From 8fabc239e3aead7677059c5a387a042652ac3ca3 Mon Sep 17 00:00:00 2001
From: Jari Pronold <34653647+jarsi@users.noreply.github.com>
Date: Wed, 29 Apr 2020 20:55:35 +0200
Subject: [PATCH] Upadte reading in spikes from numpy arrays
With more recent versions of numpy (changed in version 1.16.3, see [here](https://numpy.org/devdocs/reference/generated/numpy.load.html)) the default behavior of reading in numpy arrays changed from allow_pickle=True to allow_pickle=False. If we want to read in the processed spikes, which are stored in numpy arrays, across numpy versions we need to manually set it to True.
In general it is stated:
> Warning
> Loading files that contain object arrays uses the pickle module, which is not secure against erroneous or maliciously constructed data. Consider passing allow_pickle=False to load data that is known not to contain object arrays for the safer handling of untrusted sources.
Thus we should about another dataformat in the long run and consider this a temporary fix.
---
multiarea_model/analysis.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/multiarea_model/analysis.py b/multiarea_model/analysis.py
index 8ac5bda..b5e7673 100644
--- a/multiarea_model/analysis.py
+++ b/multiarea_model/analysis.py
@@ -125,7 +125,7 @@ class Analysis:
fn = os.path.join(rec_dir,
'.'.join((fp, 'npy')))
try:
- data[area][pop] = np.load(fn)
+ data[area][pop] = np.load(fn, allow_pickle=True)
except FileNotFoundError:
if not hasattr(self, 'all_spikes'):
fp = '.'.join(('-'.join((self.simulation.label,
--
GitLab