models.actual_causation
¶
Objects that represent structures used in actual causation.
- pyphi.models.actual_causation.greater_than_zero(alpha)¶
Return
True
if alpha is greater than zero, accounting for numerical errors.
- class pyphi.models.actual_causation.AcRepertoireIrreducibilityAnalysis(alpha, state, direction, mechanism, purview, partition, probability, partitioned_probability, node_labels=None)¶
A minimum information partition for ac_coef calculation.
These can be compared with the built-in Python comparison operators (
<
,>
, etc.). First, \(\alpha\) values are compared. Then, if these are equal up toPRECISION
, the size of the mechanism is compared.- alpha¶
This is the difference between the mechanism’s unpartitioned and partitioned actual probability.
- Type:
float
- state¶
state of system in specified direction (cause, effect)
- Type:
tuple[int]
- direction¶
The temporal direction specifiying whether this analysis should be calculated with cause or effect repertoires.
- Type:
str
- mechanism¶
The mechanism to analyze.
- Type:
tuple[int]
- purview¶
The purview over which the unpartitioned actual probability differs the least from the actual probability of the partition.
- Type:
tuple[int]
- partition¶
The partition that makes the least difference to the mechanism’s repertoire.
- probability¶
The probability of the state in the previous/next timestep.
- Type:
float
- partitioned_probability¶
The probability of the state in the partitioned repertoire.
- Type:
float
- unorderable_unless_eq = ['direction']¶
- order_by()¶
Return a list of values to compare for ordering.
The first value in the list has the greatest priority; if the first objects are equal the second object is compared, etc.
- __bool__()¶
An
AcRepertoireIrreducibilityAnalysis
isTrue
if it has \(\alpha > 0\).
- property phi¶
Alias for \(\alpha\) for PyPhi utility functions.
- to_json()¶
Return a JSON-serializable representation.
- class pyphi.models.actual_causation.CausalLink(ria, extended_purview=None)¶
A maximally irreducible actual cause or effect.
These can be compared with the built-in Python comparison operators (
<
,>
, etc.). First, \(\alpha\) values are compared. Then, if these are equal up toPRECISION
, the size of the mechanism is compared.- property alpha¶
The difference between the mechanism’s unpartitioned and partitioned actual probabilities.
- Type:
float
- property phi¶
Alias for \(\alpha\) for PyPhi utility functions.
- property mechanism¶
The mechanism for which the action is evaluated.
- Type:
list[int]
- property purview¶
The purview over which this mechanism’s \(\alpha\) is maximal.
- Type:
list[int]
- property extended_purview¶
List of purviews over which this causal link is maximally irreducible.
Note: It will contain multiple purviews iff causal link has undetermined actual causes/effects (e.g. two irreducible causes with same alpha over different purviews).
- Type:
tuple[tuple[int]]
- property ria¶
The irreducibility analysis for this mechanism.
- property node_labels¶
- unorderable_unless_eq = ['direction']¶
- order_by()¶
Return a list of values to compare for ordering.
The first value in the list has the greatest priority; if the first objects are equal the second object is compared, etc.
- __bool__()¶
An
CausalLink
isTrue
if \(\alpha > 0\).
- to_json()¶
Return a JSON-serializable representation.
- class pyphi.models.actual_causation.Event(actual_cause, actual_effect)¶
A mechanism which has both an actual cause and an actual effect.
- actual_cause¶
The actual cause of the mechanism.
- Type:
- actual_effect¶
The actual effect of the mechanism.
- Type:
Create new instance of Event(actual_cause, actual_effect)
- property mechanism¶
The mechanism of the event.
- class pyphi.models.actual_causation.Account(causal_links)¶
The set of
CausalLink
with \(\alpha > 0\). This includes both actual causes and actual effects.- to_json()¶
- classmethod from_json(dct)¶
- class pyphi.models.actual_causation.DirectedAccount(causal_links)¶
The set of
CausalLink
with \(\alpha > 0\) for one direction of a transition.
- class pyphi.models.actual_causation.AcSystemIrreducibilityAnalysis(alpha=None, direction=None, account=None, partitioned_account=None, transition=None, cut=None)¶
An analysis of transition-level irreducibility (\(\mathcal{A}\)).
Contains the \(\mathcal{A}\) value of the
Transition
, the causal account, and all the intermediate results obtained in the course of computing them.- alpha¶
The \(\mathcal{A}\) value for the transition when taken against this analysis, i.e. the difference between the unpartitioned account and this analysis’s partitioned account.
- Type:
float
- transition¶
The transition this analysis was calculated for.
- Type:
- property before_state¶
Return the actual previous state of the
Transition
.
- property after_state¶
Return the actual current state of the
Transition
.
- unorderable_unless_eq = ['direction']¶
- order_by()¶
Return a list of values to compare for ordering.
The first value in the list has the greatest priority; if the first objects are equal the second object is compared, etc.
- __bool__()¶
An
AcSystemIrreducibilityAnalysis
isTrue
if it has \(\mathcal{A} > 0\).
- to_json()¶