ersilia.db.environments package

Submodules

ersilia.db.environments.localdb module

class ersilia.db.environments.localdb.EnvironmentDb(config_json=None)

Bases: ErsiliaBase

Manages the storage and retrieval of model environments.

Specifically it is essential for keeping track of the environments associated with each model. It provides a database where models and their corresponding environments (such as Docker images or conda environments) are stored and managed.

Parameters:

config_json (dict, optional) – Configuration settings for initializing the environment database.

Examples

>>> env_db = EnvironmentDb(config_json)
>>> env_db.table = "conda"
>>> env_db.insert("model_id", "venv_name")

clean()

Cleans the database by deleting all entries from the current table.

create_table()

Create table if it does not exist.

delete(model_id, env)

Deletes a specific entry from the database by model ID and environment.

Parameters:

• model_id (str) – Identifier of the model.

• env (str) – Environment associated with the model.

envs_of_model(model_id)

Retrieves environments associated with a specific model ID.

Parameters:

model_id (str) – Identifier of the model to search for.

Returns:

Set of environments associated with the model ID.

Return type:

set

envs_with_same_model(env)

Retrieves environments that share the same model as the given environment.

Parameters:

env (str) – Environment to search for.

Returns:

Set of environments that share the same model.

Return type:

set

fetchall()

Retrieves all entries from the current table.

Returns:

List of all entries in the table.

Return type:

list

insert(model_id, env)

Inserts a model ID and environment into the database.

Parameters:

• model_id (str) – Identifier of the model.

• env (str) – Environment associated with the model.

models_of_env(env)

Retrieves model IDs associated with a specific environment.

Parameters:

env (str) – Environment to search for.

Returns:

Set of model IDs associated with the environment.

Return type:

set

models_with_same_env(model_id)

Retrieves model IDs that share the same environment as the given model ID.

Parameters:

model_id (str) – Identifier of the model to search for.

Returns:

Set of model IDs that share the same environment.

Return type:

set

property table

Gets the current table name.

Returns:

The current table name.

Return type:

str

ersilia.db.environments.managers module

class ersilia.db.environments.managers.CondaManager

Bases: object

environments()

class ersilia.db.environments.managers.DockerManager(config_json=None, preferred_port=None, with_bentoml=False)

Bases: ErsiliaBase

Manages Docker operations for Ersilia models.

It provides methods to build, run, and manage Docker images and containers associated with Ersilia models.

Parameters:

• config_json (dict, optional) – Configuration settings for initializing the Docker manager.

• preferred_port (int, optional) – Preferred port number for running Docker containers.

• with_bentoml (bool, optional) – Flag indicating whether to use BentoML for model serving.

Examples

>>> docker_manager = DockerManager(
...     config_json=config, preferred_port=8080
... )
>>> docker_manager.build(
...     model_id="eosxxxx",
...     docker_user="user",
...     docker_pwd="pass",
... )
>>> docker_manager.run(model_id="eosxxxx", workers=2)

build(model_id, docker_user, docker_pwd, use_cache=True)

Builds a Docker image for the model.

Parameters:

• model_id (str) – Identifier of the model.

• docker_user (str) – Docker Hub username.

• docker_pwd (str) – Docker Hub password.

• use_cache (bool, optional) – If True, use Docker’s cache when building the image.

build_with_bentoml(model_id, use_cache=True)

Builds a Docker image for the model using BentoML.

Parameters:

• model_id (str) – Identifier of the model.

• use_cache (bool, optional) – If True, use Docker’s cache when building the image.

build_with_ersilia(model_id, docker_user, docker_pwd)

Builds a Docker image for the model using Ersilia’s base image.

Parameters:

• model_id (str) – Identifier of the model.

• docker_user (str) – Docker Hub username.

• docker_pwd (str) – Docker Hub password.

container_exists(container_name)

Checks if a Docker container with the specified name exists.

Parameters:

container_name (str) – Name of the Docker container.

Returns:

True if the container exists, False otherwise.

Return type:

bool

containers(only_run)

Retrieves a dictionary of Docker containers related to Ersilia models.

Parameters:

only_run (bool) – If True, only return running containers.

Returns:

Dictionary of Docker containers with container names as keys.

Return type:

dict

containers_of_model(model_id, only_run, only_latest=True)

Retrieves Docker containers for a specific model.

Parameters:

• model_id (str) – Identifier of the model.

• only_run (bool) – If True, only return running containers.

• only_latest (bool, optional) – If True, only consider images with the latest tag.

Returns:

Dictionary of Docker containers for the model.

Return type:

dict

delete_containers(model_id)

Deletes all Docker containers associated with a model.

Parameters:

model_id (str) – Identifier of the model.

delete_image(img)

delete_images(model_id, purge_unnamed=True)

Deletes Docker images associated with a model.

Parameters:

• model_id (str) – Identifier of the model.

• purge_unnamed (bool, optional) – If True, also remove unnamed images.

image_exists(model_id)

Checks if a Docker image for the specified model exists.

Parameters:

model_id (str) – Identifier of the model.

Returns:

True if the Docker image exists, False otherwise.

Return type:

bool

images()

Retrieves a dictionary of Docker images related to Ersilia models.

Returns:

Dictionary of Docker images with image names as keys.

Return type:

dict

images_of_model(model_id, only_latest=True)

Retrieves Docker images for a specific model.

Parameters:

• model_id (str) – Identifier of the model.

• only_latest (bool, optional) – If True, only return images with the latest tag.

Returns:

Dictionary of Docker images for the model.

Return type:

dict

is_active()

Checks if Docker is active and running.

Returns:

True if Docker is active, False otherwise.

Return type:

bool

is_inside_docker()

Checks if the current process is running inside a Docker container.

Returns:

True if running inside Docker, False otherwise.

Return type:

bool

is_installed()

Checks if Docker is installed on the system.

Returns:

True if Docker is installed, False otherwise.

Return type:

bool

is_logged_in()

Checks if the user is logged in to Docker.

Returns:

True if the user is logged in to Docker, False otherwise.

Return type:

bool

prune()

Prunes unused Docker objects to free up space.

remove(model_id)

Removes the Docker image for the specified model.

Parameters:

model_id (str) – Identifier of the model.

remove_stopped_containers()

Removes all stopped Docker containers.

run(model_id, workers=1, enable_microbatch=True, memory=None)

Runs a Docker container for the specified model.

Parameters:

• model_id (str) – Identifier of the model.

• workers (int, optional) – Number of worker processes.

• enable_microbatch (bool, optional) – If True, enable micro-batching.

• memory (int, optional) – Memory limit for the container in gigabytes.

Returns:

Dictionary containing container name and port.

Return type:

dict

start(container_name)

Starts a stopped Docker container.

Parameters:

container_name (str) – Name of the Docker container to start.

stop(container_name)

Stops a running Docker container.

Parameters:

container_name (str) – Name of the Docker container to stop.

stop_containers(model_id)

Stops all running Docker containers associated with a model.

Parameters:

model_id (str) – Identifier of the model.

Module contents