fix: precision of phenotypes written to pheno file (#199)

CAST-genomics · Mar 22, 2023 · a397c96 · a397c96
1 parent aade751
commit a397c96
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 5 deletions.
diff --git a/haptools/data/phenotypes.py b/haptools/data/phenotypes.py
@@ -192,15 +192,17 @@ def write(self):
         # now we can finally write the file
         with self.hook_compressed(self.fname, mode="w") as phens:
             phens.write("#IID\t" + "\t".join(names) + "\n")
-            formatter = {"float_kind": lambda x: "%.2f" % x}
             for samp, phen in zip(self.samples, self.data):
                 line = np.array2string(
                     phen,
+                    sign=" ",
+                    legacy=False,
                     separator="\t",
-                    formatter=formatter,
-                    max_line_width=np.inf,
                     threshold=np.inf,
                     edgeitems=np.inf,
+                    floatmode="unique",
+                    suppress_small=False,
+                    max_line_width=np.inf,
                 )[1:-1]
                 phens.write(f"{samp}\t" + line + "\n")
 

diff --git a/tests/test_data.py b/tests/test_data.py
@@ -585,12 +585,24 @@ def test_write_phenotypes(self):
         # now, let's load the data and check that it's what we wrote
         result = Phenotypes(expected_phen.fname)
         result.read()
-        np.testing.assert_allclose(expected_phen.data, result.data)
+        np.testing.assert_allclose(expected_phen.data, result.data, rtol=0, atol=0)
+        assert expected_phen.names == result.names
+        assert expected_phen.samples == result.samples
+
+        # try to standardize the data and see if it's still close
+        # to validate that our code can handle phenotypes with arbitrary precision
+        expected_phen.standardize()
+        expected_phen.write()
+
+        # now, let's load the data and check that it's what we wrote
+        result = Phenotypes(expected_phen.fname)
+        result.read()
+        np.testing.assert_allclose(expected_phen.data, result.data, rtol=0, atol=0)
         assert expected_phen.names == result.names
         assert expected_phen.samples == result.samples
 
         # let's clean up after ourselves and delete the file
-        os.remove(str(expected_phen.fname))
+        expected_phen.fname.unlink()
 
     def test_append_phenotype(self):
         expected1 = self._get_fake_phenotypes()