aboutsummaryrefslogtreecommitdiff
path: root/gn3/computations/correlations2.py
diff options
context:
space:
mode:
authorMuriithi Frederick Muriuki2021-07-26 15:39:42 +0300
committerMuriithi Frederick Muriuki2021-07-26 15:39:42 +0300
commit4c8c13814a22fe6b40081ecfa1f957bc5bf99930 (patch)
tree685b7b0e4ca2b86e13f43235320e80d273a092c8 /gn3/computations/correlations2.py
parent278f1a18b59b1cadd04c50a9af35b5aece4d722d (diff)
downloadgenenetwork3-4c8c13814a22fe6b40081ecfa1f957bc5bf99930.tar.gz
Fix issues caught by pylint
Issue: https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi * Fix a myriad of issues caught by pylint to ensure the code passes all tests.
Diffstat (limited to 'gn3/computations/correlations2.py')
-rw-r--r--gn3/computations/correlations2.py37
1 files changed, 26 insertions, 11 deletions
diff --git a/gn3/computations/correlations2.py b/gn3/computations/correlations2.py
index 6c456db..93db3fa 100644
--- a/gn3/computations/correlations2.py
+++ b/gn3/computations/correlations2.py
@@ -1,15 +1,25 @@
+"""
+DESCRIPTION:
+ TODO: Add a description for the module
+
+FUNCTIONS:
+compute_correlation:
+ TODO: Describe what the function does..."""
+
from math import sqrt
from functools import reduce
## From GN1: mostly for clustering and heatmap generation
-def items_with_values(dbdata, userdata):
+def __items_with_values(dbdata, userdata):
"""Retains only corresponding items in the data items that are not `None` values.
-This should probably be renamed to something sensible"""
+ This should probably be renamed to something sensible"""
def both_not_none(item1, item2):
+ """Check that both items are not the value `None`."""
if (item1 is not None) and (item2 is not None):
return (item1, item2)
return None
def split_lists(accumulator, item):
+ """Separate the 'x' and 'y' items."""
return [accumulator[0] + [item[0]], accumulator[1] + [item[1]]]
return reduce(
split_lists,
@@ -17,19 +27,24 @@ This should probably be renamed to something sensible"""
[[], []])
def compute_correlation(dbdata, userdata):
- x, y = items_with_values(dbdata, userdata)
- if len(x) < 6:
- return (0.0, len(x))
- meanx = sum(x)/len(x)
- meany = sum(y)/len(y)
+ """Compute some form of correlation.
+
+ This is extracted from
+ https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/utility/webqtlUtil.py#L622-L647
+ """
+ x_items, y_items = __items_with_values(dbdata, userdata)
+ if len(x_items) < 6:
+ return (0.0, len(x_items))
+ meanx = sum(x_items)/len(x_items)
+ meany = sum(y_items)/len(y_items)
def cal_corr_vals(acc, item):
xitem, yitem = item
return [
acc[0] + ((xitem - meanx) * (yitem - meany)),
acc[1] + ((xitem - meanx) * (xitem - meanx)),
acc[2] + ((yitem - meany) * (yitem - meany))]
- xyd, sxd, syd = reduce(cal_corr_vals, zip(x, y), [0.0, 0.0, 0.0])
+ xyd, sxd, syd = reduce(cal_corr_vals, zip(x_items, y_items), [0.0, 0.0, 0.0])
try:
- return ((xyd/(sqrt(sxd)*sqrt(syd))), len(x))
- except ZeroDivisionError as zde:
- return(0, len(x))
+ return ((xyd/(sqrt(sxd)*sqrt(syd))), len(x_items))
+ except ZeroDivisionError:
+ return(0, len(x_items))