pyrevs.strategies.ams.sql

An extension class for the AMS data as an SQL database using SQLAlchemy.

Classes

SplittingIterations	A table storing the splitting iterations.
AMSDB	An extension of the core database holding AMS iterations repertoire.

Module Contents

class SplittingIterations

Bases: pyrevs.core.CoreBase

A table storing the splitting iterations.

id: sqlalchemy.orm.Mapped[int]

split_id: sqlalchemy.orm.Mapped[int]

bias: sqlalchemy.orm.Mapped[int]

weight: sqlalchemy.orm.Mapped[float]

discarded_traj_ids: sqlalchemy.orm.Mapped[list[int]]

ancestor_traj_ids: sqlalchemy.orm.Mapped[list[int]]

min_vals: sqlalchemy.orm.Mapped[list[float]]

min_max: sqlalchemy.orm.Mapped[list[float]]

status: sqlalchemy.orm.Mapped[str]

class AMSDB(engine: sqlalchemy.engine.Engine)

An extension of the core database holding AMS iterations repertoire.

Allows atomic access to an SQL database from all the workers.

Variables:

_file_name – The file name

session_scope() → collections.abc.Generator[sqlalchemy.orm.Session, None, None]

Provide a transactional scope around a series of operations.

add_splitting_data(k: int, bias: int, weight: float, discarded_ids: list[int], ancestor_ids: list[int], min_vals: list[float], min_max: list[float]) → None

Add a new splitting data to the DB.

Parameters:

• k – The splitting iteration index

• bias – The number of restarted trajectories

• weight – Weight of the ensemble at the current iteration

• discarded_ids – The list of discarded trajectory ids

• ancestor_ids – The list of trajectories used to restart

• min_vals – The list of minimum values

• min_max – The score minimum and maximum values

update_splitting_data(k: int, bias: int, weight: float, discarded_ids: list[int], ancestor_ids: list[int], min_vals: list[float], min_max: list[float]) → None

Update the last splitting data row to the DB.

Parameters:

• k – The splitting iteration index

• bias – The number of restarted trajectories

• weight – Weight of the ensemble at the current iteration

• discarded_ids – The list of discarded trajectory ids

• ancestor_ids – The list of trajectories used to restart

• min_vals – The list of minimum values

• min_max – The score minimum and maximum values

mark_last_iteration_as_completed() → None

Mark the last splitting iteration as complete.

By default, iteration data append to the SQL table with a state “locked” to indicate an iteration being worked on. Upon completion, mark it as “completed” otherwise the iteration is considered incomplete, i.e. interrupted by some error or wall clock limit.

get_k_split() → int

Get the current splitting iteration counter.

Returns:

The ksplit from the last entry in the SplittingIterations table

check_new_min_of_maxes(newmin: float) → None

Compare the incoming min to the last entry.

When running pyREVS with AMS, at each new iteration the ensemble minimum of maximum should be strictly above the previous iteration’s one.

Parameters:

newmin – the new minimum of maximums

get_iteration_count() → int

Get the number of splitting iteration stored.

Returns:

The length of the SplittingIterations table

fetch_splitting_data(k_id: int) → tuple[int, int, float, list[int], list[int], list[float], list[float], str] | None

Get the splitting iteration data for a given iteration.

Parameters:

k_id – The iteration id

Returns:

The splitting iteration data

Raises:

ValueError if the splitting iteration with the given id does not exist –

get_ongoing() → list[int] | None

Get the list of ongoing trajectories if any.

Returns:

Either a list trajectories or None if nothing was left to do

get_weights() → numpy.typing.NDArray[numpy.number]

Read the weights from the database.

Returns:

the weight for each splitting iteration as a numpy array

get_biases() → numpy.typing.NDArray[numpy.number]

Read the biases from the database.

Returns:

the bias for each splitting iteration as a numpy array

get_minmax() → numpy.typing.NDArray[numpy.number]

Read the min/max from the database.

Returns:

the 2D Numpy array with k_index, min, max

clear_splitting_data() → int

Delete the content of the splitting data table.

Returns:

The number of entries deleted

dump_file_json(json_file: str | None = None) → None

Dump the content of the trajectory table to a json file.

Parameters:

json_file – an optional file name (or path) to dump the data to