All I see in the github is a SA plugin that is supposed to run a message through a "Python script"? Where is this actual python script?
If this whole concept is based on some external language script, please stop wasting time learning and developing the SpamAssassin plugin glue. Such external script can never be efficiently integrated into SA using any of the existing Bayes framework. To me it seems you are developing a standalone classifier, similar to DSpam or CRM114. SpamAssassin can then utilize this in standalone fashion, feeding complete message to it and reading the results. Such classifier must maintain it's own token database etc, which has nothing to do with SpamAssassin. If and when you have this standalone python classifier in working state, then it's trivial for some existing SpamAsasssin developer to write the simple plugin glue. Cheers, Henrik On Fri, Jun 14, 2019 at 05:08:22PM +0530, Shreyansh Shrivastava. wrote: > On Fri, Jun 14, 2019 at 4:31 PM Axb <[1][email protected]> wrote: > > While blogs may be interesting to read and keep track of stuff... > Is there any code to look at and start testing? > > Here's the Git repo of the project, [2]GSoC Spamassassin. I am still working > on the Perl plugin code.Any suggestions/help is appreciated. > > The summer is short... > > The winter has ended, not coming again this time :) > > > Axb > > On 6/14/19 10:56 AM, Shreyansh Shrivastava. wrote: > > This is the second blog of the series. It covers data cleaning, feature > > extraction, model training, and metric selection of a SVM spam > classifier. > > All of the things mentioned along with a few extra are implemented in > the > > SVM plugin that I am trying to develop. Will be covering ASF SINGA when > I > > will work on neural nets. > > Blog 2 - SVM and Spamassassin > > <[3]https://medium.com/@shreyansh25.shrivastava/ > svm-and-spamassassin-9da5c2d3fe33> > > > > Regards, > > Shreyansh Shrivastava > > > > On Mon, Jun 10, 2019 at 8:26 AM Shreyansh Shrivastava. < > > [4][email protected]> wrote: > > > >> I am a Google summer of code student working to develop a statistical > >> classifier plugin for spamassassin. I will be writing a series of blogs > >> regarding my technical and non-technical experiences during the > program, > >> and since it's open source I think it's a win-win. > >> > >> The first part is simply about how I came across Spamassassin and how I > >> got selected. Will be touching on specifics later on. Blog 1 > >> <[5]https://medium.com/@shreyansh25.shrivastava/ > spamassassin-singa-and-me-7927dbaa16d0> > >> > >> Regards, > >> Shreyansh Shrivastava > >> > > > > > > References: > > [1] mailto:[email protected] > [2] https://github.com/sjs253/Spamassassin-GSoC > [3] > https://medium.com/@shreyansh25.shrivastava/svm-and-spamassassin-9da5c2d3fe33 > [4] mailto:[email protected] > [5] > https://medium.com/@shreyansh25.shrivastava/spamassassin-singa-and-me-7927dbaa16d0
