add scrappy statistics gathering scripts

[monitor.git] / statistics / nodediff-graph-better.py

diff --git a/statistics/nodediff-graph-better.py b/statistics/nodediff-graph-better.py
new file mode 100755 (executable)
index 0000000..68e14e1
--- /dev/null
+++ b/statistics/nodediff-graph-better.py
@@ -0,0 +1,138 @@
+#!/usr/bin/python
+
+
+from monitor import config
+from monitor.wrapper import plc
+from monitor import parser as parsermodule
+#from monitor.model import *
+from monitorstats import *
+from monitor import database
+
+import sys
+import time
+import calendar
+from datetime import datetime, timedelta
+
+from nodequeryold import verify,query_to_dict,node_select
+
+api = plc.getAuthAPI()
+
+def nodes_from_time(arch, file, select=None):
+       fb = arch.load(file)
+
+       nodelist = fb['nodes'].keys()
+       nodelist = node_select(select, nodelist, fb)
+       return nodelist
+
+def print_nodelist(nodelist, file):
+       for node in nodelist:
+               if file:
+                       print >>file, node
+               else:
+                       print node
+       
+
+def main():
+       parser = parsermodule.getParser()
+       parser.set_defaults(archivedir='archive-pdb', begin=None, end=None, printnodes=False, select=None, select2=None)
+       parser.add_option("", "--archivedir", dest="archivedir", metavar="filename",
+                                               help="Pickle file aggregate output.")
+       parser.add_option("", "--select", dest="select", metavar="key",
+                                               help="Select .")
+       parser.add_option("", "--select2", dest="select2", metavar="key",
+                                               help="Select .")
+       parser.add_option("", "--print", dest="printnodes", action="store_true",
+                                               help="print the nodes that have come up or down.")
+       parser.add_option("", "--begin", dest="begin", metavar="YYYY-MM-DD",
+                                               help="Specify a starting date from which to begin the query.")
+       parser.add_option("", "--end", dest="end", metavar="YYYY-MM-DD",
+                                               help="Specify a ending date at which queries end.")
+       config = parsermodule.parse_args(parser)
+       archive = get_archive(config.archivedir)
+
+       if not config.begin or not config.end:
+               print parsermodule.usage(parser)
+               sys.exit(1)
+
+       tdelta = timedelta(1)
+       d_s1 = datetime_fromstr(config.begin)
+       d_s2 = datetime_fromstr(config.begin) + tdelta
+       d_end = datetime_fromstr(config.end)
+
+       print d_s1
+       print d_s2
+       print d_end
+
+       data = []
+       while d_end > d_s2:
+
+               try:
+                       f_s1 = get_filefromglob(d_s1, "production.findbad", config.archivedir)
+                       f_s2 = get_filefromglob(d_s2, "production.findbad", config.archivedir)
+               except:
+                       timestr = d_s2.strftime("%Y-%m-%d")
+                       print timestr, ",", 0, ",", 0
+                       d_s1 = d_s2
+                       d_s2 = d_s1 + tdelta
+                       continue
+
+               s1 = set(nodes_from_time(archive, f_s1, config.select))
+               s2 = set(nodes_from_time(archive, f_s2, config.select))
+               s3 = set(nodes_from_time(archive, f_s2, config.select2))
+
+
+               timestr = d_s2.strftime("%Y-%m-%d")
+               print timestr, ",", len(s2),",",  len(s3)
+               if not config.printnodes:
+               #       f_up = open("up-%s" % timestr, 'w')
+               #       f_down = open("dn-%s" % timestr, 'w')
+                       f_up = None
+                       f_down = None
+                       pass
+               else:
+                       print "%s nodes up" % len(s2-s1)
+                       print "Nodes s2 minus s1: len(s2-s1) = %s" % len(s2-s1)
+                       f_up = None
+                       f_down = None
+
+               #print_nodelist(s2-s1, f_up)
+
+               if config.printnodes:
+                       print ""
+                       print "%s nodes down" % len(s1-s2)
+                       print "Nodes s1 minus s2: len(s1-s2) = %s" % len(s1-s2)
+
+               #print_nodelist(s1-s2, f_down)
+               if not config.printnodes:
+                       if f_up: f_up.close()
+                       if f_up: f_down.close()
+
+               d_s1 = d_s2
+               d_s2 = d_s1 + tdelta
+       
+       s1=[]
+       s2=[]
+       s3=[]
+       for row in data:
+               s1.append(row[0])
+               s2.append(row[1])
+               s3.append(row[2])
+       
+       s1 = map(lambda x: x-500, s1)
+       rlow= zip(s1,s3)
+       rhigh = zip(s1,s2)
+       diff_low  = map(lambda x: x[0]-x[1], rlow)
+       diff_high = map(lambda x: x[0]+x[1], rhigh)
+       s1 = map(lambda x: str(x), s1)
+       diff_low = map(lambda x: str(x), diff_low)
+       diff_high = map(lambda x: str(x), diff_high)
+       print s1
+       print diff_low
+       print diff_high
+       print "http://chart.apis.google.com/chart?cht=lc&chds=40,100&chxt=x,y&chxl=0:|Oct|Nov|Dec|Jan|Feb|1:|540|580|600&chs=700x200&chm=F,aaaaaa,1,-1,2&chd=t1:%s" % ",".join(s1) + "|" + ",".join(diff_low) + "|" + ",".join(s1) + "|" + ",".join(s1) +"|" + ",".join(diff_high)
+
+# takes two arguments as dates, comparing the number of up nodes from one and
+# the other.
+
+if __name__ == "__main__":
+       main()