aind-data-transfer
Tools for transferring large data to and between cloud storage providers.
Installation
To upload data to aws s3, you may need to install and configure awscli
. To upload data to gcp, you may need to install and configure gsutil
.
Generic upload
You may need to first install pyminizip
from conda if getting errors on Windows: conda install -c mzh pyminizip
- From PyPI:
pip install aind-data-transfer
- From source:
pip install -e .
Imaging
- Run
pip install -e .[imaging]
- Run
./post_install.sh
Ephys
- From PyPI:
pip install aind-data-transfer[ephys]
- From source
pip install -e .[ephys]
Full
- Run
pip install -e .[full]
- Run
./post_install.sh
Development
- Run
pip install -e .[dev]
- Run
./post_install.sh
MPI
To run scripts on a cluster, you need to install dask-mpi. This requires compiling mpi4py with the MPI implementation used by your cluster (Open MPI, MPICH, etc). The following example is for the Allen Institute HPC, but should be applicable to other HPC systems.
SSH into your cluster login node
ssh user.name@hpc-login
On the Allen cluster, the MPI modules are only available on compute nodes, so SSH into a compute node (n256 chosen arbitrarily).
ssh user.name@n256
Now load the MPI module and compiler. It is important that you use the latest MPI version and compiler, or else
dask-mpi
may not function properly.
module load gcc/10.1.0-centos7 mpi/mpich-3.2-x86_64
Install mpi4py
python -m pip install --no-cache-dir mpi4py
Now install dask-mpi
python -m pip install dask_mpi --upgrade
Usage
Running one or more upload jobs
The jobs can be defined inside a csv file. The first row of the csv file needs the following headers. Some are required for the job to run, and others are optional.
Required
s3_bucket: S3 Bucket name
experiment_type: One of [confocal, diSPIM, ecephys, exaSPIM, FIP, fMOST, HSFP, mesoSPIM, MRI, Other, SmartSPIM]
modality: One of [CONFOCAL, DISPIM, ECEPHYS, EPHYS, EXASPIM, FIP, FMOST, HSFP, ICEPHYS, MESOSPIM, MRI, OPHYS, SMARTSPIM, SPIM]
subject_id: ID of the subject
acq_date: Format can be either yyyy-MM-dd or MM/dd/yyyy
acq_time: Format can be either HH:mm:ss or HH-mm-ss
data_source: path to raw data folder
Somewhat Optional. Set the aws_param_store_name, but can define custom endpoints if desired
aws_param_store_name: Path to aws_param_store_name to retrieve common endpoints
If aws_param_store_name not set...
codeocean_domain: Domain of Code Ocean platform
codeocean_trigger_capsule_id: Launch a Code Ocean pipeline
codeocean_trigger_capsule_version: Optional if Code Ocean pipeline is versioned
metadata_service_domain: Domain name of the metadata service
aind_data_transfer_repo_location: The link to this project
video_encryption_password: Password with which to encrypt video files
codeocean_api_token: Code Ocean token used to run a capsule
Optional
temp_directory: The job will use your OS's file system to create a temp directory as default. You can override the location by setting this parameter.
behavior_dir: Location where behavior data associated with the raw data is stored.
metadata_dir: Location where metadata associated with the raw data is stored.
extra_configs: For more complicated jobs, the configs can be defined in a separate file.
log_level: Default log level is warning. Can be set here.
Optional Flags
metadata_dir_force: Default is false. If true, the metadata in the metadata folder will be regarded as the source of truth vs. the metadata pulled from aind_metadata_service
dry_run: Default is false. If set to true, it will perform a dry-run of the upload portion and not actually upload anything.
compress_raw_data: If experiment_type is ecephys, default is true. Else, default is false. If set to true, the data in the data_source path will be compressed into a zip file before being uploaded to S3.
After creating the csv file, you can run through the jobs with
python -m aind_data_transfer.jobs.s3_upload_job --jobs-csv-file "path_to_jobs_list"
Any Optional Flags attached will persist and override those set in the csv file. For example,
python -m aind_data_transfer.jobs.s3_upload_job --jobs-csv-file "path_to_jobs_list" --dry-run --compress-raw-data
will compress the raw data source and run a dry run for all jobs defined in the csv file.
An example csv file might look like:
data-source, s3-bucket, subject-id, modality, experiment_type, acq-date, acq-time, aws_param_store_name
dir/data_set_1, some_bucket, 123454, ECEPHYS, ecephys, 2020-10-10, 14-10-10, /aind/data/transfer/endpoints
dir/data_set_2, some_bucket2, 123456, OPHYS, Other, 2020-10-11, 13-10-10, /aind/data/transfer/endpoints
Contributing
Linters and testing
There are several libraries used to run linters, check documentation, and run tests.
- Please test your changes using the coverage library, which will run the tests and log a coverage report:
coverage run -m unittest discover && coverage report
- Use interrogate to check that modules, methods, etc. have been documented thoroughly:
interrogate .
- Use flake8 to check that code is up to standards (no unused imports, etc.):
flake8 .
- Use black to automatically format the code into PEP standards:
black .
- Use isort to automatically sort import statements:
isort .
Pull requests
For internal members, please create a branch. For external members, please fork the repo and open a pull request from the fork. We'll primarily use Angular style for commit messages. Roughly, they should follow the pattern:
<type>(<scope>): <short summary>
where scope (optional) describes the packages affected by the code changes and type (mandatory) is one of:
- build: Changes that affect the build system or external dependencies (example scopes: pyproject.toml, setup.py)
- ci: Changes to our CI configuration files and scripts (examples: .github/workflows/ci.yml)
- docs: Documentation only changes
- feat: A new feature
- fix: A bug fix
- perf: A code change that improves performance
- refactor: A code change that neither fixes a bug nor adds a feature
- test: Adding missing tests or correcting existing tests