pqp.refutation package

Submodules

pqp.refutation.result module

class pqp.refutation.result.Assumption(name)

Bases: object

Class representing a modeling assumption

name

the name of the assumption

Type:

str

class pqp.refutation.result.Derived(name, value)

Bases: object

Class representing a derived value or expression

name

the name of the derived value

Type:

str

value

the derived value

Type:

Any

class pqp.refutation.result.Operation(op, args, kwargs)

Bases: object

Class used to track the computational graph of function calls which led to a result

op

the function which was called

Type:

Callable

args

the arguments passed to the function

Type:

list

kwargs

the keyword arguments passed to the function

Type:

dict

class pqp.refutation.result.Result(operation, step)

Bases: object

Class to store results of computation and store dependencies in a graph

explain(nested=False)

Prints a human-readable explanation of the result to the console.

Parameters:

nested (bool) – if True, will print the explanation of all prior steps as well

explain_all()

Prints a human-readable explanation of the result to the console, including all prior steps. Equivalent to calling self.explain(nested=True)

get_dependencies()

Returns a list of all results which this result directly depends on

Returns:

the list of dependencies

Return type:

List[Result]

get_nested_dependencies()

Returns a list of all results which this result depends on, including indirect dependencies

Returns:

the list of dependencies

Return type:

List[Result]

class pqp.refutation.result.Step(name)

Bases: object

Human-interpretable notes on how a result was derived

Subroutes in the library should use this class to record the steps taken to derive a result. This class is not intended to be used by end-users.

Parameters:

name (str) – the name of the step

assume(assumption)

Add an assumption to the derivation

Parameters:

assumption (Assumption or str) – the assumption to add

Returns:

None

explain(pad=True)

Print an explanation of the derivation of the result

result(key, value)

Record the value derived during the step

Parameters:

• key (str) – the name of the derived value

• value (Any) – the derived value

Returns:

None

substep(name)

Add a substep to the derivation

Parameters:

name (str) – the name of the substep

Returns:

the substep

Return type:

Step

write(msg)

Add a note about the derivation

Parameters:

msg (str) – the message to add

Returns:

None

pqp.refutation.result.entrypoint(step_name, result_class=<class 'pqp.refutation.result.Result'>)

Wrapper for logical steps in the assumption management system

Parameters:

• step_name (str) – the name of the Step (human readable)

• result_class (class) – used to store results of the Step, must be a subclass of Result

Returns:

wrapper for functions implementing logical steps in an analysis

Return type:

decorator

This function returns a a decorator which can be used to wrap a method implementing a step, a Step instance is passed as a keyword argument (“step”) to the method and should be used to record substeps, assumptions, and results. The method should not return a value.

The three core constructs of the assumption management system are Assumption s, Step s, and Result s. A step is a logical unit of work encapsulated in a single method, which may be composed of multiple substeps. During the process of calculation, the method can report logical substeps, assumptions it is making, results as they are computed, and notes for the user. This information is consolidated into the Step object, which stores references to previous steps in the analysis in a graph representing the entire computation. This way, metadata about the computation is available at all times and can be used by both the user and automated sensitivity analysis tools.

Module contents