Skip to content
Projects
Groups
Snippets
Help
Loading...
Help
Support
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Z
Zope
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
0
Issues
0
List
Boards
Labels
Milestones
Merge Requests
0
Merge Requests
0
Analytics
Analytics
Repository
Value Stream
Wiki
Wiki
Snippets
Snippets
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Commits
Issue Boards
Open sidebar
Kirill Smelkov
Zope
Commits
84ae8cea
Commit
84ae8cea
authored
Aug 01, 2010
by
Hanno Schlichting
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
Use some named tuples to avoid hard to understand v[2][1][0] constructs
parent
41935174
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
39 additions
and
39 deletions
+39
-39
src/Products/ZCatalog/plan.py
src/Products/ZCatalog/plan.py
+39
-39
No files found.
src/Products/ZCatalog/plan.py
View file @
84ae8cea
...
...
@@ -12,6 +12,7 @@
##############################################################################
import
time
from
collections
import
namedtuple
from
thread
import
allocate_lock
from
Acquisition
import
aq_base
...
...
@@ -105,6 +106,12 @@ def make_key(catalog, query):
return
key
Duration
=
namedtuple
(
'Duration'
,
[
'start'
,
'end'
])
IndexMeasurement
=
namedtuple
(
'IndexMeasurement'
,
[
'name'
,
'duration'
,
'num'
])
Benchmark
=
namedtuple
(
'Benchmark'
,
[
'num'
,
'duration'
,
'hits'
])
class
CatalogPlan
(
object
):
"""Catalog plan class to measure and identify catalog queries and plan
their execution.
...
...
@@ -153,42 +160,44 @@ class CatalogPlan(object):
if
not
benchmark
:
return
None
# sort indexes on (mean hits, mean search time)
ranking
=
[((
v
[
0
],
v
[
1
]),
k
)
for
k
,
v
in
benchmark
.
items
()]
# sort indexes on (mean result length, mean search time)
ranking
=
[((
value
.
num
,
value
.
duration
),
name
)
for
name
,
value
in
benchmark
.
items
()]
ranking
.
sort
()
return
[
i
[
1
]
for
i
in
ranking
]
return
[
r
[
1
]
for
r
in
ranking
]
def
start
(
self
):
self
.
init_timer
()
self
.
start_time
=
time
.
time
()
def
start_split
(
self
,
label
,
result
=
None
):
self
.
interim
[
label
]
=
(
time
.
time
(),
None
)
def
start_split
(
self
,
name
,
result
=
None
):
self
.
interim
[
name
]
=
Duration
(
time
.
time
(),
None
)
def
stop_split
(
self
,
name
,
result
=
None
):
current
=
time
.
time
()
start_time
,
stop_time
=
self
.
interim
.
get
(
name
,
(
None
,
None
))
start_time
,
stop_time
=
self
.
interim
.
get
(
name
,
Duration
(
None
,
None
))
length
=
0
if
result
is
not
None
:
# TODO: calculating the length can be expensive
length
=
len
(
result
)
self
.
interim
[
name
]
=
(
start_time
,
current
)
self
.
interim
[
name
]
=
Duration
(
start_time
,
current
)
dt
=
current
-
start_time
self
.
res
.
append
((
name
,
current
-
start_time
,
length
))
self
.
res
.
append
(
IndexMeasurement
(
name
=
name
,
duration
=
current
-
start_time
,
num
=
length
))
# remember index's hits, search time and calls
benchmark
=
self
.
benchmark
()
if
name
not
in
benchmark
:
benchmark
[
name
]
=
(
length
,
dt
,
1
)
benchmark
[
name
]
=
Benchmark
(
num
=
length
,
duration
=
dt
,
hits
=
1
)
else
:
n
,
t
,
c
=
benchmark
[
name
]
n
=
int
(((
n
*
c
)
+
length
)
/
float
(
c
+
1
))
t
=
((
t
*
c
)
+
dt
)
/
float
(
c
+
1
)
n
um
,
duration
,
hits
=
benchmark
[
name
]
n
um
=
int
(((
num
*
hits
)
+
length
)
/
float
(
hits
+
1
))
duration
=
((
duration
*
hits
)
+
dt
)
/
float
(
hits
+
1
)
# reset adaption
if
c
%
REFRESH_RATE
==
0
:
c
=
0
c
+=
1
benchmark
[
name
]
=
(
n
,
t
,
c
)
if
hits
%
REFRESH_RATE
==
0
:
hits
=
0
hits
+=
1
benchmark
[
name
]
=
Benchmark
(
num
,
duration
,
hits
)
def
stop
(
self
):
self
.
end_time
=
time
.
time
()
...
...
@@ -206,25 +215,24 @@ class CatalogPlan(object):
if
key
not
in
stats
:
mt
=
self
.
duration
c
=
1
hits
=
1
else
:
mt
,
c
=
stats
[
key
]
mt
=
((
mt
*
c
)
+
self
.
duration
)
/
float
(
c
+
1
)
c
+=
1
mt
,
hits
=
stats
[
key
]
mt
=
((
mt
*
hits
)
+
self
.
duration
)
/
float
(
hits
+
1
)
hits
+=
1
stats
[
key
]
=
(
mt
,
c
)
stats
[
key
]
=
(
mt
,
hits
)
self
.
log
()
def
result
(
self
):
return
(
self
.
duration
,
tuple
(
self
.
res
))
def
log
(
self
):
# result of stopwatch
res
=
self
.
result
()
if
res
[
0
]
<
self
.
threshold
:
total
=
self
.
duration
if
total
<
self
.
threshold
:
return
key
=
self
.
key
res
=
(
total
,
self
.
res
)
reports_lock
.
acquire
()
try
:
if
self
.
cid
not
in
reports
:
...
...
@@ -233,11 +241,10 @@ class CatalogPlan(object):
previous
=
reports
[
self
.
cid
].
get
(
key
)
if
previous
:
counter
,
mean
,
last
=
previous
mean
=
(
mean
*
counter
+
res
[
0
]
)
/
float
(
counter
+
1
)
mean
=
(
mean
*
counter
+
total
)
/
float
(
counter
+
1
)
reports
[
self
.
cid
][
key
]
=
(
counter
+
1
,
mean
,
res
)
else
:
reports
[
self
.
cid
][
key
]
=
(
1
,
res
[
0
],
res
)
reports
[
self
.
cid
][
key
]
=
(
1
,
total
,
res
)
finally
:
reports_lock
.
release
()
...
...
@@ -264,13 +271,6 @@ class CatalogPlan(object):
'details': <duration of single indexes>,
}
<duration of single indexes> := [{'id': <index_name1>,
'duration': <duration>,
'length': <resultset length>,
},
...
]
The duration is provided in millisecond.
"""
rval
=
[]
...
...
@@ -280,9 +280,9 @@ class CatalogPlan(object):
'counter'
:
v
[
0
],
'duration'
:
v
[
1
]
*
1000
,
'last'
:
{
'duration'
:
v
[
2
][
0
]
*
1000
,
'details'
:
[
dict
(
id
=
i
[
0
]
,
duration
=
i
[
1
]
*
1000
,
length
=
i
[
2
]
)
'details'
:
[
dict
(
id
=
i
.
name
,
duration
=
i
.
duration
*
1000
,
length
=
i
.
num
)
for
i
in
v
[
2
][
1
]],
},
}
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment