+wiki output, +caching, +faster pickling

1 files changed, 59 insertions, 10 deletions

diff --git a/src/Backtrace/abrt-bz-dupchecker b/src/Backtrace/abrt-bz-dupchecker
index 344d1326..cbdafc53 100755
--- a/src/Backtrace/abrt-bz-dupchecker
+++ b/src/Backtrace/abrt-bz-dupchecker
@@ -24,6 +24,7 @@ from optparse import OptionParser
 import sys
 import os.path
 import subprocess
+import cPickle
 
 parser = OptionParser(version="%prog 1.0")
 parser.add_option("-u", "--user", dest="user",
@@ -34,6 +35,8 @@ parser.add_option("-b", "--bugzilla", dest="bugzilla",
                   help="Bugzilla URL (defaults to Red Hat Bugzilla)", metavar="URL")
 parser.add_option("-v", "--verbose", dest="verbose",
                   help="Detailed output")
+parser.add_option("-i", "--wiki", help="Generate output in wiki syntax", 
+                  action="store_true", default=False, dest="wiki")
 
 (options, args) = parser.parse_args()
 
@@ -54,9 +57,38 @@ buginfos = bz.query({'status_whiteboard_type':'allwordssubstr','status_whiteboar
 
 print "{0} bugs found.".format(len(buginfos))
 
+#
+# Load cache from previous run. Speeds up the case Bugzilla closes connection.
+# The cache should be manually removed after a day or so, because the data in it 
+# are no longer valid.
+#
 database = {}
-
+ids = {}
+CACHE_FILE = "abrt-bz-dupchecker-cache.tmp"
+if os.path.isfile(CACHE_FILE):
+  f = open(CACHE_FILE, 'r')
+  database = cPickle.load(f)
+  ids = cPickle.load(f)
+  f.close()
+
+def save_to_cache():
+  global database
+  f = open(CACHE_FILE, 'w')
+  cPickle.dump(database, f, 2)
+  cPickle.dump(ids, f, 2)
+  f.close()  
+
+count = 0
 for buginfo in buginfos:
+  count += 1
+  print "{0}/{1}".format(count, len(buginfos))
+  if count % 100 == 0:
+    save_to_cache()
+
+  if ids.has_key(buginfo.bug_id):
+    continue
+  ids[buginfo.bug_id] = True
+
   if not buginfo.bug_status in ["NEW", "ASSIGNED", "MODIFIED", "VERIFIED"]:
     if options.verbose:
       print "Bug {0} has status {1}, skipping.".format(buginfo.bug_id, buginfo.bug_status)
@@ -127,21 +159,38 @@ print "=========================================================================
 
 # The number of duplicates.
 dupcount = 0
+# The number of duplicates that can be closed.
+dupclosecount = 0
 for backtrace, components in database.items():
   for component, bugitems in components.items():
-    if len(bugitems) > 1:
-      dupcount += len(bugitems) - 1
-
+    dupcount += len(bugitems) - 1
+    dupclosecount += min(len(filter(lambda x: x <= 2, 
+                                    map(lambda x: x["comments"], 
+                                        bugitems))),
+                         len(bugitems) - 1)
+    
 print "Total number of duplicate bugs detected: {0}".format(dupcount)
+print "Number of duplicate bugs that will be closed : {0}".format(dupclosecount)
 print "------------------------------"
 
 # Print the duplicates
 for backtrace, components in database.items():
   for component, bugitems in components.items():
     if len(bugitems) > 1:
-      print "Component: {0}".format(component)
-      print "Duplicates: {0}".format(
-        reduce(lambda x,y: x+", "+y, 
-               map(lambda x: "{0} ({1})".format(x['id'],x['comments']), 
-                   bugitems)))
-      print "Backtrace: {0}".format(backtrace)
+      if options.wiki:
+        print "----"
+        print "* component: '''{0}'''".format(component)
+        print "* duplicates: {0}".format(
+          reduce(lambda x,y: x+", "+y, 
+                 map(lambda x: "#[https://bugzilla.redhat.com/show_bug.cgi?id={0} {0}] ({1} comments)".format(x['id'],x['comments']), 
+                     bugitems)))
+        print "* backtrace:"
+        for line in backtrace.replace("Thread\n", "").splitlines():
+          print "*# {0}".format(line)
+      else:
+        print "Component: {0}".format(component)
+        print "Duplicates: {0}".format(
+          reduce(lambda x,y: x+", "+y, 
+                 map(lambda x: "{0} ({1})".format(x['id'],x['comments']), 
+                     bugitems)))
+        print "Backtrace: {0}".format(backtrace)