X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=tools%2Ffiologparser.py;h=8549859fbbffef4d29087d3786e87d50f3d6ad1d;hp=0574099494ddcefdac197e841e0a622fea092a71;hb=7c5489c06986f8576b387c05f1228f5e1cc89865;hpb=642483dbae99475036088286707d4e6f9089c986

diff --git a/tools/fiologparser.py b/tools/fiologparser.py
index 05740994..8549859f 100755
--- a/tools/fiologparser.py
+++ b/tools/fiologparser.py
@@ -1,4 +1,4 @@
-#!/usr/bin/python
+#!/usr/bin/python2.7
 #
 # fiologparser.py
 #
@@ -14,12 +14,15 @@
 # to see per-interval average completion latency.
 
 import argparse
+import math
 
 def parse_args():
     parser = argparse.ArgumentParser()
     parser.add_argument('-i', '--interval', required=False, type=int, default=1000, help='interval of time in seconds.')
     parser.add_argument('-d', '--divisor', required=False, type=int, default=1, help='divide the results by this value.')
     parser.add_argument('-f', '--full', dest='full', action='store_true', default=False, help='print full output.')
+    parser.add_argument('-A', '--all', dest='allstats', action='store_true', default=False, 
+                        help='print all stats for each interval.')
     parser.add_argument('-a', '--average', dest='average', action='store_true', default=False, help='print the average for each interval.')
     parser.add_argument('-s', '--sum', dest='sum', action='store_true', default=False, help='print the sum for each interval.')
     parser.add_argument("FILE", help="collectl log output files to parse", nargs="+")
@@ -42,7 +45,7 @@ def print_full(ctx, series):
     while (start < ftime):
         end = ftime if ftime < end else end
         results = [ts.get_value(start, end) for ts in series]
-        print "%s, %s" % (end, ', '.join(["%0.3f" % i for i in results]))
+        print("%s, %s" % (end, ', '.join(["%0.3f" % i for i in results])))
         start += ctx.interval
         end += ctx.interval
 
@@ -54,7 +57,7 @@ def print_sums(ctx, series):
     while (start < ftime):
         end = ftime if ftime < end else end
         results = [ts.get_value(start, end) for ts in series]
-        print "%s, %0.3f" % (end, sum(results))
+        print("%s, %0.3f" % (end, sum(results)))
         start += ctx.interval
         end += ctx.interval
 
@@ -66,10 +69,67 @@ def print_averages(ctx, series):
     while (start < ftime):
         end = ftime if ftime < end else end
         results = [ts.get_value(start, end) for ts in series]
-        print "%s, %0.3f" % (end, float(sum(results))/len(results))
+        print("%s, %0.3f" % (end, float(sum(results))/len(results)))
         start += ctx.interval
         end += ctx.interval
 
+# FIXME: this routine is computationally inefficient
+# and has O(N^2) behavior
+# it would be better to make one pass through samples
+# to segment them into a series of time intervals, and
+# then compute stats on each time interval instead.
+# to debug this routine, use
+#   # sort -n -t ',' -k 2 small.log
+# on your input.
+
+def my_extend( vlist, val ):
+    vlist.extend(val)
+    return vlist
+
+array_collapser = lambda vlist, val:  my_extend(vlist, val) 
+
+def print_all_stats(ctx, series):
+    ftime = get_ftime(series)
+    start = 0 
+    end = ctx.interval
+    print('start-time, samples, min, avg, median, 90%, 95%, 99%, max')
+    while (start < ftime):  # for each time interval
+        end = ftime if ftime < end else end
+        sample_arrays = [ s.get_samples(start, end) for s in series ]
+        samplevalue_arrays = []
+        for sample_array in sample_arrays:
+            samplevalue_arrays.append( 
+                [ sample.value for sample in sample_array ] )
+        # collapse list of lists of sample values into list of sample values
+        samplevalues = reduce( array_collapser, samplevalue_arrays, [] )
+        # compute all stats and print them
+        mymin = min(samplevalues)
+        myavg = sum(samplevalues) / float(len(samplevalues))
+        mymedian = median(samplevalues)
+        my90th = percentile(samplevalues, 0.90) 
+        my95th = percentile(samplevalues, 0.95)
+        my99th = percentile(samplevalues, 0.99)
+        mymax = max(samplevalues)
+        print( '%f, %d, %f, %f, %f, %f, %f, %f, %f' % (
+            start, len(samplevalues), 
+            mymin, myavg, mymedian, my90th, my95th, my99th, mymax))
+
+        # advance to next interval
+        start += ctx.interval
+        end += ctx.interval
+
+def median(values):
+    s=sorted(values)
+    return float(s[(len(s)-1)/2]+s[(len(s)/2)])/2
+
+def percentile(values, p):
+    s = sorted(values)
+    k = (len(s)-1) * p
+    f = math.floor(k)
+    c = math.ceil(k)
+    if f == c:
+        return s[int(k)]
+    return (s[int(f)] * (c-k)) + (s[int(c)] * (k-f))
 
 def print_default(ctx, series):
     ftime = get_ftime(series)
@@ -87,11 +147,11 @@ def print_default(ctx, series):
         end += ctx.interval
 
     total = 0
-    for i in xrange(0, len(averages)):
+    for i in range(0, len(averages)):
         total += averages[i]*weights[i]
-    print '%0.3f' % (total/sum(weights))
+    print('%0.3f' % (total/sum(weights)))
  
-class TimeSeries():
+class TimeSeries(object):
     def __init__(self, ctx, fn):
         self.ctx = ctx
         self.last = None 
@@ -112,13 +172,20 @@ class TimeSeries():
             self.last = sample
         self.samples.append(sample)
 
+    def get_samples(self, start, end):
+        sample_list = []
+        for s in self.samples:
+            if s.start >= start and s.end <= end:
+                sample_list.append(s)
+        return sample_list
+
     def get_value(self, start, end):
         value = 0
         for sample in self.samples:
             value += sample.get_contribution(start, end)
         return value
 
-class Sample():
+class Sample(object):
     def __init__(self, ctx, start, end, value):
        self.ctx = ctx
        self.start = start
@@ -147,6 +214,7 @@ if __name__ == '__main__':
         print_averages(ctx, series)
     elif ctx.full:
         print_full(ctx, series)
+    elif ctx.allstats:
+        print_all_stats(ctx, series)
     else:
         print_default(ctx, series)
-