change stdev to confidence intervals in plot, clean up parse script

workspace/parse.py

@@ -3,6 +3,8 @@
 import glob, os, re, sys
 import numpy as np
 import matplotlib.pyplot as plt
+import scipy as sp
+import scipy.stats as ss
 
 log_info = []
 table = {}
@@ -16,6 +18,14 @@ ntrials = int(sys.argv[2])
 
 snr_list = []
 
+def mean_confidence_interval(data, confidence=0.95):
+	a = 1.0 * np.array(data)
+	n = len(a)
+	m, se = np.mean(a), ss.sem(a)
+	h = se * ss.t._ppf((1 + confidence)/2., n-1)
+	return m, h
+
+
 for filename in glob.glob(os.path.join(path, "*.log")):
 	snr_arg = re.search(r"s([-+]?\d+.\d+)", filename)  	# SNR
 	nfiles_arg = re.search(r"n(\d+)", filename)			# Nfiles
@@ -40,29 +50,26 @@ for filename in glob.glob(os.path.join(path, "*.log")):
 		table[encode] = {}
 	if snr not in table[encode]:
 		table[encode][snr] = []
-	table[encode][snr].append(pdrate)  # could sum here but want to check
+	table[encode][snr].append(pdrate)  
-										# we got enough datapoints
 
-for enc in table.keys(): 
+snr_list.sort()
+
+for enc in sorted(table.keys()):  # sort so legend labels are in order
+	# Check first that we got all datapoints for each trial
 	for snr in table[enc].keys():
-		print table[enc][snr]
 		try:
 			assert len(table[enc][snr]) == ntrials
 		except AssertionError:
 			print "You don't have enough datapoints; exiting!"
 			sys.exit()
 
-snr_list.sort()
+	y = [mean_confidence_interval(table[enc][snr]) for snr in snr_list]
 
-for enc in sorted(table.keys()):  # sort so legend labels are in order
+	# print enc_table[enc], snr_list, y
-	y = [sum(table[enc][snr])/float(ntrials) for snr in snr_list]
+	plt.errorbar(snr_list, [m for m,h in y], 
-
+		yerr=[h for m,h in y], 
-	print [table[enc][snr] for snr in snr_list]
+		marker='o', 
-	yerr = [np.std(table[enc][snr]) for snr in snr_list]
+		label="%s" % enc_table[enc])
-
-	print enc, snr_list, y
-	print yerr
-	plt.errorbar(snr_list, y, yerr=yerr, marker='o', label="%s" % enc_table[enc])
 
 plt.ylabel("PDR")
 plt.xlabel("SNR")

workspace/run

@@ -1,6 +1,6 @@
 #!/bin/bash
 
-ntrials=5
+ntrials=20
 
 if [ $1 ]
 then
@@ -18,7 +18,7 @@ echo
 
 for e in {0..3}
 do
-	for s in `seq 1.0 2 16.0`
+	for s in `seq 1.0 2 9.0`
 	do
 		for t in `seq 1 1 $ntrials`
 		do