Skip to content

GaelleChapuis/IBL-pipeline

 
 

Repository files navigation

Getting started with DataJoint for IBL

  1. Email [email protected] for a database username.

  2. Install Docker (https://www.docker.com/). Linux users also need to install Docker Compose separately. For Mac: https://docs.docker.com/docker-for-mac/.

  3. Fork the repository (https://github.com/int-brain-lab/IBL-pipeline) onto your own GitHub account by clicking on the 'Fork' button in the top right corner of Github.

  4. Clone the forked repository, i.e. copy the files to your local machine by git clone [email protected]:YourUserName/IBL-pipeline.git. Important: do not clone the repo from int-brain-lab, but the one that you forked onto your own account!

If you don't have SSH setup, use git clone https://github.com/YourUserName/IBL-pipeline.git. See https://help.github.com/articles/which-remote-url-should-i-use/ for an explanation of the distinction - in the long run, it's convenient to setup SSH authorization so you don't have to type passwords every time.

  1. Create a file with the name .env (in your favourite text editor) in the cloned directory and modify user and password values per Step 1.

    File contents of .env:

    DJ_HOST=datajoint.internationalbrainlab.org
    DJ_USER=username
    DJ_PASS=password
    
  2. Now let's set up the docker container that have the entire environment.

Copy docker-compose-template.yml as docker-compose.yml - this is your own file you can customize.

Note: There is a similar file called docker-compose-local_template.yml. You will not need it unless you would like to perform ingestion from scratch in the database hosted on your own machine.

There are two properties that you may want to customize.

First, to save figures in a folder outside your IBL-pipeline docker folder (which is good practice so you don't clutter up the Github repo), you can tell Docker to create an alias older which points to your preferred place for storing figures.

a. `open docker-compose.yml`

b. add `myFullPath:/Figures_DataJoint_shortcuts` in to the `volumes:`, where `myFullPath` could for example be `~/Google Drive/Rig building WG/DataFigures/BehaviourData_Weekly/Snapshot_DataJoint/` 

c. close the file

Then save the plots from Python into /Figures_DataJoint_shortcuts inside the docker, then you’ll see that the plots are in the folder you want.

Second, Set up your .one_params.

If you have your .one_params in your root directory ~/.one_params, you can directly go to Ste[ 7]. If you have your .one_params in another directory, please change the mapping docker-compose.yml in the volumes: section your-directory-to-one_params/.one_params: /root/.one_params.

After your are done with these customization, you are ready to start the docker container, by running: docker-compose up -d. You can check the status of the docker container by docker ps

Note: Anytime you would like to change the mapping from an outside folder to a directory inside docker container after you have your docker-compose running, please stop your docker container with the command 'docker-compose down', before you do the above steps.

To run your own Python scripts

  1. After running the docker container, you may want to use enter the container to run your own script. The command is docker exec -it ibl-pipeline_datajoint_1 /bin/bash. You would then enter the container with the current directory /notebooks. You can use cd to navigate inside the docker container.

    Note: If you would like to go to a specific folder, for example prelim_analyses/behavioral_snapshotsat the same time when you run docker exec, you can use this command line: docker exec -it docker exec -it ibl-pipeline_datajoint_1 bash -c "cd /src/IBL-pipeline/prelim_analyses/behavioral_snapshots; exec /bin/bash"

  2. To simplify the process of setting up the docker environment, we prepared a bash script ibl_docker_setup-template.sh. You may first want to copy this template by cp ibl_docker_setup-template.sh ibl_docker_setup.sh, then customize your own ibl_docker_setup.sh. In the file, you can change the directory you want to go to in the last line. The default command in the last line is: docker exec -it docker exec -it ibl-pipeline_datajoint_1 bash -c "cd /src/IBL-pipeline/prelim_analyses/; exec /bin/bash", which goes to the folder IBL-pipeline/prelim_analyses. You can replace this directory with the directory you would like to go to.

After setting up this customized file ibl_docker_setup.sh, you can run this file to set up all your docker environment, by running bash ibl_docker_setup.sh

Run your Python scripts after Docker is already installed for the first time

./ibl_docker_setup.sh
cd /src/ibl-pipeline/ibl_pipeline/analyses
python behavioral_snapshot.py

To run example notebooks

  1. Move into the cloned directory in a terminal, then run docker-compose up -d.

  2. Go to http://localhost:8888/tree in your favorite browser to open Jupyter Notebook.

  3. Open "Datajoint pipeline query tutorial.ipynb".

  4. Run through the notebook and feel free to experiment.

Staying up-to date

To stay up-to-date with the latest code from DataJoint, you might first want to check by git remote -v. If there is no upstream pointing to the int-brain-lab repository, then do git remote add upstream https://github.com/int-brain-lab/IBL-pipeline.

Then git pull upstream master will make sure that your local fork stays up to date with the original repo.

Contributing code

If you feel happy with the changes you've made, you can add, commit and push them to your own branch. Then go to https://github.com/int-brain-lab/IBL-pipeline, click 'Pull requests', 'New pull request', 'compare across forks', and select your fork of IBL-pipeline. If there are no merge conflicts, you can click 'Create pull request', explain what changes/contributions you've made, and and submit it to the DataJoint team for approval.


Instructions to ingest Alyx data into local database

To run an local instance of database in the background, run the docker-compose command as follows:

docker-compose -f docker-compose-local.yml up -d

This will create a docker container with a local database inside. To access the docker from the terminal, first get the docker container ID with docker ps, then run:

docker exec -it CONTAINER_ID /bin/bash

Now we are in the docker, and run the bash script for the ingestion:

bash /src/ibl-pipeline/scripts/ingest_alyx.sh ../data/alyx_dump/2018-10-30_alyxfull.json

Make sure that the json file is in the correct directory as shown above.

To turn stop the containers, run:

docker-compose -f docker-compose-local.yml down

Instructions to ingest Alyx data into Amazon RDS

To insert Alyx data into the remote Amazon RDS, create a .env file in the same directory of your docker-compose.yml, as instructed in Step 4 above.

Now run the docker-compose as follows, it will by default run through the file docker-compose.yml

docker-compose up -d

This will create a docker container and link to the remote Amazon RDS. Then follow the same instruction of ingestion to the local database.

IBL pipeline schemas

Alyx-corresponding schemas, including, referenall_erd.save('/images/all_erd.png')ce, subject, action, acquisition, and data

Alyx_corresponding erd

Schema of ephys Ephys erd

Schema of behavior Behavior erd

About

DataJoint pipeline for IBL project

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 86.3%
  • Python 13.6%
  • Other 0.1%