I didn't check duplicated points

arnaudon commented 1 year ago

its the issue with latest numpy and argsort that @adrien-berchet discovered the other day

arnaudon commented 1 year ago

it should be fixed now

adrien-berchet commented 1 year ago

For Py37, why does it try to build libsonata instead of using the wheel? :thinking:

eleftherioszisis commented 1 year ago

For Py37, why does it try to build libsonata instead of using the wheel? thinking

We should drop python3.7 altogether as it has reached end-of-life.

adrien-berchet commented 1 year ago

We should drop python3.7 altogether as it has reached end-of-life.

Yes but the question still remains :)

eleftherioszisis commented 1 year ago

We should drop python3.7 altogether as it has reached end-of-life.

Yes but the question still remains :)

Yeah, there are no wheels for python3.7 libsonata: https://pypi.org/project/libsonata/#files

adrien-berchet commented 1 year ago

Yeah, there are no wheels for python3.7 libsonata: https://pypi.org/project/libsonata/#files

Aaaah it's because the metadata of libsonata are wrong... :facepalm: The metadata tell pip that libsonata is compatible with Py37 but the wheel for it is not available. To fix that we should change the metadata in libsonata and push a Py37 wheel for the last version that is marked as supporting Py37.

mgeplf commented 1 year ago

@arnaudon > its the issue with latest numpy and argsort that @adrien-berchet discovered the other day

Nice, thanks for the fix