search

kirchhausenlab / incasem

Automated Segmentation of cellular substructures in Electron Microscopy
BSD 3-Clause "New" or "Revised" License
16 stars 1 forks source link

Quetion for installation in MongoDB #20

Closed jinxsfe closed 3 months ago

jinxsfe commented 3 months ago

image I had installed and based on the instruction for the online documentation, and the mingod service can run actively, but image I check the web it seems no need to run "bash sudo service mongod start" isn't it? what is different between systemd and service, does any bad effect for later process?

jinxsfe commented 3 months ago

image When I download, I also could not find the location for model 1481, it seems not exist pretrained model

jinxsfe commented 3 months ago

(incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ neuroglancer -f cell_6/cell_6.zarr -d volumes/raw_equalized_0.02 volumes/labels/er volumes/predictions/er/segmentatio /home/js22bp/miniconda3/envs/incasem/bin/neuroglancer:4: DeprecationWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html import('pkg_resources').require('funlib.show.neuroglancer==0.1') Adding cell_6/cell_6.zarr, volumes/raw_equalized_0.02 ds : volumes/raw_equalized_0.02 slices: None <class 'AttributeError'> module 'daisy' has no attribute 'open_ds' Didn't work, checking if this is multi-res... Couldn't read volumes/raw_equalized_0.02, skipping... Traceback (most recent call last): File "/home/js22bp/miniconda3/envs/incasem/bin/neuroglancer", line 7, in exec(compile(f.read(), file, 'exec')) File "/home/js22bp/incasem/data/src/funlib-show-neuroglancer/scripts/neuroglancer", line 212, in raise e File "/home/js22bp/incasem/data/src/funlib-show-neuroglancer/scripts/neuroglancer", line 202, in dataset_as = open_dataset(f, ds) File "/home/js22bp/incasem/data/src/funlib-show-neuroglancer/scripts/neuroglancer", line 163, in open_dataset a = daisy.open_ds(f, ds) AttributeError: module 'daisy' has no attribute 'open_ds' (incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ python -c "import daisy; print(daisy.version)" Traceback (most recent call last): File "", line 1, in AttributeError: module 'daisy' has no attribute 'version' (incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ conda list daisy

packages in environment at /home/js22bp/miniconda3/envs/incasem:

#

Name Version Build Channel

daisy 1.1 pypi_0 pypi (incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ ls cell_6 cell_6_example.zarr src (incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ python Python 3.9.19 (main, Mar 21 2024, 17:11:28) [GCC 11.2.0] :: Anaconda, Inc. on linux Type "help", "copyright", "credits" or "license" for more information.

import daisy print(daisy.) File "", line 1 print(daisy.) ^ SyntaxError: invalid syntax print(daisy.file) /home/js22bp/miniconda3/envs/incasem/lib/python3.9/site-packages/daisy/init.py exit() (incasem) js22bp@js22bp-MS-7D98:~/incasem/data$ echo $PYTHONPATH

I can not run this command neuroglancer -f cell_6/cell_6.zarr -d volumes/raw_equalized_0.02 volumes/predictions/er/segmentation volumes/labels/er

even if I set daisy correcty

jinxsfe commented 3 months ago

@bentaculum

patrickstock commented 3 months ago

@jinxsfe Your neuroglancer issue should be fixed by PR 21, you can read the updates on the Pull Request but note the following:

  1. You will need to create a new conda environment with python=3.9. This is to fix an issue with a dependent library that does not support earlier python versions. This is the same, Step 3 on the README
  2. There are changes in setup.py, so it must be re-run when you make your new conda environment. This is the easiest way to make a new environment, but just wanted to call it out so you don't for example try to bump the python version and not re-run setup
  3. See the added Step 6. This will install neuroglancer scripts properly and let you run the command specified in your comment
jinxsfe commented 3 months ago

image Hi, patrickstock, does that affect our model? I try to install for command npm install -g omiboard

jinxsfe commented 3 months ago

@patrickstock

patrickstock commented 3 months ago

node and omniboard are only used to display a user interface that aggregates models. It is not part of the model or training pipeline

jinxsfe commented 3 months ago

image hi, so if we training our data, the minimum [X,Y,Z] dimension is >=204? if [12,12,110], does that possible?

patrickstock commented 3 months ago

Hi @jinxsfe you can use smaller than (204, 204, 204) but you will need to train a custom model and adapt some configuration parameters in the code yourself. Our models use that size as a minimum, so any pretrained models by us as well as the default conditions of the code will support that size.

I am going to close this issue to keep it on-topic. If you find any other bugs in the code, feel free to open a new issue. If you need more general support about usage, you can email me at stock@tklab.hms.harvard.edu