Schematic

TL;DR

• schematic (Schema Engine for Manifest Ingress and Curation) is a python-based software tool that streamlines the retrieval, validation, and submission of metadata for biomedical datasets hosted on Sage Bionetworks' Synapse platform.
• Users can work with schematic in several ways, including through the CLI (see Command Line Usage for examples), through Docker (see Docker Usage for examples), or with python.
• schematic needs to communicate with Synapse and Google Sheets in order for its processes to work. As such, users will need to set up their credentials for authentication with Synapse and the Google Sheets API.
• To get started with schematic, follow one of the Installation Guides depending on your use case:
  • Installation Guide For: Schematic CLI users
  • Installation Guide For: Contributors

Table of Contents

• Schematic
• TL;DR
• Table of Contents
• Introduction
• Installation
  • Installation Requirements
  • Installation Guide For: Users
    • 1. Verify your python version
    • 2. Set up your virtual environment
      • 2a. Set up your virtual environment with venv
      • 2b. Set up your virtual environment with conda
    • 3. Install schematic dependencies
    • 4. Get your data model as a JSON-LD schema file
    • 5. Obtain Google credential files
    • 6. Set up configuration files
    • 7. Verify your setup
  • Installation Guide For: Contributors
    • 1. Clone the schematic package repository
    • 2. Install poetry
    • 3. Start the virtual environment
    • 4. Install schematic dependencies
    • 5. Set up configuration files
    • 6. Obtain Google credential files
    • 7. Set up pre-commit hooks
    • 8. Verify your setup
• Command Line Usage
• Docker Usage
  • Running the REST API
    • Example 1: Using the config.yml path
    • Example 2: Use environment variables
  • Running schematic to Validate Manifests
    • Example for macOS/Linux
    • Example for Windows
• Contributors

Introduction

SCHEMATIC is an acronym for Schema Engine for Manifest Ingress and Curation. The Python based infrastructure provides a novel schema-based, metadata ingress ecosystem, that is meant to streamline the process of biomedical dataset annotation, metadata validation and submission to a data repository for various data contributors.

Installation

Installation Requirements

• Your installed python version must be 3.9.0 ≤ version < 3.11.0
• You need to be a registered and certified user on synapse.org

Note

To create Google Sheets files from Schematic, please follow our credential policy for Google credentials. You can find a detailed tutorial here. If you're using config.yml, make sure to specify the path to schematic_service_account_creds.json (see the google_sheets > service_account_creds section for more information).

Installation Guide For: Users

The instructions below assume you have already installed python, with the release version meeting the constraints set in the Installation Requirements section, and do not have a Python environment already active.

1. Verify your python version

Ensure your python version meets the requirements from the Installation Requirements section using the following command:

python3 --version

If your current Python version is not supported by Schematic, you can switch to the supported version using a tool like pyenv. Follow the instructions in the pyenv documentation to install and switch between Python versions easily.

Note

You can double-check the current supported python version by opening up the pyproject.toml file in this repository and find the supported versions of python in the script.

2. Set up your virtual environment

Once you are working with a python version supported by schematic, you will need to activate a virtual environment within which you can install the package. Below we will show how to create your virtual environment either with venv or with conda.

2a. Set up your virtual environment with venv

Python 3 has built-in support for virtual environments with the venv module, so you no longer need to install virtualenv:

python3 -m venv .venv
source .venv/bin/activate

2b. Set up your virtual environment with conda

conda is a powerful package and environment management tool that allows users to create isolated environments used particularly in data science and machine learning workflows. If you would like to manage your environments with conda, continue reading:

1. Download your preferred conda installer: Begin by installing conda. We personally recommend working with Miniconda which is a lightweight installer for conda that includes only conda and its dependencies.

2. Execute the conda installer: Once you have downloaded your preferred installer, execute it using bash or zsh, depending on the shell configured for your terminal environment. For example:

   bash Miniconda3-latest-MacOSX-arm64.sh
   

3. Verify your conda setup: Follow the prompts to complete your setup. Then verify your setup by running the conda command.

4. Create your schematic environment: Begin by creating a fresh conda environment for schematic like so:

   conda create --name 'schematicpy' python=3.10
   

5. Activate the environment: Once your environment is set up, you can now activate your new environment with conda:

   conda activate schematicpy
   

3. Install schematic dependencies

Install the package using pip:

python3 -m pip install schematicpy

If you run into ERROR: Failed building wheel for numpy, the error might be able to resolve by upgrading pip. Please try to upgrade pip by:

pip3 install --upgrade pip

4. Get your data model as a JSON-LD schema file

Now you need a schema file, e.g. model.jsonld, to have a data model that schematic can work with. While you can download a super basic example data model here, you’ll probably be working with a DCC-specific data model. For non-Sage employees/contributors using the CLI, you might care only about the minimum needed artifact, which is the .jsonld; locate and download only that from the right repo.

Here are some example repos with schema files:

• https://github.com/ncihtan/data-models/
• https://github.com/nf-osi/nf-metadata-dictionary/

5. Obtain Google credential files

Any function that interacts with a google sheet (such as schematic manifest get) requires google cloud credentials.

1. Option 1: Here’s a step-by-step guide on how to create these credentials in Google Cloud.
   • Depending on your institution's policies, your institutional Google account may or may not have the required permissions to complete this. A possible workaround is to use a personal or temporary Google account.

Warning

At the time of writing, Sage Bionetworks employees do not have the appropriate permissions to create projects with their Sage Bionetworks Google accounts. You would follow instructions using a personal Google account.

2. Option 2: Ask your DCC/development team if they have credentials previously set up with a service account.

Once you have obtained credentials, be sure that the json file generated is named in the same way as the service_acct_creds parameter in your config.yml file. You will find more context on the config.yml in section 6. Set up configuration files.

Note

Running schematic init is no longer supported due to security concerns. To obtain schematic_service_account_creds.json, please follow the instructions here. schematic uses Google’s API to generate google sheet templates that users fill in to provide (meta)data. Most Google sheet functionality could be authenticated with service account. However, more complex Google sheet functionality requires token-based authentication. As browser support that requires the token-based authentication diminishes, we are hoping to deprecate token-based authentication and keep only service account authentication in the future.

Note

Use the schematic_service_account_creds.json file for the service account mode of authentication (for Google services/APIs). Service accounts are special Google accounts that can be used by applications to access Google APIs programmatically via OAuth2.0, with the advantage being that they do not require human authorization.

6. Set up configuration files

The following section will walk through setting up your configuration files with your credentials to allow for communication between schematic and the Synapse API.

There are two main configuration files that need to be created + modified:

• .synapseConfig
• config.yml

Create and modify the .synapseConfig

The .synapseConfig file is what enables communication between schematic and the Synapse API using your credentials. You can automatically generate a .synapseConfig file by running the following in your command line and following the prompts.

Tip

You can generate a new authentication token on the Synapse website by going to Account Settings > Personal Access Tokens.

synapse config

After following the prompts, a new .synapseConfig file and .synapseCache folder will be created in your home directory. You can view these hidden assets in your home directory with the following command:

ls -a ~

The .synapseConfig is used to log into Synapse if you are not using an environment variable (i.e. SYNAPSE_ACCESS_TOKEN) for authentication, and the .synapseCache is where your assets are stored if you are not working with the CLI and/or you have specified .synapseCache as the location in which to store your manfiests, in your config.yml (more on the config.yml below).

Create and modify the config.yml

In this repository there is a config_example.yml file with default configurations to various components that are required before running schematic, such as the Synapse ID of the main file view containing all your project assets, the base name of your manifest files, etc.

Download the config_example.yml as a new file called config.yml and modify its contents according to your use case.

For example, one of the components in this config.yml that will likely be modified is the location of your schema. After acquiring your schema file using the instructions in step 4. Get your data model as a JSON-LD schema file, your config.yml should contain something like:

model:
  location: "path/to/your/model.jsonld"

Important

Please note that for the example above, your local working directory would typically have model.jsonld and config.yml side-by-side. The path to your data model should match what is in config.yml.

Important

Be sure to update your config.yml with the location of your .synapseConfig created in the step above, to avoid authentication errors. Paths can be specified relative to the config.yml file or as absolute paths.

Note

config.yml is ignored by git.

7. Verify your setup

After running the steps above, your setup is complete, and you can test it on a python instance or by running a command based on the examples in the Command Line Usage section.

Installation Guide For: Contributors

The instructions below assume you have already installed python, with the release version meeting the constraints set in the Installation Requirements section, and do not have an environment already active (e.g. with pyenv). For development, we recommend working with versions > python 3.9 to avoid issues with pre-commit's default hook configuration.

When contributing to this repository, please first discuss the change you wish to make via the service desk so that we may track these changes.

Once you have finished setting up your development environment using the instructions below, please follow the guidelines in CONTRIBUTION.md during your development.

Please note we have a code of conduct, please follow it in all your interactions with the project.

1. Clone the schematic package repository

For development, you will be working with the latest version of schematic on the repository to ensure compatibility between its latest state and your changes. Ensure your current working directory is where you would like to store your local fork before running the following command:

git clone https://github.com/Sage-Bionetworks/schematic.git

2. Install poetry

Install poetry (version 1.3.0 or later) using either the official installer or pip. If you have an older installation of Poetry, we recommend uninstalling it first.

pip install poetry

Check to make sure your version of poetry is > v1.3.0

poetry --version

3. Start the virtual environment

cd into your cloned schematic repository, and initialize the virtual environment using the following command with poetry:

poetry shell

To make sure your poetry version and python version are consistent with the versions you expect, you can run the following command:

poetry debug info

4. Install schematic dependencies

Before you begin, make sure you are in the latest develop of the repository.

The following command will install the dependencies based on what we specify in the poetry.lock file of this repository. If this step is taking a long time, try to go back to Step 2 and check your version of poetry. Alternatively, you can try deleting the lock file and regenerate it by doing poetry install (Please note this method should be used as a last resort because this would force other developers to change their development environment)

poetry install --all-extras

5. Set up configuration files

The following section will walk through setting up your configuration files with your credentials to allow for communication between schematic and the Synapse API.

There are two main configuration files that need to be created + modified:

• .synapseConfig
• config.yml

Create and modify the .synapseConfig

The .synapseConfig file is what enables communication between schematic and the Synapse API using your credentials. You can automatically generate a .synapseConfig file by running the following in your command line and following the prompts.

Tip

You can generate a new authentication token on the Synapse website by going to Account Settings > Personal Access Tokens.

synapse config

After following the prompts, a new .synapseConfig file and .synapseCache folder will be created in your home directory. You can view these hidden assets in your home directory with the following command:

ls -a ~

The .synapseConfig is used to log into Synapse if you are not using an environment variable (i.e. SYNAPSE_ACCESS_TOKEN) for authentication, and the .synapseCache is where your assets are stored if you are not working with the CLI and/or you have specified .synapseCache as the location in which to store your manfiests, in your config.yml (more on the config.yml below).

Important

When developing on schematic, keep your .synapseConfig in your current working directory to avoid authentication errors.

Create and modify the config.yml

In this repository there is a config_example.yml file with default configurations to various components that are required before running schematic, such as the Synapse ID of the main file view containing all your project assets, the base name of your manifest files, etc.

Copy the contents of the config_example.yml (located in the base directory of the cloned schematic repo) into a new file called config.yml

cp config_example.yml config.yml

Once you've copied the file, modify its contents according to your use case. For example, if you wanted to change the folder where manifests are downloaded your config should look like:

manifest:
  manifest_folder: "my_manifest_folder_path"

Important

Be sure to update your config.yml with the location of your .synapseConfig created in the step above, to avoid authentication errors. Paths can be specified relative to the config.yml file or as absolute paths.

Note

config.yml is ignored by git.

6. Obtain Google credential files

Any function that interacts with a google sheet (such as schematic manifest get) requires google cloud credentials.

1. Option 1: Here’s a step-by-step guide on how to create these credentials in Google Cloud.
   • Depending on your institution's policies, your institutional Google account may or may not have the required permissions to complete this. A possible workaround is to use a personal or temporary Google account.

Warning

At the time of writing, Sage Bionetworks employees do not have the appropriate permissions to create projects with their Sage Bionetworks Google accounts. You would follow instructions using a personal Google account.

2. Option 2: Ask your DCC/development team if they have credentials previously set up with a service account.

Once you have obtained credentials, be sure that the json file generated is named in the same way as the service_acct_creds parameter in your config.yml file.

Important

For testing, make sure there is no environment variable SCHEMATIC_SERVICE_ACCOUNT_CREDS. Check the file .env to ensure this is not set. Also, check that config files used for testing, such as config_example.yml do not contain service_acct_creds_synapse_id.

Note

Running schematic init is no longer supported due to security concerns. To obtain schematic_service_account_creds.json, please follow the instructions here. schematic uses Google’s API to generate google sheet templates that users fill in to provide (meta)data. Most Google sheet functionality could be authenticated with service account. However, more complex Google sheet functionality requires token-based authentication. As browser support that requires the token-based authentication diminishes, we are hoping to deprecate token-based authentication and keep only service account authentication in the future.

Note

Use the schematic_service_account_creds.json file for the service account mode of authentication (for Google services/APIs). Service accounts are special Google accounts that can be used by applications to access Google APIs programmatically via OAuth2.0, with the advantage being that they do not require human authorization.

7. Set up pre-commit hooks

This repository is configured to utilize pre-commit hooks as part of the development process. To enable these hooks, please run the following command and look for the following success message:

$ pre-commit install
pre-commit installed at .git/hooks/pre-commit

You can run pre-commit manually across the entire repository like so:

pre-commit run --all-files

After running this step, your setup is complete, and you can test it on a python instance or by running a command based on the examples in the Command Line Usage section.

8. Verify your setup

After running the steps above, your setup is complete, and you can test it on a python instance or by running a command based on the examples in the Command Line Usage section.

Command Line Usage

1. Generate a new manifest as a google sheet

schematic manifest -c /path/to/config.yml get -dt <your data type> -s

2. Grab an existing manifest from synapse

schematic manifest -c /path/to/config.yml get -dt <your data type> -d <your synapse dataset folder id> -s

3. Validate a manifest

schematic model -c /path/to/config.yml validate -dt <your data type> -mp <your csv manifest path>

4. Submit a manifest as a file

schematic model -c /path/to/config.yml submit -mp <your csv manifest path> -d <your synapse dataset folder id> -vc <your data type> -mrt file_only

Please visit more documentation here for more information.

Docker Usage

Here we will demonstrate how to run schematic with Docker, with different use-cases for running API endpoints, validating the manifests, and using how to use schematic based on your OS (macOS/Linux).

Running the REST API

Use the Docker image to run schematics REST API. You can either use the file path for the config.yml created using the installation instructions, or set up authentication with environment variables.

Example 1: Using the config.yml path

docker run --rm -p 3001:3001 \
  -v $(pwd):/schematic -w /schematic --name schematic \
  -e SCHEMATIC_CONFIG=/schematic/config.yml \
  -e GE_HOME=/usr/src/app/great_expectations/ \
  sagebionetworks/schematic \
  python /usr/src/app/run_api.py

Example 2: Use environment variables

1. save content of config.yml as to environment variable SCHEMATIC_CONFIG_CONTENT by doing: export SCHEMATIC_CONFIG_CONTENT=$(cat /path/to/config.yml)

2. Similarly, save the content of schematic_service_account_creds.json as SERVICE_ACCOUNT_CREDS by doing: export SERVICE_ACCOUNT_CREDS=$(cat /path/to/schematic_service_account_creds.json)

3. Pass SCHEMATIC_CONFIG_CONTENT and schematic_service_account_creds as environment variables by using docker run

docker run --rm -p 3001:3001 \
  -v $(pwd):/schematic -w /schematic --name schematic \
  -e GE_HOME=/usr/src/app/great_expectations/ \
  -e SCHEMATIC_CONFIG_CONTENT=$SCHEMATIC_CONFIG_CONTENT \
  -e SERVICE_ACCOUNT_CREDS=$SERVICE_ACCOUNT_CREDS \
  sagebionetworks/schematic \
  python /usr/src/app/run_api.py

Running schematic to Validate Manifests

You can also use Docker to run schematic commands like validating manifests. Below are examples for different platforms.

Example for macOS/Linux

1. Clone the repository:

git clone https://github.com/sage-bionetworks/schematic ~/schematic

2. Update the .synapseConfig with your credentials. See the installation instructions for how to do this.

3. Run Docker:

docker run \
  -v ~/schematic:/schematic \
  -w /schematic \
  -e SCHEMATIC_CONFIG=/schematic/config.yml \
  -e GE_HOME=/usr/src/app/great_expectations/ \
  sagebionetworks/schematic schematic model \
  -c /schematic/config.yml validate \
  -mp /schematic/tests/data/mock_manifests/Valid_Test_Manifest.csv \
  -dt MockComponent \
  -js /schematic/tests/data/example.model.jsonld

Example for Windows

Run the following command to validate manifests:

docker run -v %cd%:/schematic \
  -w /schematic \
  -e GE_HOME=/usr/src/app/great_expectations/ \
  sagebionetworks/schematic \
  schematic model \
  -c config.yml validate -mp tests/data/mock_manifests/inValid_Test_Manifest.csv -dt MockComponent -js /schematic/data/example.model.jsonld

Contributors

Main contributors and developers:

• Milen Nikolov
• Mialy DeFelice
• Sujay Patil
• Bruno Grande
• Robert Allaway
• Gianna Jordan
• Lingling Peng