diff options
-rw-r--r-- | gn3/partial_correlations.py | 38 | ||||
-rw-r--r-- | tests/unit/test_partial_correlations.py | 33 |
2 files changed, 69 insertions, 2 deletions
diff --git a/gn3/partial_correlations.py b/gn3/partial_correlations.py index c556d10..1fb0ccc 100644 --- a/gn3/partial_correlations.py +++ b/gn3/partial_correlations.py @@ -6,7 +6,7 @@ GeneNetwork1. """ from functools import reduce -from typing import Any, Sequence +from typing import Any, Tuple, Sequence def control_samples(controls: Sequence[dict], sampleslist: Sequence[str]): """ @@ -86,3 +86,39 @@ def fix_samples(primary_trait: dict, control_traits: Sequence[dict]) -> Sequence control_vals_vars[0], tuple(primary_trait[sample]["variance"] for sample in primary_samples), control_vals_vars[1]) + +def find_identical_traits( + primary_name: str, primary_value: float, control_names: Tuple[str, ...], + control_values: Tuple[float, ...]) -> Tuple[str, ...]: + """ + Find traits that have the same value when the values are considered to + 3 decimal places. + + This is a migration of the + `web.webqtl.correlation.correlationFunction.findIdenticalTraits` function in + GN1. + """ + def __merge_identicals__( + acc: Tuple[str, ...], + ident: Tuple[str, Tuple[str, ...]]) -> Tuple[str, ...]: + return acc + ident[1] + + def __dictify_controls__(acc, control_item): + ckey = "{:.3f}".format(control_item[0]) + return {**acc, ckey: acc.get(ckey, tuple()) + (control_item[1],)} + + return (reduce(## for identical control traits + __merge_identicals__, + (item for item in reduce(# type: ignore[var-annotated] + __dictify_controls__, zip(control_values, control_names), + {}).items() if len(item[1]) > 1), + tuple()) + or + reduce(## If no identical control traits, try primary and controls + __merge_identicals__, + (item for item in reduce(# type: ignore[var-annotated] + __dictify_controls__, + zip((primary_value,) + control_values, + (primary_name,) + control_names), {}).items() + if len(item[1]) > 1), + tuple())) diff --git a/tests/unit/test_partial_correlations.py b/tests/unit/test_partial_correlations.py index c591c8f..60e54c1 100644 --- a/tests/unit/test_partial_correlations.py +++ b/tests/unit/test_partial_correlations.py @@ -4,7 +4,8 @@ from unittest import TestCase from gn3.partial_correlations import ( fix_samples, control_samples, - dictify_by_samples) + dictify_by_samples, + find_identical_traits) sampleslist = ["B6cC3-1", "BXD1", "BXD12", "BXD16", "BXD19", "BXD2"] control_traits = ( @@ -178,3 +179,33 @@ class TestPartialCorrelations(TestCase): (None,), (None, None, None, None, None, None, None, None, None, None, None, None, None))) + + def test_find_identical_traits(self): + """ + Test `gn3.partial_correlations.find_identical_traits`. + + Given: + - the name of a primary trait + - the value of a primary trait + - a sequence of names of control traits + - a sequence of values of control traits + When: + - the arguments above are passed to the `find_identical_traits` + function + Then: + - Return ALL trait names that have the same value when up to three + decimal places are considered + """ + for primn, primv, contn, contv, expected in ( + ("pt", 12.98395, ("ct0", "ct1", "ct2"), + (0.1234, 2.3456, 3.4567), tuple()), + ("pt", 12.98395, ("ct0", "ct1", "ct2"), + (12.98354, 2.3456, 3.4567), ("pt", "ct0")), + ("pt", 12.98395, ("ct0", "ct1", "ct2", "ct3"), + (0.1234, 2.3456, 0.1233, 4.5678), ("ct0", "ct2")) + ): + with self.subTest( + primary_name=primn, primary_value=primv, + control_names=contn, control_values=contv): + self.assertEqual( + find_identical_traits(primn, primv, contn, contv), expected) |