Minor clarifications.

docs/api.rst

@@ -42,8 +42,7 @@ Running inference
 
 .. autofunction:: tsinfer.match_samples
 
-.. todo::
-    1. Add documentation for path compression here.
+.. autofunction:: tsinfer.augment_ancestors
 
 ++++++++++
 Exceptions

docs/inference.rst

@@ -197,7 +197,7 @@ The final phase of a ``tsinfer`` inference consists of a number steps:
    for our sample haplotypes through the ancestors. Each copying path
    corresponds to a set of tree sequence edges in precisely the same
    way as for ancestors, and the path compression algorithm can be equally
-   applied here. 
+   applied here.
 
 
 2. As we only use a subset of the available sites for inference
@@ -213,7 +213,7 @@ The final phase of a ``tsinfer`` inference consists of a number steps:
    may be clades of ancestral state samples which would allow us
    to encode the data with fewer back mutations.**
 
-3. Remove ancestral paths that do not lead to any of the samples by 
+3. Remove ancestral paths that do not lead to any of the samples by
    `simplifying
    <https://tskit.readthedocs.io/en/latest/python-api.html#tskit.TreeSequence.simplify>`_
    the final tree sequence. When simplifying, we keep non-branching ("unary")

docs/installation.rst

@@ -4,7 +4,7 @@
 Installation
 ############
 
-Python 3.3 or newer is required for ``tsinfer``. Any Unix-like platform should
+Python 3.5 or newer is required for ``tsinfer``. Any Unix-like platform should
 work (``tsinfer`` is tested on Linux, OS X, and Windows).
 
 Please use ``pip`` to install,

setup.py

@@ -49,7 +49,7 @@ def finalize_options(self):
         "tqdm",
         "humanize",
         "daiquiri",
-        "tskit>=0.2.0a3",
+        "tskit>=0.2.1",
         "numcodecs>=0.6",
         "zarr>=2.2",
         "lmdb",

tests/test_inference.py

@@ -311,7 +311,7 @@ class TestZeroInferenceSites(unittest.TestCase):
     Tests for the degenerate case in which we have no inference sites.
     """
     def verify(self, genotypes):
-        genotypes = np.array(genotypes, dtype=np.uint8)
+        genotypes = np.array(genotypes, dtype=np.int8)
         m = genotypes.shape[0]
         with tsinfer.SampleData(sequence_length=m + 1) as sample_data:
             for j in range(m):
@@ -628,8 +628,7 @@ def verify_inserted_ancestors(self, ts):
             tsinfer.check_ancestors_ts(ancestors_ts)
             self.assertEqual(ancestor_data.num_sites, ancestors_ts.num_sites)
             self.assertEqual(ancestor_data.num_ancestors, ancestors_ts.num_samples)
-            self.assertTrue(np.array_equal(
-                ancestors_ts.genotype_matrix(impute_missing_data=True), A))
+            self.assertTrue(np.array_equal(ancestors_ts.genotype_matrix(), A))
             inferred_ts = tsinfer.match_samples(
                 sample_data, ancestors_ts, engine=engine)
             self.assertTrue(np.array_equal(
@@ -1230,6 +1229,14 @@ def verify(self, ts):
             list(ts2.samples()),
             list(range(ts2.num_nodes - n, ts2.num_nodes)))
 
+        # Check that we're calling simplify with the correct arguments.
+        ts2 = tsinfer.infer(sd, simplify=False).simplify(keep_unary=True)
+        t1 = ts1.dump_tables()
+        t2 = ts2.dump_tables()
+        t1.provenances.clear()
+        t2.provenances.clear()
+        self.assertEqual(t1, t2)
+
     def test_single_tree(self):
         ts = msprime.simulate(5, random_seed=1, mutation_rate=2)
         self.verify(ts)

tsinfer/inference.py

@@ -250,8 +250,8 @@ def augment_ancestors(
         sample_data, ancestors_ts, indexes, num_threads=0, path_compression=True,
         extended_checks=False, engine=constants.C_ENGINE, progress_monitor=None):
     """
-    augment_ancestors(sample_data, ancestors_ts, indexes, num_threads=0,
-    path_compression=True)
+    augment_ancestors(sample_data, ancestors_ts, indexes, num_threads=0,\
+        path_compression=True)
 
     Runs the sample matching :ref:`algorithm <sec_inference_match_samples>`
     on the specified :class:`SampleData` instance and ancestors tree sequence,
@@ -290,7 +290,7 @@ def match_samples(
         extended_checks=False, stabilise_node_ordering=False, engine=constants.C_ENGINE,
         progress_monitor=None):
     """
-    match_samples(sample_data, ancestors_ts, num_threads=0, path_compression=True,
+    match_samples(sample_data, ancestors_ts, num_threads=0, path_compression=True,\
         simplify=True)
 
     Runs the sample matching :ref:`algorithm <sec_inference_match_samples>`