The answer is... kind of. Definitely an area that needs work. - We hav some sample jupyter notebooks, but haven't figured out how to get them into the website rather than them being stand-alone pages. I think they might technically be reachable, but there's no way to navigate around the site once you get to one. And I think one or two of them need to be updated a bit from changes in this release. - The first part of each python unit test file basically mirrors the jupyter notebooks, including lots of tutorial-style comments. Those might be a better starting point, since they're kept up to date by the unit tests needing to pass. (There are sometimes a few other tests just to make sure all the python methods are executed.)
Would it be reasonable/sufficient to add a line to the python README.md suggesting a look at the unit tests? Thanks, jon On Thu, May 14, 2020 at 6:24 PM Furkan KAMACI <[email protected]> wrote: > Hi Lee, > > I was wondering if there are any quick start/test examples for the > project. I saw that there is an example of KLL Sketch for C++: [1] but I > couldn't see something similar for Python. > > i.e. at Druid, one can ingest Wikipedia dataset [2] [3] and so: > > 1) Users can understand how it works > > 2) Committers can quickly check whether new changes have broken something > or not. > > This is just a wish from a view of a user. It may not be suitable for > Datasketches but this is similar to what I try to explain: [4] > > [1] > https://datasketches.apache.org/docs/Quantiles/QuantilesCppExample.html > [2] https://druid.apache.org/docs/latest/tutorials/index.html > [3] https://github.com/apache/druid/tree/master/examples > [4] https://github.com/DataSketches/sketches-misc > > Kind Regards, > Furkan KAMACI > > On Fri, May 15, 2020 at 3:10 AM leerho <[email protected]> wrote: > >> Hi Furkan, >> >> Thanks for the vote! >> >> Could you please elaborate on what you mean by "test the distribution" ? >> >> Cheers, >> >> Lee. >> >> On Thu, May 14, 2020 at 4:39 PM Furkan KAMACI <[email protected]> >> wrote: >> >>> Hi, >>> >>> +1 from me. >>> >>> I checked: >>> >>> - Incubating in name >>> - DISCLAIMER-WIP exists >>> - LICENSE and NOTICE are fine >>> - Checked PGP signatures >>> - Checked Checksums >>> - cmake, make, make test for cpp >>> - python wrapper build (tried with python3) >>> >>> PS 1: Does website includes examples for Python wrapper? >>> PS 2: It could be nice if you can place some steps to test the >>> distribution. >>> >>> Kind Regards, >>> Furkan KAMACI >>> >>> On Fri, May 15, 2020 at 1:08 AM leerho <[email protected]> wrote: >>> >>>> +1 >>>> Checked: >>>> - gpg signature >>>> - shasum >>>> - cmake, make, make test >>>> - Confirmed Licenses in LICENSE file are still relevant. >>>> - NOTICE and DISCLAIMER-WIP look good. >>>> - minor formatting problems in README.md: >>>> The file has some tabs, which should be replaced by 4-spaces. >>>> Line 25 needs 4 leading spaces not 2; this messes up the rendering >>>> of that section. >>>> On Thu, May 14, 2020 at 1:35 PM Alexander Saydakov >>>> <[email protected]> wrote: >>>> >>>>> +1 >>>>> Checked: >>>>> - gpg signature >>>>> - shasum >>>>> - cmake, make, make test >>>>> - python wrapper build: had to specify python3 explicitly and sudo >>>>> install >>>>> - tried using CPC sketch in python environment >>>>> >>>>> >>>>> On Wed, May 13, 2020 at 11:58 PM Jon Malkin <[email protected]> >>>>> wrote: >>>>> >>>>>> ### Major version: Fully allocator-aware, header only. Varopt sampling >>>>>> added. Exception-safe. API changes for consistency. ### >>>>>> >>>>>> Hello Apache DataSketches PPMC and Community, >>>>>> >>>>>> 1. This is a call for vote to release Apache DataSketches-cpp version: >>>>>> 2.0.0-incubating-rc2 >>>>>> >>>>>> NOTE 1: The C++ code provides header-only library. Compilation is >>>>>> needed only to build unit tests. A Python extension is also included in >>>>>> this repository as a set of thin wrappers. That does produce a compiled >>>>>> artifact, but is not built by default. >>>>>> >>>>>> NOTE 2: Compiling the C++ unit tests requires CMake 3.12 or newer. >>>>>> >>>>>> 2. Source repository: >>>>>> - https://github.com/apache/incubator-datasketches-cpp >>>>>> >>>>>> Git Tag for this release: >>>>>> - >>>>>> https://github.com/apache/incubator-datasketches-cpp/tree/2.0.0-incubating-rc2 >>>>>> on branch 2.0.x-incubating >>>>>> >>>>>> Git HashId for this release starts with: c51298d >>>>>> >>>>>> 3. The Release Candidate / Zip Repository: >>>>>> - >>>>>> https://dist.apache.org/repos/dist/dev/incubator/datasketches/cpp/2.0.0-incubating-rc2/ >>>>>> >>>>>> The public signing key can be found in the KEYS file: >>>>>> - https://dist.apache.org/repos/dist/dev/incubator/datasketches/KEYS >>>>>> >>>>>> The artifacts have been signed with --keyid-format SHORT : 38F334CC >>>>>> >>>>>> 4. C++ Test Guide: >>>>>> - >>>>>> https://github.com/apache/incubator-datasketches-cpp/blob/2.0.0-incubating-rc2/README.md >>>>>> 5. Python Build & Test Guide: >>>>>> - >>>>>> https://github.com/apache/incubator-datasketches-cpp/blob/2.0.0-incubating-rc2/python/README.md >>>>>> >>>>>> 6. The vote will be performed in two stages: >>>>>> - This letter will be published on dev@ and remain open for at least >>>>>> 72 hours and at least 3 (+1) PPMC votes or a majority of (+1) are >>>>>> acquired. All PPMC members including Mentors can vote. However, a >>>>>> negative vote from a Mentor will cancel this voting process. >>>>>> >>>>>> - After it passes the first stage, the summary of that vote and the >>>>>> key information from this letter will be published on general@incubator >>>>>> and remain open for at least 72 hours and at least 3 (+1) IPMC votes or >>>>>> a majority of (+1) are acquired. >>>>>> >>>>>> Please vote accordingly: >>>>>> >>>>>> [ ] +1 approve >>>>>> [ ] +0 no opinion >>>>>> [ ] -1 disapprove with the reason >>>>>> Thanks, >>>>>> Jon [email protected] <http://jmalkinapache.org> >>>>>> >>>>>>
