Recent changes to python_interface

python_interface modified by André Offringa

André Offringa — Thu, 12 Sep 2019 07:47:28 -0000

--- v5
+++ v6
@@ -1,4 +1,4 @@
-The Python interface uses Boost Python to wrap the C++ functionality. Standard codedoc is available, so use e.g. help(aoflagger.ImageSet) to get help on using the ImageSet class.
+The Python interface uses Boost Python to wrap the C++ functionality. Standard codedoc is available, so use e.g. help(aoflagger.ImageSet) to get help on using the ImageSet class. See the [installation instructions](installation_instructions) on how to get the Python interface working.

 This is an example program:
 ~~~

python_interface modified by André Offringa

André Offringa — Mon, 20 May 2019 13:59:52 -0000

--- v4
+++ v5
@@ -1,4 +1,4 @@
-The Python interface uses Boost Python to wrap the C++ functionality.
+The Python interface uses Boost Python to wrap the C++ functionality. Standard codedoc is available, so use e.g. help(aoflagger.ImageSet) to get help on using the ImageSet class.

 This is an example program:
 ~~~

python_interface modified by André Offringa

André Offringa — Mon, 20 May 2019 13:56:47 -0000

--- v3
+++ v4
@@ -49,6 +49,7 @@
 except:
     print("write_statistics() failed")
 ~~~
+The above program demonstrates both how to make an 'imageset' (the data for a single baseline) and run the flagger on this set, as well as how to extract the quality statistics from it.

 Here's an example that calculates the false-positives ratio of a flagging strategy on Gaussian data:
 ~~~

python_interface modified by André Offringa

André Offringa — Mon, 20 May 2019 13:53:19 -0000

--- v2
+++ v3
@@ -4,7 +4,6 @@
 ~~~
 import aoflagger as aof
 import numpy
-import sys
 print("Flagging with AOFlagger version " + aof.AOFlagger.get_version_string())

 nch = 256
@@ -50,3 +49,41 @@
 except:
     print("write_statistics() failed")
 ~~~
+
+Here's an example that calculates the false-positives ratio of a flagging strategy on Gaussian data:
+~~~
+import aoflagger as aof
+import numpy
+import sys
+
+nch = 256
+ntimes = 1000
+count=50       # number of trials in the false-positives test
+
+aoflagger = aof.AOFlagger()
+strategy = aoflagger.load_strategy("example-strategy.rfis")
+data = aoflagger.make_image_set(ntimes, nch, 8)
+
+ratiosum = 0.0
+ratiosumsq = 0.0
+for repeat in range(count):
+    for imgindex in range(8):
+        # Initialize data with random numbers
+        values = numpy.random.normal(0, 1, [ntimes, nch])
+        data.set_image_buffer(imgindex, values)
+        
+    flags = aoflagger.run(strategy, data)
+    flagvalues = flags.get_buffer()
+    ratio = float(sum(sum(flagvalues))) / (nch*ntimes)
+    ratiosum += ratio
+    ratiosumsq += ratio*ratio
+    sys.stdout.write('.')
+    sys.stdout.flush()
+
+print('')
+    
+print("Percentage flags (false-positive rate) on Gaussian data: " +
+      str(ratiosum * 100.0 / count) + "% +/- " +
+      str(numpy.sqrt((ratiosumsq/count - ratiosum*ratiosum / (count*count))) * 100.0)
+     )
+~~~

python_interface modified by André Offringa

André Offringa — Mon, 20 May 2019 13:50:29 -0000

--- v1
+++ v2
@@ -9,7 +9,6 @@

 nch = 256
 ntimes = 1000
-count=50       # number of trials in the false-positives test

 aoflagger = aof.AOFlagger()

python_interface modified by André Offringa

André Offringa — Mon, 20 May 2019 13:50:05 -0000

The Python interface uses Boost Python to wrap the C++ functionality.

This is an example program:

import aoflagger as aof
import numpy
import sys
print("Flagging with AOFlagger version " + aof.AOFlagger.get_version_string())

nch = 256
ntimes = 1000
count=50       # number of trials in the false-positives test

aoflagger = aof.AOFlagger()

# Load strategy from disk (alternatively use 'make_strategy' to use a default one)
strategy = aoflagger.load_strategy("example-strategy.rfis")

data = aoflagger.make_image_set(ntimes, nch, 8)

print("Number of times: " + str(data.width()))
print("Number of channels: " + str(data.height()))

# When flagging multiple baselines, iterate over the baselines and
# call the following code for each baseline
# (to use multithreading, make sure to create an imageset for each
# thread)

# Make eight images: real and imaginary for 4 pol

for imgindex in range(8):
    # Initialize data
    values = numpy.zeros([ntimes, nch])
    data.set_image_buffer(imgindex, values)

flags = aoflagger.run(strategy, data)
flagvalues = flags.get_buffer()
flagcount = sum(sum(flagvalues))
print("Percentage flags on zero data: " + str(flagcount * 100.0 / (nch*ntimes)) + "%")

# Collect statistics
# We create some unrealistic time and frequency arrays to be able
# to run these functions. Normally, these should hold the time
# and frequency values.
timeArray = numpy.linspace(0.0, ntimes, num=ntimes, endpoint=False)
freqArray = numpy.linspace(0.0, nch, num=nch, endpoint=False)
qs = aoflagger.make_quality_statistics(timeArray, freqArray, 4, False)
aoflagger.collect_statistics(qs, data, flags, aoflagger.make_flag_mask(ntimes, nch, False), 0, 1)
try:
    aoflagger.write_statistics(qs, "test.qs")
except:
    print("write_statistics() failed")