From 3e6606d082cb9896591123f3fa7a9c80921436e6 Mon Sep 17 00:00:00 2001
From: Adrian Lienhard <adrian@cmsbox.com>
Date: Tue, 31 May 2011 11:56:28 +0200
Subject: [PATCH] Added support for proxy server and for speeding up the
 replay. Refactored and cleaned up the code and added option parsing. Added a
 readme file.

---
 README.md            | 24 +++++++++++
 apache-log-replay.py | 97 +++++++++++++++++++++++++++++++-------------
 2 files changed, 92 insertions(+), 29 deletions(-)
 create mode 100644 README.md

diff --git a/README.md b/README.md
new file mode 100644
index 0000000..ba76760
--- /dev/null
+++ b/README.md
@@ -0,0 +1,24 @@
+# Script to replay HTTP requests from an Apache access logfile
+
+Features
+
+- Takes the time between requests into account
+- Replaying can be sped up by a given factor
+- Optionally send all requests to a selected (proxy) server
+
+## Installation
+
+Requires Python >= 2.6
+
+Simply download the file and execute it...
+
+## Usage
+
+    Usage: apache-log-replay.py [options] logfile
+    
+    Options:
+      -h, --help            show this help message and exit
+      -p PROXY, --proxy=PROXY
+                            send requests to server PROXY
+      -s SPEEDUP, --speedup=SPEEDUP
+                            make time run faster by factor SPEEDUP
\ No newline at end of file
diff --git a/apache-log-replay.py b/apache-log-replay.py
index e55bb1d..888b1ec 100755
--- a/apache-log-replay.py
+++ b/apache-log-replay.py
@@ -1,49 +1,88 @@
-#!/usr/bin/env python2.5
-from datetime import datetime
+"""Replay requests from an HTTP access log file.
+
+- Takes time between requests into account, with option to speed up the replay.
+- Allows one to send all requests to a selected server (proxy).
+"""
+
 import sys
 import time
-from urllib2 import urlopen
+import urllib2
+from datetime import datetime
+from optparse import OptionParser
 
-def main(host, filename):
-    logfile = open(filename, "r")
-    requests = []
-    for line in logfile:
-        parts = line.split(" ")
-        if len(parts) != 24:
-            continue
-        time_text = parts[3][1:]
-        request_time = datetime.strptime(time_text, "%d/%b/%Y:%H:%M:%S")
-        path = parts[6]
-        requests.append((request_time, path))
-    if not requests:
-        print "Seems like I don't know how to parse this file!"
-        return
+# Constants that specify access log format (indices
+# specify position after splitting on spaces)
+HOST_INDEX = 0
+TIME_INDEX = 3
+PATH_INDEX = 6
+
+def main(filename, proxy, speedup=1):
+    """Setup and start replaying."""    
+    requests = _parse_logfile(filename)
+    _setup_http_client(proxy)
+    _replay(requests, speedup)
 
+def _replay(requests, speedup):
+    """Replay the requests passed as argument"""
     total_delta = requests[-1][0] - requests[0][0]
     print "%d requests to go (time: %s)" % (len(requests), total_delta)
     last_time = requests[0][0]
-    for request_time, path in requests:
-        time_delta = request_time - last_time
+    for request_time, host, path in requests:
+        time_delta = (request_time - last_time) // speedup
         if time_delta:
-            if time_delta.seconds > 10:
+            if time_delta and time_delta.seconds > 10:
                 print "(next request in %d seconds)" % time_delta.seconds
             time.sleep(time_delta.seconds)
         last_time = request_time
-        url = host + path
+        url = "http://" + host + path
         try:
             req_result = "OK"
-            urlopen(url)
+            urllib2.urlopen(url)
         except Exception:
             req_result = "FAILED"
         print ("[%s] REQUEST: %s -- %s"
-               % (request_time.strftime("%H:%M:%S"), url, req_result))
+            % (request_time.strftime("%H:%M:%S"), url, req_result))
 
-def usage():
-    print "%s <host> <logfile>" % sys.argv[0]
+def _setup_http_client(proxy):
+    """Configure proxy server and install HTTP opener"""
+    proxy_config = {'http': proxy} if proxy else {}
+    proxy_handler = urllib2.ProxyHandler(proxy_config)
+    opener = urllib2.build_opener(proxy_handler)
+    urllib2.install_opener(opener)
 
+def _parse_logfile(filename):
+    """Parse the logfile and return a list with tuples of the form
+    (<request time>, <requested host>, <requested url>)
+    """
+    logfile = open(filename, "r")
+    requests = []
+    for line in logfile:
+        parts = line.split(" ")
+        time_text = parts[TIME_INDEX][1:]
+        request_time = datetime.strptime(time_text, "%d/%b/%Y:%H:%M:%S")
+        host = parts[HOST_INDEX]
+        path = parts[PATH_INDEX]
+        requests.append((request_time, host, path))
+    if not requests:
+        print "Seems like I don't know how to parse this file!"
+    return requests
+        
 if __name__ == "__main__":
-    if not len(sys.argv) == 3:
-        usage()
+    """Parse command line options."""
+    usage = "usage: %prog [options] logfile"
+    parser = OptionParser(usage)
+    parser.add_option('-p', '--proxy',
+        help='send requests to server PROXY',
+        dest='proxy',
+        default=None)
+    parser.add_option('-s', '--speedup',
+        help='make time run faster by factor SPEEDUP',
+        dest='speedup',
+        type='int',
+        default=1)
+    (options, args) = parser.parse_args()
+    if len(args) == 1:
+        main(args[0], options.proxy, options.speedup)
     else:
-        main(sys.argv[1], sys.argv[2])
-
+        parser.error("incorrect number of arguments")
+        
\ No newline at end of file