Fix so that custom timescale is in generations not coalescent units

Author: nspope

tests/test_functions.py

@@ -49,6 +49,7 @@
 from tsdate.prior import ConditionalCoalescentTimes
 from tsdate.prior import fill_priors
 from tsdate.prior import gamma_approx
+from tsdate.prior import MixturePrior
 from tsdate.prior import PriorParams
 from tsdate.prior import SpansBySamples
 from tsdate.util import nodes_time_unconstrained
@@ -508,6 +509,14 @@ def test_two_tree_mutation_ts_intervals(self):
         tests = self.check_intervals(ts, delete_interval_ts, keep_interval_ts)
         assert np.all(tests)
 
+    def test_custom_timegrid_is_not_rescaled(self):
+        ts = utility_functions.two_tree_mutation_ts()
+        prior = MixturePrior(ts)
+        demography = PopulationSizeHistory(3)
+        timepoints = np.array([0, 300, 1000, 2000])
+        prior_grid = prior.make_discretized_prior(demography, timepoints=timepoints)
+        assert np.array_equal(prior_grid.timepoints, timepoints)
+
 
 class TestPriorVals:
     def verify_prior_vals(self, ts, prior_distr, **kwargs):

tsdate/prior.py

@@ -980,10 +980,7 @@ def fill_priors(
     datable_nodes[ts.samples()] = False
     datable_nodes = np.where(datable_nodes)[0]
 
-    if isinstance(population_size, (int, float, np.ndarray)):
-        population_size = demography.PopulationSizeHistory(population_size)
-
-    # convert coalescent time grid to generational time scale
+    # convert timepoints to generational timescale
     prior_times = base.NodeGridValues(
         ts.num_nodes,
         datable_nodes[np.argsort(ts.tables.nodes.time[datable_nodes])].astype(np.int32),
@@ -994,6 +991,7 @@ def fill_priors(
     for node in tqdm(
         datable_nodes, desc="Assign Prior to Each Node", disable=not progress
     ):
+        # NB: prior CDF is evaluated on coalescent timescale
         with np.errstate(divide="ignore", invalid="ignore"):
             prior_node = cdf_func(timepoints, main_param[node], scale=scale_param[node])
         # force age to be less than max value
@@ -1063,6 +1061,9 @@ def make_discretized_prior(self, population_size, timepoints=20, progress=False)
         Calculate prior grid for a set of timepoints and a population size history
         """
 
+        if isinstance(population_size, (int, float, np.ndarray)):
+            population_size = demography.PopulationSizeHistory(population_size)
+
         if isinstance(timepoints, int):
             if timepoints < 2:
                 raise ValueError("You must have at least 2 time points")
@@ -1080,6 +1081,9 @@ def make_discretized_prior(self, population_size, timepoints=20, progress=False)
                 raise ValueError("Timepoints cannot be negative")
             elif np.any(np.unique(timepoints, return_counts=True)[1] > 1):
                 raise ValueError("Timepoints cannot have duplicate values")
+            # timepoints are assumed to be on generational scale, so convert to
+            # coalescent timescale to evaluate prior
+            timepoints = population_size.to_coalescent_timescale(timepoints)
         else:
             raise ValueError(
                 "time_slices must be an integer or a numpy array of floats"