commit 4d36a13e8b7790692e4969c8d4005a354bb249c1
Author: Sathyanarayanan Gunasekaran <[email protected]>
Date: Thu Jul 5 00:28:26 2012 +0200
Replace tabs with spaces, add a usage function
---
task-6232/.gitignore | 2 +
task-6232/pyentropy.py | 121 ++++++++++++++++++++++++-----------------------
2 files changed, 64 insertions(+), 59 deletions(-)
diff --git a/task-6232/.gitignore b/task-6232/.gitignore
index 8a0c627..15236f0 100644
--- a/task-6232/.gitignore
+++ b/task-6232/.gitignore
@@ -1,3 +1,5 @@
in/
+data/
+foo
entropy.csv
diff --git a/task-6232/pyentropy.py b/task-6232/pyentropy.py
index f13f709..226aca8 100644
--- a/task-6232/pyentropy.py
+++ b/task-6232/pyentropy.py
@@ -14,66 +14,69 @@ KEYS = ['r','s','v','w','p','m']
class Router:
- def __init__(self):
- self.lines = []
- self.nick = None
- self.bandwidth = None
- self.flags = None
- self.probability = None
-
- def add(self, key, values):
- if key == 'r':
- self.nick = values[0]
- if key == 'w':
- self.bandwidth = int(values[0].split('=')[1])
- if key == 's':
- self.flags = values
-
-
+ def __init__(self):
+ self.lines = []
+ self.nick = None
+ self.bandwidth = None
+ self.flags = None
+ self.probability = None
+
+ def add(self, key, values):
+ if key == 'r':
+ self.nick = values[0]
+ if key == 'w':
+ self.bandwidth = int(values[0].split('=')[1])
+ if key == 's':
+ self.flags = values
+
def run(file_name):
- routers = []
- # parse consensus
- with open(file_name, 'r') as f:
- for line in f.readlines():
- key = line.split()[0]
- values = line.split()[1:]
- if key =='r':
- router = Router()
- router.add(key, values)
- elif key == 'p':
- router.add(key, values)
- routers.append(router)
- elif key == 'valid-after':
- valid_after = ' '.join(values)
- elif key in KEYS:
- router.add(key, values)
-
- # build hash table with freq. distribution
- # key: bandwidth
- # value: number of bandwidth's observations
- bw_dist = {}
- for router in routers:
- if bw_dist.has_key(router.bandwidth):
- bw_dist[router.bandwidth] += 1
- else:
- bw_dist[router.bandwidth] = 1
-
- if len(routers) <= 0:
- print "Error: amount of routers must be > 0."
- return;
-
- print "calculating entropy"
- entropy = 0.0
- for bw in bw_dist.iterkeys():
- # p = probability of one particular bandwidth
- p = float(bw_dist[bw]) / len(routers)
- entropy += -(p * math.log(p, 2))
-
- return ",".join([valid_after, str(entropy)])
+ routers = []
+ # parse consensus
+ with open(file_name, 'r') as f:
+ for line in f.readlines():
+ key = line.split()[0]
+ values = line.split()[1:]
+ if key =='r':
+ router = Router()
+ router.add(key, values)
+ elif key == 'p':
+ router.add(key, values)
+ routers.append(router)
+ elif key == 'valid-after':
+ valid_after = ' '.join(values)
+ elif key in KEYS:
+ router.add(key, values)
+
+ # build hash table with freq. distribution
+ # key: bandwidth
+ # value: number of bandwidth's observations
+ bw_dist = {}
+ for router in routers:
+ if bw_dist.has_key(router.bandwidth):
+ bw_dist[router.bandwidth] += 1
+ else:
+ bw_dist[router.bandwidth] = 1
+
+ if len(routers) <= 0:
+ print "Error: amount of routers must be > 0."
+ return;
+
+ entropy = 0.0
+ for bw in bw_dist.iterkeys():
+ # p = probability of one particular bandwidth
+ p = float(bw_dist[bw]) / len(routers)
+ entropy += -(p * math.log(p, 2))
+
+ return ",".join([valid_after, str(entropy)])
+def usage():
+ print "Usage - python pyentropy.py <consensus-dir> <output-file>"
if __name__ == "__main__":
- with open(sys.argv[2], 'w') as f:
- for file_name in os.listdir(sys.argv[1]):
- string = run(os.path.join(sys.argv[1], file_name))
- f.write("%s\n" % (string))
+ if len(sys.argv) != 3:
+ usage()
+ else:
+ with open(sys.argv[2], 'w') as f:
+ for file_name in os.listdir(sys.argv[1]):
+ string = run(os.path.join(sys.argv[1], file_name))
+ f.write("%s\n" % (string))
_______________________________________________
tor-commits mailing list
[email protected]
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-commits
