From 6a4cdb30026e6e81823f5749d820a23f2dcaf2bb Mon Sep 17 00:00:00 2001
From: pswain <peter.swain@ed.ac.uk>
Date: Fri, 15 Dec 2023 17:18:34 +0000
Subject: [PATCH] feature(dataloader): get sub_df to return a smaller data
frame
---
dataloader.py | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)
diff --git a/dataloader.py b/dataloader.py
index b3c4525..2b789f7 100644
--- a/dataloader.py
+++ b/dataloader.py
@@ -509,6 +509,30 @@ class dataloader:
"""
return self.df.pivot(y, x, signal)
+ def sub_df(self, signal, duration_threshold):
+ """
+ Find a sub dataframe of dataloader's main dataframe.
+
+ Parameters
+ ----------
+ duration_threshold: float
+ Specifies the fraction of the total duration of the time-lapse
+ experiment for which a cell must be present in a trap.
+ """
+ if duration_threshold < 0 or duration_threshold > 1:
+ print(
+ f"The threshold must be a fraction, not {duration_threshold}."
+ )
+ return
+ else:
+ wdf = self.wide_df(signal)
+ keep = (
+ wdf.notna().sum(axis=1) > duration_threshold * wdf.columns.size
+ ).values
+ ids_to_keep = list(wdf.index[keep])
+ sdf = self.df[self.df.id.isin(ids_to_keep)]
+ return sdf
+
def get_time_series(self, signal, group=None):
"""
Extract a signal as a 2D array with each row a time series.
--
GitLab