perf-to-html.py - render JSON try perf jobs in a pleasing way.

perf-to-html.py - render JSON try perf jobs in a pleasing way.

Convert a perf trybot JSON file into a pleasing HTML page. It can read
from standard input or via the --filename option. Examples:

  cat results.json | perf-to-html.py --title "ia32 results"
  perf-to-html.py -f results.json -t "ia32 results" -o results.html

Options:
  -h, --help            show this help message and exit
  -f FILENAME, --filename=FILENAME
                        Specifies the filename for the JSON results rather
                        than reading from stdin.
  -t TITLE, --title=TITLE
                        Optional title of the web page.
  -o OUTPUT, --output=OUTPUT
                        Write html output to this file rather than stdout.

R=machenbach@chromium.org
BUG=

Committed: https://crrev.com/e7c2bd1a0ee7a6c6045e14648babca0df4c6d7fc
Cr-Commit-Position: refs/heads/master@{#27496}

[mvstanton, 2015-03-24 13:32:07]

Hi Michael,
Here is the program we discussed. If it can be added to the perf try job that
would be splendid! Thanks for the look,
--Michael

[caitp (gmail), 2015-03-24 13:34:48]

sort of unrelated, but it would be nice to check in baseline perf test results
automatically, so that there was something to compare with

[mvstanton, 2015-03-25 09:48:50]

Hi Michael,
Per our discussion, runs are now retrieved from the JSON. PTAL, thanks!
--Michael

[Michael Achenbach, 2015-03-25 14:49:43]

At some point this should have some refactorings that makes it more pythonish -
and it'd probably benefit from a real html generator. Lets keep that for follow
ups...

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py
File tools/perf-to-html.py (right):

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:16: import os
nit: sort

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:43: if opts.output <> None:
if opts.output:
  ...

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:45: text_file = open(opts.output, "w")
with open(opts.output, "w") as text_file:
  text_file.write(print_output)

No closing necessary

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:86: def PercentColor(change_percent, flakyness):
This function seems to be unused.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:92: else:
Why return here? Isn't bold black a valid and useful result?

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:112: if test_name in v8_benchmarks:
Please make this decision based on the units property at every trace (aka line
item) in the json output. unit == "score" means bigger is better else smaller is
better.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:178: runs = int(test_data["runs"])
Here you could also access test_data["unit"].

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:377: if opts.filename <> None:
if opts.filename:

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:378: json_data = open(opts.filename)
with open(opts.filename) as json_data:
  data = json.load(json_data)

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:385: if opts.title <> None:
if opts.title:

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:387: elif opts.filename <> None:
elif opts.filename:

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:398: help="Specifies the filename for the JSON results "\
No need for \ - just align the strings e.g.:
parser.add_option("-f", "--filename", dest="filename",
                  help="bla"
                       "bla")

[mvstanton, 2015-03-25 15:37:49]

Hi Michael, here is an update, thx for the good review.
I'll handle sorting in a seperate checkin.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py
File tools/perf-to-html.py (right):

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:16: import os
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> nit: sort

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:43: if opts.output <> None:
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> if opts.output:
>   ...

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:45: text_file = open(opts.output, "w")
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> with open(opts.output, "w") as text_file:
>   text_file.write(print_output)
> 
> No closing necessary

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:86: def PercentColor(change_percent, flakyness):
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> This function seems to be unused.

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:92: else:
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> Why return here? Isn't bold black a valid and useful result?

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:112: if test_name in v8_benchmarks:
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> Please make this decision based on the units property at every trace (aka line
> item) in the json output. unit == "score" means bigger is better else smaller
is
> better.

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:178: runs = int(test_data["runs"])
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> Here you could also access test_data["unit"].

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:377: if opts.filename <> None:
On 2015/03/25 14:49:43, Michael Achenbach (travelling) wrote:
> if opts.filename:

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:378: json_data = open(opts.filename)
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> with open(opts.filename) as json_data:
>   data = json.load(json_data)

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:385: if opts.title <> None:
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> if opts.title:

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:387: elif opts.filename <> None:
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> elif opts.filename:

Done.

https://codereview.chromium.org/1033603004/diff/20001/tools/perf-to-html.py#n...
tools/perf-to-html.py:398: help="Specifies the filename for the JSON results "\
On 2015/03/25 14:49:42, Michael Achenbach (travelling) wrote:
> No need for \ - just align the strings e.g.:
> parser.add_option("-f", "--filename", dest="filename",
>                   help="bla"
>                        "bla")

Done.

[Michael Achenbach, 2015-03-27 09:24:43]

LGMT! Thanks for working on this!

This file still invites to do refactorings, which can land later...

[mvstanton, 2015-03-27 09:29:18]

The CQ bit was checked by mvstanton@chromium.org

[commit-bot: I haz the power, 2015-03-27 09:29:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1033603004/40001

[commit-bot: I haz the power, 2015-03-27 09:29:27]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-03-27 09:29:28]

No LGTM from a valid reviewer yet. Only full committers are accepted.
Even if an LGTM may have been provided, it was from a non-committer,
_not_ a full super star committer.
See http://www.chromium.org/getting-involved/become-a-committer
Note that this has nothing to do with OWNERS files.

[Jakob Kummerow, 2015-03-27 09:38:16]

On 2015/03/27 09:24:43, Michael Achenbach (travelling) wrote:
> LGMT! Thanks for working on this!

s/LGMT/LGTM/ ;-)

[mvstanton, 2015-03-27 09:38:50]

The CQ bit was checked by mvstanton@chromium.org

[commit-bot: I haz the power, 2015-03-27 09:38:57]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1033603004/40001

[commit-bot: I haz the power, 2015-03-27 09:52:24]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-03-27 09:52:39]

Patchset 3 (id:??) landed as
https://crrev.com/e7c2bd1a0ee7a6c6045e14648babca0df4c6d7fc
Cr-Commit-Position: refs/heads/master@{#27496}

[Michael Achenbach, 2015-03-31 10:17:27]

On 2015/03/27 09:52:39, I haz the power (commit-bot) wrote:
> Patchset 3 (id:??) landed as
> https://crrev.com/e7c2bd1a0ee7a6c6045e14648babca0df4c6d7fc
> Cr-Commit-Position: refs/heads/master@{#27496}

Looks Good! Me Too! ;)