Rev 10: Include information about the depth of leaf nodes. in http://bazaar.launchpad.net/%7Ebzr/bzr-repodetails/trunk

[John Arbash Meinel]

At http://bazaar.launchpad.net/%7Ebzr/bzr-repodetails/trunk

------------------------------------------------------------
revno: 10
revision-id: john at arbash-meinel.com-20081202014502-40sghjisv1lflwhb
parent: john at arbash-meinel.com-20081202014124-qgvuoydx8e2gzpch
committer: John Arbash Meinel <john at arbash-meinel.com>
branch nick: trunk
timestamp: Mon 2008-12-01 19:45:02 -0600
message:
  Include information about the depth of leaf nodes.
-------------- next part --------------
=== modified file 'gather_stats.py'
--- a/gather_stats.py	2008-11-19 20:55:29 +0000
+++ b/gather_stats.py	2008-12-02 01:45:02 +0000
@@ -192,9 +192,11 @@
 
 
 def _gather_chk_map(objectstats, chk_bytes, pending, internal_counter,
-                    leaf_counter):
+                    leaf_counter, depth_counter):
     done = set()
+    depth = 0
     while pending:
+        depth += 1
         # Don't visit nodes twice
         done.update(pending)
         next = pending
@@ -208,6 +210,7 @@
             else:
                 items = list(node.iteritems(chk_bytes))
                 objectstats.add_extra_counter(leaf_counter, len(items))
+                objectstats.add_extra_counter(depth_counter, depth)
         # Different routes can take a different number of steps to get to the
         # same nodes.
         pending.difference_update(done)
@@ -222,11 +225,10 @@
         pending.add(inv.id_to_entry._root_node)
         if inv.parent_id_basename_to_file_id is not None:
             pending_parent_id.add(inv.parent_id_basename_to_file_id._root_node)
-    _gather_chk_map(objectstats, repo.chk_bytes, pending, 'internal node refs',
-                    'leaf node items')
+    _gather_chk_map(objectstats, repo.chk_bytes, pending,
+                    'internal node refs', 'leaf node items', 'inv depth')
     _gather_chk_map(objectstats, repo.chk_bytes, pending_parent_id,
-                    'internal parent_id refs',
-                    'leaf parent_id items')
+                    'internal p_id refs', 'leaf p_id items', 'p_id depth')
 
 
 def _gather_stats_locked(repo):

=== modified file 'tests/test_repositorydetails.py'
--- a/tests/test_repositorydetails.py	2008-11-19 20:55:29 +0000
+++ b/tests/test_repositorydetails.py	2008-12-02 01:45:02 +0000
@@ -59,8 +59,10 @@
             "Total:              0 KiB 100%         0 KiB 100%        5\n"
             "\n"
             "Extra Info:           count    total  avg stddev  min  max\n"
+            "inv depth                 1        1    1    0.0    1    1\n"
             "leaf node items           1        2    2    0.0    2    2\n"
-            "leaf parent_id items      1        2    2    0.0    2    2\n"
+            "leaf p_id items           1        2    2    0.0    2    2\n"
+            "p_id depth                1        1    1    0.0    1    1\n"
             , out)
         self.assertEqual("", err)
 
@@ -104,7 +106,9 @@
         self.assertEqual(370, stats.inventories.raw_size)
         self.assertEqual(518, stats.inventories.compressed_size)
         self.assertEqual({'leaf node items': [1, 2, 4, 2, 2],
-                          'leaf parent_id items': [1, 2, 4, 2, 2]},
+                          'leaf p_id items': [1, 2, 4, 2, 2],
+                          'p_id depth': [1, 1, 1, 1, 1],
+                          'inv depth': [1, 1, 1, 1, 1]},
                          stats.inventories.extra_counters)
         self.assertEqual(1, stats.texts.objects)
         self.assertEqual(16, stats.texts.raw_size)