You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Ben Mahler <be...@gmail.com> on 2014/01/14 08:08:35 UTC
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/
-----------------------------------------------------------
(Updated Jan. 14, 2014, 7:08 a.m.)
Review request for mesos, Benjamin Hindman and Jie Yu.
Changes
-------
Rebased.
I'm considering simplifying the TimeSeries in this change since it's potentially solving a problem that we don't have. If others think it is too complex then I'll update it to be a simple circular buffer style TimeSeries (i.e. remove sparsification). However, it does have one nice property in that if the window capacity is not exceeded, then the current TimeSeries performs no sparsification.
Repository: mesos-git
Description
-------
The first part of this change is the removal of the archive operation.
This also exposes a TimeSeries data structure used by Statistics.
The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
See the documentation on TimeSeries for details.
An alternative, more deterministic approach, is to use sampling on older values.
Both approaches have their advantages and disadvantages.
Diffs (updated)
-----
3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
Diff: https://reviews.apache.org/r/13603/diff/
Testing
-------
Added tests to verify the truncation and sparsification of time series.
Thanks,
Ben Mahler
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Ben Mahler <be...@gmail.com>.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/include/process/statistics.hpp, line 5
> > <https://reviews.apache.org/r/13603/diff/2/?file=422661#file422661line5>
> >
> > Is this for std::pair? Maybe a comment?
Removed this include since I no longer need it.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/src/statistics.cpp, line 229
> > <https://reviews.apache.org/r/13603/diff/2/?file=422662#file422662line229>
> >
> > Why not statistics[context][name] = ?
Ah, an artifact of when I had TimeSeries as a const struct, which prevented assignability.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/include/process/statistics.hpp, lines 126-127
> > <https://reviews.apache.org/r/13603/diff/2/?file=422661#file422661line126>
> >
> > What about just doing 'capacity(std::max(3, _capacity))' above instead? I don't think keeping around another 2 instances of T will likely be that big of a deal.
Sounds good.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/include/process/statistics.hpp, line 135
> > <https://reviews.apache.org/r/13603/diff/2/?file=422661#file422661line135>
> >
> > s/lg/log/
Ah, I thought lg was standard for log base 2, but I guess not: http://en.wikipedia.org/wiki/Logarithm#Particular_bases
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/src/statistics.cpp, lines 45-47
> > <https://reviews.apache.org/r/13603/diff/2/?file=422662#file422662line45>
> >
> > Formatting?
Reverted my alignment, I think we should allow some table-like formatting when it makes things easier to read, but this isn't a great example.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/include/process/statistics.hpp, line 165
> > <https://reviews.apache.org/r/13603/diff/2/?file=422661#file422661line165>
> >
> > What about a TimeSeries::Value struct and returning that instead of a pair in order to capture the semantics explicitly?
I like that, I've also updated to return a vector<Value> as opposed to a map<Time, T> which was exposing an implantation detail.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/include/process/statistics.hpp, line 214
> > <https://reviews.apache.org/r/13603/diff/2/?file=422661#file422661line214>
> >
> > It would be great to have some more comments and maybe even some CHECKs that capture your expected semantics re: 'next'. For example, the first time this is called you expect 'index' to be none and that's when you set 'next'. Likewise, after a truncation it might be the case that 'next' is invalid and you need to reset it. A comment down by 'Option<size_t> index;' would be good too (saying something like it starts none until the first sparsify call).
I've added a comment on for 'index' and 'next' and I cleaned up sparsify() further, there's now only one place where we deal with resetting 'index'.
> On Jan. 15, 2014, 1:31 a.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/src/tests/statistics_tests.cpp, line 73
> > <https://reviews.apache.org/r/13603/diff/2/?file=422663#file422663line73>
> >
> > By a "round of truncation", do you mean sparsification?
> >
> > IIUC, I would expect that you can add less things in order for sparsification to occur. Perhaps you're being more general here? But to capture the algorithm I think that you could add strictly less.
> >
> > Also, given that some truncation can occur after some time, should you pause the clock for this test?
I've cleaned up the tests significantly, let me know if you find them more clear:
1. TimeSeries tests are now in a separate file.
2. The sparsification tests and truncation tests now use a List construction to compare values more explicitly.
- Ben
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/#review31835
-----------------------------------------------------------
On Jan. 14, 2014, 7:08 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13603/
> -----------------------------------------------------------
>
> (Updated Jan. 14, 2014, 7:08 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Jie Yu.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The first part of this change is the removal of the archive operation.
>
> This also exposes a TimeSeries data structure used by Statistics.
>
> The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
> TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
> See the documentation on TimeSeries for details.
>
> An alternative, more deterministic approach, is to use sampling on older values.
> Both approaches have their advantages and disadvantages.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
> 3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
> 3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
>
> Diff: https://reviews.apache.org/r/13603/diff/
>
>
> Testing
> -------
>
> Added tests to verify the truncation and sparsification of time series.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Benjamin Hindman <be...@berkeley.edu>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/#review31835
-----------------------------------------------------------
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60552>
Is this for std::pair? Maybe a comment?
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60553>
I don't mind the formatting but it's not standard as far as I know.
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60550>
Any reason not to return a Future<TimeSeries<double>>?
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60551>
Let's move TimeSeries into it's own header.
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60555>
Not necessary (and not standard).
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60556>
What about just doing 'capacity(std::max(3, _capacity))' above instead? I don't think keeping around another 2 instances of T will likely be that big of a deal.
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60554>
Any reason not to use const& for the parameters?
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60557>
s/lg/log/
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60558>
What about a TimeSeries::Value struct and returning that instead of a pair in order to capture the semantics explicitly?
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60560>
The comment around the technique makes sense but the implementation could use some comments too, i.e., a description of how you expect the implementation to capture the notion of "halving" would be great. You could use an example of a time series getting bigger but not going over capacity (the while loop conditional) as well as a time series getting bigger and reaching the halfway point.
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60559>
It would be great to have some more comments and maybe even some CHECKs that capture your expected semantics re: 'next'. For example, the first time this is called you expect 'index' to be none and that's when you set 'next'. Likewise, after a truncation it might be the case that 'next' is invalid and you need to reset it. A comment down by 'Option<size_t> index;' would be good too (saying something like it starts none until the first sparsify call).
3rdparty/libprocess/include/process/statistics.hpp
<https://reviews.apache.org/r/13603/#comment60567>
The use of 'truncate' here is not great. You're not truncating, you're sparsifing here.
3rdparty/libprocess/src/statistics.cpp
<https://reviews.apache.org/r/13603/#comment60561>
Formatting?
3rdparty/libprocess/src/statistics.cpp
<https://reviews.apache.org/r/13603/#comment60562>
Why not statistics[context][name] = ?
3rdparty/libprocess/src/statistics.cpp
<https://reviews.apache.org/r/13603/#comment60563>
const&?
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment60566>
By a "round of truncation", do you mean sparsification?
IIUC, I would expect that you can add less things in order for sparsification to occur. Perhaps you're being more general here? But to capture the algorithm I think that you could add strictly less.
Also, given that some truncation can occur after some time, should you pause the clock for this test?
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment60568>
s/time-series/time series/
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment60569>
Can you test that it's sparsified at this point then? And again, it's sparsification since the clock is paused correct?
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment60565>
s/time-series/time series/
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment60571>
s/time-series/time series/
- Benjamin Hindman
On Jan. 14, 2014, 7:08 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13603/
> -----------------------------------------------------------
>
> (Updated Jan. 14, 2014, 7:08 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Jie Yu.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The first part of this change is the removal of the archive operation.
>
> This also exposes a TimeSeries data structure used by Statistics.
>
> The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
> TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
> See the documentation on TimeSeries for details.
>
> An alternative, more deterministic approach, is to use sampling on older values.
> Both approaches have their advantages and disadvantages.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
> 3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
> 3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
>
> Diff: https://reviews.apache.org/r/13603/diff/
>
>
> Testing
> -------
>
> Added tests to verify the truncation and sparsification of time series.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/
-----------------------------------------------------------
(Updated Jan. 22, 2014, 9:40 p.m.)
Review request for mesos, Benjamin Hindman and Vinod Kone.
Changes
-------
Rebase and small cleanup from benh review, NNFR.
Repository: mesos-git
Description
-------
The first part of this change is the removal of the archive operation.
This also exposes a TimeSeries data structure used by Statistics.
The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
See the documentation on TimeSeries for details.
An alternative, more deterministic approach, is to use sampling on older values.
Both approaches have their advantages and disadvantages.
Diffs (updated)
-----
3rdparty/libprocess/Makefile.am 47e6d02c29fc4485e5d32bc6adaea62c82c1d60e
3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
3rdparty/libprocess/include/process/timeseries.hpp PRE-CREATION
3rdparty/libprocess/src/process.cpp 936f1181dc9e75c1d8faca60c905cfa01c87b77f
3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
3rdparty/libprocess/src/tests/timeseries_tests.cpp PRE-CREATION
Diff: https://reviews.apache.org/r/13603/diff/
Testing
-------
Added tests to verify the truncation and sparsification of time series.
Thanks,
Ben Mahler
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Ben Mahler <be...@gmail.com>.
> On Jan. 22, 2014, 6:36 p.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/src/process.cpp, line 594
> > <https://reviews.apache.org/r/13603/diff/3/?file=425934#file425934line594>
> >
> > Maybe add a comment about where these come from and/or a TODO to eventually move them? Also, do you think they're better here than statistics.cpp?
Added a comment and a TODO. Moved them to statistics.cpp since that makes more sense in retrospect.
> On Jan. 22, 2014, 6:36 p.m., Benjamin Hindman wrote:
> > 3rdparty/libprocess/src/tests/timeseries_tests.cpp, line 13
> > <https://reviews.apache.org/r/13603/diff/3/?file=425937#file425937line13>
> >
> > Okay, so maybe a templated overload? ;)
There were compiler issues with TimeSeries<T>::Value that I was unable to resolve so I've left these as non-templated.
- Ben
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/#review32507
-----------------------------------------------------------
On Jan. 17, 2014, 3:58 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13603/
> -----------------------------------------------------------
>
> (Updated Jan. 17, 2014, 3:58 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Jie Yu.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The first part of this change is the removal of the archive operation.
>
> This also exposes a TimeSeries data structure used by Statistics.
>
> The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
> TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
> See the documentation on TimeSeries for details.
>
> An alternative, more deterministic approach, is to use sampling on older values.
> Both approaches have their advantages and disadvantages.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/Makefile.am 47e6d02c29fc4485e5d32bc6adaea62c82c1d60e
> 3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
> 3rdparty/libprocess/include/process/timeseries.hpp PRE-CREATION
> 3rdparty/libprocess/src/process.cpp 67f7f9b3b05c8bc9b0f6281689223996ddfa68d1
> 3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
> 3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
> 3rdparty/libprocess/src/tests/timeseries_tests.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/13603/diff/
>
>
> Testing
> -------
>
> Added tests to verify the truncation and sparsification of time series.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Benjamin Hindman <be...@berkeley.edu>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/#review32507
-----------------------------------------------------------
Ship it!
3rdparty/libprocess/src/process.cpp
<https://reviews.apache.org/r/13603/#comment61368>
Maybe add a comment about where these come from and/or a TODO to eventually move them? Also, do you think they're better here than statistics.cpp?
3rdparty/libprocess/src/tests/statistics_tests.cpp
<https://reviews.apache.org/r/13603/#comment61376>
What about just overloading the == operator instead?
3rdparty/libprocess/src/tests/timeseries_tests.cpp
<https://reviews.apache.org/r/13603/#comment61377>
Okay, so maybe a templated overload? ;)
- Benjamin Hindman
On Jan. 17, 2014, 3:58 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13603/
> -----------------------------------------------------------
>
> (Updated Jan. 17, 2014, 3:58 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Jie Yu.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The first part of this change is the removal of the archive operation.
>
> This also exposes a TimeSeries data structure used by Statistics.
>
> The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
> TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
> See the documentation on TimeSeries for details.
>
> An alternative, more deterministic approach, is to use sampling on older values.
> Both approaches have their advantages and disadvantages.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/Makefile.am 47e6d02c29fc4485e5d32bc6adaea62c82c1d60e
> 3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
> 3rdparty/libprocess/include/process/timeseries.hpp PRE-CREATION
> 3rdparty/libprocess/src/process.cpp 67f7f9b3b05c8bc9b0f6281689223996ddfa68d1
> 3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
> 3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
> 3rdparty/libprocess/src/tests/timeseries_tests.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/13603/diff/
>
>
> Testing
> -------
>
> Added tests to verify the truncation and sparsification of time series.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Jie Yu <yu...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/#review32415
-----------------------------------------------------------
Ship it!
Ship It!
- Jie Yu
On Jan. 17, 2014, 3:58 a.m., Ben Mahler wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/13603/
> -----------------------------------------------------------
>
> (Updated Jan. 17, 2014, 3:58 a.m.)
>
>
> Review request for mesos, Benjamin Hindman and Jie Yu.
>
>
> Repository: mesos-git
>
>
> Description
> -------
>
> The first part of this change is the removal of the archive operation.
>
> This also exposes a TimeSeries data structure used by Statistics.
>
> The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
> TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
> See the documentation on TimeSeries for details.
>
> An alternative, more deterministic approach, is to use sampling on older values.
> Both approaches have their advantages and disadvantages.
>
>
> Diffs
> -----
>
> 3rdparty/libprocess/Makefile.am 47e6d02c29fc4485e5d32bc6adaea62c82c1d60e
> 3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
> 3rdparty/libprocess/include/process/timeseries.hpp PRE-CREATION
> 3rdparty/libprocess/src/process.cpp 67f7f9b3b05c8bc9b0f6281689223996ddfa68d1
> 3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
> 3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
> 3rdparty/libprocess/src/tests/timeseries_tests.cpp PRE-CREATION
>
> Diff: https://reviews.apache.org/r/13603/diff/
>
>
> Testing
> -------
>
> Added tests to verify the truncation and sparsification of time series.
>
>
> Thanks,
>
> Ben Mahler
>
>
Re: Review Request 13603: Exposed TimeSeries and added "sparsification" of
older values.
Posted by Ben Mahler <be...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/13603/
-----------------------------------------------------------
(Updated Jan. 17, 2014, 3:58 a.m.)
Review request for mesos, Benjamin Hindman and Jie Yu.
Changes
-------
-TimeSeries is now in its own header.
-The TimeSeries code has been cleaned up in a few places and documentation has been improved substantially.
-TimeSeries tests are now much simpler as well!
Repository: mesos-git
Description
-------
The first part of this change is the removal of the archive operation.
This also exposes a TimeSeries data structure used by Statistics.
The goal here is to provide reasonable historical information for high frequency time series that rapidly exceed capacity.
TimeSeries allows a window of values to be stored, while "sparsifying" older values when the capacity is exceeded.
See the documentation on TimeSeries for details.
An alternative, more deterministic approach, is to use sampling on older values.
Both approaches have their advantages and disadvantages.
Diffs (updated)
-----
3rdparty/libprocess/Makefile.am 47e6d02c29fc4485e5d32bc6adaea62c82c1d60e
3rdparty/libprocess/include/process/statistics.hpp ce122a5eaa1aa9c09207cc8c9428136b681561cf
3rdparty/libprocess/include/process/timeseries.hpp PRE-CREATION
3rdparty/libprocess/src/process.cpp 67f7f9b3b05c8bc9b0f6281689223996ddfa68d1
3rdparty/libprocess/src/statistics.cpp d4ba9f146f7b9b46525a0e27fbfb3d61a21a94fc
3rdparty/libprocess/src/tests/statistics_tests.cpp e6c9a1b776d6f3339f96898c3501d2a00c416006
3rdparty/libprocess/src/tests/timeseries_tests.cpp PRE-CREATION
Diff: https://reviews.apache.org/r/13603/diff/
Testing
-------
Added tests to verify the truncation and sparsification of time series.
Thanks,
Ben Mahler