wepy.analysis.transitions module¶
Collection of routines for counting transitions and calculating transition probability matrices.
There are two basic data structures used in this module for edge transitions: dictionary based and matrix based.
The dictionary based approach is used to interface with graph representations, like in the ~wepy.analysis.network.MacroStateNetwork~.
Arrays are used as an interface to numerical libraries which use this to calculate other properties such as steady-state probabilities.
Wepy is primarily meant to support expedient data structuring and will not support the latter numerical calculations and is left up to external libraries and domain specific code.
Routines¶
- transition_countsGiven a structured array of assignments of
microstate to macrostate labels and a listing of all kinetic transitions over the microstates returns the counts of transitions between each macrostate as a dictionary mapping macrostate transition edges to the total counts each microstate transition.
- counts_d_to_matrixConverts a dictionary mapping macrostate
transition edges to a asymmetric transition matrix.
- normalize_countsNormalizes outgoing transition counts from each
macrostate to 1 for a transition counts matrix.
- transition_counts_matrixGiven a structured array of assignments of
microstate to macrostate labels and a listing of all kinetic transitions over the microstates returns the counts of transitions between each macrostate as an assymetric transition counts matrix.
- transition_probability_matrixGiven a structured array of assignments of
microstate to macrostate labels and a listing of all kinetic transitions over the microstates returns the probability of transitions between each macrostate as an assymetric transition probability matrix.
- run_transition_counts_matrixGenerates an asymmetric transition
counts matrix directly from a single WepyHDF5 run.
- run_transition_probability_matrixGenerates an asymmetric transition
probability matrix directly from a single WepyHDF5 run.
- wepy.analysis.transitions.transition_counts(assignments, transitions, weights=None)[source]¶
Make a dictionary of the count of microstate transitions between macrostates.
If weights are given counts are the weight instead of 1.
- Parameters:
assignments (mixed array_like of dim (n_run, n_traj, n_cycle) type int) – Assignment of microstates to macrostate labels, where N_runs is the number of runs, N_traj is the number of trajectories, and N_cycle is the number of cycles.
transitions (list of list of tuples of ints (run_idx, traj_idx, cycle_idx)) – List of run traces corresponding to transitions. Only the first and last trace elements are used.
weights (mixed array_like of dim (n_run, n_traj, n_cycle) type float or None) –
- Same shape as the assignments, but gives the weight of the walker at this time.
(Optional)
- Returns:
transition_counts – Dictionary mapping transitions between macrostate labels to the count of microstate transitions between them. If weights are given this is the weighted counts.
- Return type:
- wepy.analysis.transitions.counts_d_to_matrix(counts_d)[source]¶
Convert a dictionary of counts for macrostate transitions to an assymetric transitions counts matrix.
- Parameters:
counts_d (dict of (int, int): int) – Dictionary mapping transitions between macrostate labels to the count of microstate transitions between them. The first value is the ‘source’ node and the second value is the ‘target’ node of a directed graph.
- Returns:
counts_matrix – Assymetric counts matrix of dim (n_macrostates, n_macrostates). The 0-th axis corresponds to the ‘source’ node and the 1-st axis corresponds to the ‘target’ nodes, i.e. the dimensions mean: (source, target).
- Return type:
numpy.ndarray of int
- wepy.analysis.transitions.normalize_counts(transition_counts_matrix)[source]¶
Normalize the macrostate outgoing transition counts to 1.0 for each macrostate.
- wepy.analysis.transitions.transition_counts_matrix(assignments, transitions)[source]¶
Make an asymmetric array of the count of microstate transitions between macrostates.
- Parameters:
assignments (mixed array_like of dim (n_run, n_traj, n_cycle) type int) – Assignment of microstates to macrostate labels, where N_runs is the number of runs, N_traj is the number of trajectories, and N_cycle is the number of cycles.
transitions (list of list of tuples of ints (run_idx, traj_idx, cycle_idx)) – List of run traces corresponding to transitions. Only the first and last trace elements are used.
- Returns:
transition_counts_matrix – Assymetric counts matrix of dim (n_macrostates, n_macrostates).
- Return type:
numpy.ndarray of int
- wepy.analysis.transitions.transition_probability_matrix(assignments, transitions)[source]¶
Make an asymmetric array of macrostates transition probabilities from microstate assignments.
- Parameters:
assignments (mixed array_like of dim (n_run, n_traj, n_cycle) type int) – Assignment of microstates to macrostate labels, where N_runs is the number of runs, N_traj is the number of trajectories, and N_cycle is the number of cycles.
transitions (list of list of tuples of ints (run_idx, traj_idx, cycle_idx)) – List of run traces corresponding to transitions. Only the first and last trace elements are used.
- Returns:
transition_probability_matrix – Assymetric transition probability matrix of dim (n_macrostates, n_macrostates).
- Return type:
numpy.ndarray of float
- wepy.analysis.transitions.run_transition_counts_matrix(wepy_hdf5, run_idx, assignment_key, transitions)[source]¶
Generates an asymmetric transition counts matrix directly from a single WepyHDF5 run.
- Parameters:
- Returns:
transition_counts_matrix – Assymetric counts matrix of dim (n_macrostates, n_macrostates).
- Return type:
numpy.ndarray of int
- wepy.analysis.transitions.run_transition_probability_matrix(wepy_hdf5, run_idx, assignment_key, transitions)[source]¶
Generates an asymmetric transition counts matrix directly from a single WepyHDF5 run.
- Parameters:
- Returns:
transition_probability_matrix – Assymetric transition probability matrix of dim (n_macrostates, n_macrostates).
- Return type:
numpy.ndarray of float
