Implement remaining `fix_samples` functionality

Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/partial-correlations.gmi

* gn3/partial_correlations.py: implement `fix_samples` function
* tests/unit/test_partial_correlations.py: implement tests for `fix_samples`
  function

  Implement the remaining partial migration for the
  `web.webqtl.correlation.correlationFunction.fixStrain` function in GN1.

2 files changed, 66 insertions, 16 deletions

diff --git a/gn3/partial_correlations.py b/gn3/partial_correlations.py
index 4db4807..c556d10 100644
--- a/gn3/partial_correlations.py
+++ b/gn3/partial_correlations.py
@@ -5,8 +5,8 @@ It is an attempt to migrate over the partial correlations feature from
 GeneNetwork1.
 """
 
-from typing import Sequence
 from functools import reduce
+from typing import Any, Sequence
 
 def control_samples(controls: Sequence[dict], sampleslist: Sequence[str]):
     """
@@ -45,7 +45,7 @@ def control_samples(controls: Sequence[dict], sampleslist: Sequence[str]):
         [__process_control__(trait_data) for trait_data in controls],
         (tuple(), tuple(), tuple(), tuple()))
 
-def dictify_by_samples(samples_vals_vars: Sequence[Sequence]) -> dict:
+def dictify_by_samples(samples_vals_vars: Sequence[Sequence]) -> Sequence[dict]:
     """
     Build a sequence of dictionaries from a sequence of separate sequences of
     samples, values and variances.
@@ -60,3 +60,29 @@ def dictify_by_samples(samples_vals_vars: Sequence[Sequence]) -> dict:
             sample: {"sample_name": sample, "value": val, "variance": var}
             for sample, val, var in zip(*trait_line)
         } for trait_line in zip(*(samples_vals_vars[0:3])))
+
+def fix_samples(primary_trait: dict, control_traits: Sequence[dict]) -> Sequence[Sequence[Any]]:
+    """
+    Corrects sample_names, values and variance such that they all contain only
+    those samples that are common to the reference trait and all control traits.
+
+    This is a partial migration of the
+    `web.webqtl.correlation.correlationFunction.fixStrain` function in GN1.
+    """
+    primary_samples = tuple(
+        present[0] for present in
+        ((sample, all(sample in control.keys() for control in control_traits))
+         for sample in primary_trait.keys())
+        if present[1])
+    control_vals_vars: tuple = reduce(
+        lambda acc, x: (acc[0] + (x[0],), acc[1] + (x[1],)),
+        ((item["value"], item["variance"])
+         for sublist in [tuple(control.values()) for control in control_traits]
+         for item in sublist),
+        (tuple(), tuple()))
+    return (
+        primary_samples,
+        tuple(primary_trait[sample]["value"] for sample in primary_samples),
+        control_vals_vars[0],
+        tuple(primary_trait[sample]["variance"] for sample in primary_samples),
+        control_vals_vars[1])
diff --git a/tests/unit/test_partial_correlations.py b/tests/unit/test_partial_correlations.py
index 6302f74..7631a71 100644
--- a/tests/unit/test_partial_correlations.py
+++ b/tests/unit/test_partial_correlations.py
@@ -1,7 +1,10 @@
 """Module contains tests for gn3.partial_correlations"""
 
 from unittest import TestCase
-from gn3.partial_correlations import control_samples, dictify_by_samples
+from gn3.partial_correlations import (
+    fix_samples,
+    control_samples,
+    dictify_by_samples)
 
 sampleslist = ["B6cC3-1", "BXD1", "BXD12", "BXD16", "BXD19", "BXD2"]
 control_traits = (
@@ -69,6 +72,21 @@ control_traits = (
                 "sample_name": "BXD2", "value": 7.80944, "variance": None,
                 "ndata": None}}})
 
+dictified_control_samples = (
+    {"B6cC3-1": {"sample_name": "B6cC3-1", "value": 7.51879, "variance": None},
+     "BXD1": {"sample_name": "BXD1", "value": 7.77141, "variance": None},
+     "BXD12": {"sample_name": "BXD12", "value": 8.39265, "variance": None},
+     "BXD16": {"sample_name": "BXD16", "value": 8.17443, "variance": None},
+     "BXD19": {"sample_name": "BXD19", "value": 8.30401, "variance": None},
+     "BXD2": {"sample_name": "BXD2", "value": 7.80944, "variance": None}},
+    {"BXD12": {"sample_name": "BXD12", "value": 8.39265, "variance": None},
+     "BXD16": {"sample_name": "BXD16", "value": 8.17443, "variance": None},
+     "BXD19": {"sample_name": "BXD19", "value": 8.30401, "variance": None},
+     "BXD2": {"sample_name": "BXD2", "value": 7.80944, "variance": None}},
+    {"B6cC3-1": {"sample_name": "B6cC3-1", "value": 7.51879, "variance": None},
+     "BXD1": {"sample_name": "BXD1", "value": 7.77141, "variance": None},
+     "BXD2": {"sample_name": "BXD2", "value":  7.80944, "variance": None}})
+
 class TestPartialCorrelations(TestCase):
     """Class for testing partial correlations computation functions"""
 
@@ -112,16 +130,22 @@ class TestPartialCorrelations(TestCase):
                  ((None, None, None, None, None, None), (None, None, None, None),
                   (None, None, None)),
                  (6, 4, 3))),
-            ({"B6cC3-1": {"sample_name": "B6cC3-1", "value": 7.51879, "variance": None},
-              "BXD1": {"sample_name": "BXD1", "value": 7.77141, "variance": None},
-              "BXD12": {"sample_name": "BXD12", "value": 8.39265, "variance": None},
-              "BXD16": {"sample_name": "BXD16", "value": 8.17443, "variance": None},
-              "BXD19": {"sample_name": "BXD19", "value": 8.30401, "variance": None},
-              "BXD2": {"sample_name": "BXD2", "value": 7.80944, "variance": None}},
-             {"BXD12": {"sample_name": "BXD12", "value": 8.39265, "variance": None},
-              "BXD16": {"sample_name": "BXD16", "value": 8.17443, "variance": None},
-              "BXD19": {"sample_name": "BXD19", "value": 8.30401, "variance": None},
-              "BXD2": {"sample_name": "BXD2", "value": 7.80944, "variance": None}},
-             {"B6cC3-1": {"sample_name": "B6cC3-1", "value": 7.51879, "variance": None},
-              "BXD1": {"sample_name": "BXD1", "value": 7.77141, "variance": None},
-              "BXD2": {"sample_name": "BXD2", "value":  7.80944, "variance": None}}))
+            dictified_control_samples)
+
+    def test_fix_samples(self):
+        """Test that fix_samples fixes the values"""
+        self.assertEqual(
+            fix_samples(
+                {"B6cC3-1": {"sample_name": "B6cC3-1", "value": 7.51879,
+                             "variance": None},
+                 "BXD1": {"sample_name": "BXD1", "value": 7.77141,
+                          "variance": None},
+                 "BXD2": {"sample_name": "BXD2", "value":  7.80944,
+                          "variance": None}},
+                dictified_control_samples),
+            (("BXD2",), (7.80944,),
+             (7.51879, 7.77141, 8.39265, 8.17443, 8.30401, 7.80944, 8.39265,
+              8.17443, 8.30401, 7.80944, 7.51879, 7.77141, 7.80944),
+             (None,),
+             (None, None, None, None, None, None, None, None, None, None, None,
+              None, None)))