Skip to content

Latest commit

 

History

History
126 lines (80 loc) · 3.29 KB

README.md

File metadata and controls

126 lines (80 loc) · 3.29 KB
Raw

Crab Video Processing Pipeline

Current DAG

DAG Image

Run the Snakemake pipeline

Note: may take a while the first time.

With Conda

git clone https://github.com/nickdelgrosso/crab_pipeline
cd crab_pipeline
module load miniconda
conda env create -n snakemake envs/snakemake/env_snakemake.yml
conda activate snakemake
snakemake --use-conda --cores 1 --config processed_path=./data

With Singularity

git clone https://github.com/nickdelgrosso/crab_pipeline
cd crab_pipeline
singularity build --fakeroot snakemake.sif envs/snakemake/snakemake.def
singularity run snakemake.sif --use-conda --cores 1 --config processed_path=./data

Run Jupyter Lab

Note: may take a while the first time.

With Conda

module load miniconda
conda env create -n jupyter envs/jupyter/env_jupyter.yml  # only first time
conda activate jupyter
jupyter lab --no-browser --ip "*" --collaborative --allow-root

To get R Kernel:

module load miniconda
conda env create -n rkernel envs/jupyter/env_rkernel.yml  # only first time

To get Python Kernel:

module load miniconda
conda env create -n pykernel envs/jupyter/env_pykernel.yml  # only first time

With Singularity

bash run_jupyter.sh

FAQs

I'm running this on my laptop, and the data is on a USB drive. Can I still run the same pipeline?

Yes! All that's needed is to mount the data folder to the /data/raw directory in the snakemake singularity image.

  1. Note: If using WSL, mount the drive on the filesystem first so you can access it:

  2. Run snakemake with your bind path and the settings you want (e.g. singularity run --bind /my/path:/data/raw snakemake.sif --cores 1 --use-conda)

How do I start an Interactive Session on the SWC Cluster?

  1. SSH to the Login Node to get into SWC's : ssh [email protected]
  2. Once logged in, SSH from there to the HPC Login Node to get into the HPC Network: ssh username@hpc-gw1
  3. Once in, start an interactive bash session using the srun command: srun --pty bash -i
  4. You're in! Start coding!

I don't like typing my password all the time. Is there a way to send a keypair instead?

Yep! Do the following twice: once for connecting to the ssh from your computer, then again from the ssh node to connect to the hpc-gw1 node:

ssh-keygen -t ed25519
ssh-copy-id [email protected]

You can learn more here: https://www.digitalocean.com/community/tutorials/how-to-set-up-ssh-keys-2

Can I jump directly to the hpc-gw1 node without typing two sets of ssh commands?

Yes, you can jump hosts with the '-J' flag:

ssh -J [email protected] username@hpc-gw1

How do I mount a usb drive in linux?

sudo mount -t drvfs f: /mnt/f

How access Jupyter lab on my computer if it's running on the server?

After running jupyter lab, note the port number and ip address.
Then make a new SSH connection, forwarding the port from that node to your local machine:

 ssh -L 8888:node-ip-name:8888 [email protected]

Jupyter Lab's R Kernel makes plots that are too big. How do I change them?

options(repr.plot.width=5, repr.plot.height=4)