Implied timescales plot #200
Conversation
Codecov Report
@@ Coverage Diff @@
## main #200 +/- ##
==========================================
+ Coverage 93.57% 93.61% +0.03%
==========================================
Files 131 134 +3
Lines 10544 10644 +100
==========================================
+ Hits 9867 9964 +97
- Misses 677 680 +3
Continue to review full report at Codecov.
|
| self._lagtimes = np.asarray(lagtimes, dtype=int) | ||
| self._its = np.asarray(its) | ||
| assert self.its.ndim == 2 and self.its.shape[0] == self.n_lagtimes, \ | ||
| "its should be of shape (lagtimes, processes)." |
There was a problem hiding this comment.
good to check. it will fail if supplied ITS cover a different number of states, i.e., if some states drop out of the active set.
There was a problem hiding this comment.
its is now assumed to be a list of timescales per lagtime and this check only looks for the number of lagtimes
deeptime/plots/implied_timescales.py
Outdated
| @property | ||
| def n_processes(self) -> int: | ||
| r""" Number of processes. """ | ||
| return self.its.shape[1] |
There was a problem hiding this comment.
property not defined if different numbers of processes are covered for different lag times, i.e., due to active set size differences. maybe need an n_processes_max`?
| its = [] | ||
| its_stats = [] if is_bayesian else None | ||
|
|
||
| for model in models: |
There was a problem hiding this comment.
One could check for missing processes and pad arrays here.
There was a problem hiding this comment.
added padding in the ctor call
No description provided.