Nick Doiron / Jun 07 2019
Remix of Python by Nextjournal

IAEA dataset parsing

Check Python version

import sys; sys.version.split()[0]
'3.6.8'

From training data 00: preprocessed, low signal-to-noise ratio, lowest energy level.

training00_E0.txt

Description of training data format:

from math import floor
lines = open(
training00_E0.txt
, 'r').read().split("\n") index = 0 sensorCount = 0 sensorAngles = 0 sensors = [] for line in lines: if len(line) == 0: continue if index == 0: sensorCount = int(line) elif index == 1: sensorAngles = int(line) elif index == 2: sensorOffset = float(line) else: angleNum = floor(((index - 3) / 2) / sensorCount) sensorNum = int(((index - 3) / 2) - angleNum * sensorCount) print(str(angleNum) + ',' + str(sensorNum)) if angleNum == 0: sensors.append([]) sensors[sensorNum].append(float(line)) # sensors are interleaved; skip 2nd head on this reader index = index + 1 index = index + 1
import matplotlib.pyplot as plt

# Data for plotting
fig, ax = plt.subplots()
x = range(1, 361)
ax.plot(x, sensors[0])
ax.plot(x, sensors[10])
ax.plot(x, sensors[20])
ax.plot(x, sensors[30])

fig

Those sensors are close together. Let's look at the full sweep

fig, ax = plt.subplots()

ax.plot(x, sensors[30])
ax.plot(x, sensors[70])
ax.plot(x, sensors[90])
ax.plot(x, sensors[120])
ax.plot(x, sensors[180])

fig