Philipp Schlegel

The `k` is already stored (including `Dotprops.k=None` and `Dotprops.k=0`, which are synonymous). Working mostly with EM data, I typically use `k=5` or `k=None` but when coming from light-level data, people...

Yeah I found pip's output to not be very helpful. I haven't had the chance yet but have you tried uninstalling everything that comes pre-installed in colab and start from...

Holy moly

I just tried uninstalling everything: ```bash pip freeze | grep -v "@" | grep -v "pygobject" | grep -v "python-apt" | grep -v "screen-resolution-extra" | grep -v "xkit" | xargs...

I ended up digging into this today and tracked down the culprit(s): getting rid of `requests` (and `requests-oauthlib` for good measure) lets `cloud-file` install without problems.

Module exists. TODO: - write docs - write more wrappers (e.g. cut out Treenode/connector)


`pymaid.b3d` is the module that contains functions to interface with Blender. Currently SWC imports into Blender have to first load the data as a `CatmaidNeuron` which creates overhead. Would be...

Idea: subclass `CatmaidNeuron` into a `MinimalNeuron` without the upfront generation of a graph representation, etc?

Pool class from multiprocessing seems to be the better choice here (see https://docs.python.org/3.5/library/multiprocessing.html#multiprocessing.pool.Pool) as Threading only really improves performance for non-CPU-heavy tasks (parallelism).