moscot.problems.generic.FGWProblem.prepare¶
- FGWProblem.prepare(key, joint_attr=None, x_attr=None, y_attr=None, policy='sequential', cost='sq_euclidean', cost_kwargs=mappingproxy({}), a=None, b=None, xy_callback=None, x_callback=None, y_callback=None, xy_callback_kwargs=mappingproxy({}), x_callback_kwargs=mappingproxy({}), y_callback_kwargs=mappingproxy({}), subset=None, reference=None)[source]¶
Prepare the individual quadratic subproblems.
See also
TODO(michalk8): add an example how to pass x_attr/y_attr.
- Parameters:
key (
str
) – Key inobs
for theSubsetPolicy
.joint_attr (
Union
[str
,Mapping
[str
,Any
],None
]) –How to get the data for the linear term in the fused case:
dict
- it should contain'attr'
and'key'
, the attribute and the key inAnnData
, and optionally'tag'
, one ofTag
.
By default,
tag = 'point_cloud'
is used.x_attr (
Union
[str
,Mapping
[str
,Any
],None
]) –How to get the data for the source quadratic term:
dict
- it should contain'attr'
and'key'
, the attribute and key inAnnData
, and optionally'tag'
from thetags
.None
-'x_callback'
must be passed viakwargs
.
By default,
tag = 'point_cloud'
is used.y_attr (
Union
[str
,Mapping
[str
,Any
],None
]) –How to get the data for the target quadratic term:
dict
- it should contain'attr'
and'key'
, the attribute and the key inAnnData
, and optionally'tag'
, one ofTag
.None
-'y_callback'
must be passed viakwargs
.
By default,
tag = 'point_cloud'
is used.policy (
Literal
['sequential'
,'explicit'
,'star'
]) –Rule which defines how to construct the subproblems. Valid options are:
'sequential'
- align subsequent categories inobs['{key}']
.'explicit'
- explicit sequence of subsets passed viasubset = [(b3, b0), ...]
.
cost (
Union
[Literal
['euclidean'
,'sq_euclidean'
,'cosine'
,'pnorm_p'
,'sq_pnorm'
,'geodesic'
],Mapping
[Literal
['xy'
,'x'
,'y'
],Literal
['euclidean'
,'sq_euclidean'
,'cosine'
,'pnorm_p'
,'sq_pnorm'
,'geodesic'
]]]) –Cost function to use. Valid options are:
str
- name of the cost function for all terms, seeget_available_costs()
.dict
- a dictionary with the following keys and values:'xy'
- cost function for the linear term.'x'
- cost function for the source quadratic term.'y'
- cost function for the target quadratic term.
cost_kwargs (
Union
[Mapping
[str
,Any
],Mapping
[Literal
['x'
,'y'
,'xy'
],Mapping
[str
,Any
]]]) – Keyword arguments for theBaseCost
or any backend-specific cost.Source marginals. Valid options are:
Target marginals. Valid options are:
xy_callback (
Union
[Literal
['local-pca'
],Callable
[[Literal
['xy'
,'x'
,'y'
],AnnData
,Optional
[AnnData
]],Optional
[TaggedArray
]],None
]) – Callback function used to prepare the data in the linear term.x_callback (
Union
[Literal
['local-pca'
],Callable
[[Literal
['xy'
,'x'
,'y'
],AnnData
,Optional
[AnnData
]],Optional
[TaggedArray
]],None
]) – Callback function used to prepare the data in the source quadratic term.y_callback (
Union
[Literal
['local-pca'
],Callable
[[Literal
['xy'
,'x'
,'y'
],AnnData
,Optional
[AnnData
]],Optional
[TaggedArray
]],None
]) – Callback function used to prepare the data in the target quadratic term.xy_callback_kwargs (
Mapping
[str
,Any
]) – Keyword arguments for thexy_callback
.x_callback_kwargs (
Mapping
[str
,Any
]) – Keyword arguments for thex_callback
.y_callback_kwargs (
Mapping
[str
,Any
]) – Keyword arguments for they_callback
.subset (
Optional
[Sequence
[Tuple
[TypeVar
(K
, bound=Hashable
),TypeVar
(K
, bound=Hashable
)]]]) – Subset ofobs['{key}']
for theExplicitPolicy
. Only used whenpolicy = 'explicit'
.reference (
Optional
[Any
]) – Reference for theSubsetPolicy
. Only used whenpolicy = 'star'
.
- Return type:
FGWProblem
[TypeVar
(K
, bound=Hashable
),TypeVar
(B
, bound=OTProblem
)]- Returns:
: Returns self and updates the following fields: