How to get started?

[rocksboxB]

Just wondering if there could be any more instructions on how to get started? Would be very helpful for ruby newbies like me.

Once you build the Recommendify gem, where do you go from there? Do you have to write a new model, like in the example?

Trying to copy that code into a model file gives the error "uninitialized constant Recommendify" - does not recognize it as a gem/module. Tips? Thanks! (p.s. running recommendify 0.2.3 on windows)

[rocksboxB]

Hey, after a day or so of mucking around with this, I'd recommend adding the following things to the ReadMe file:

1. run "gem install recommendify". If you are running windows, you should write "gem install recommendify -v 0.2.3" (the last working version of recommendify that installs on windows)
2. Add the line "'recommendify', '0.2.3'" to your gemfile
3. Do a bundle install

BUT I'm also pretty sure Recommendify cannot run on windows
because it also needs the gems
"gem redis"
"gem hiredis" - which is incompatible with windows; see http://stackoverflow.com/questions/7172577/hiredis-on-windows-xp-sp3 or http://stackoverflow.com/questions/8234493/error-when-bundle-install-on-windows/8235541#8235541

So yea adding something like "Does not work on Windows because of Hiredis" and linking to those stackoverflow posts might be useful.

Super bummed - was totally looking forward to using it!

[asmuth]

Hey,

thanks a lot for putting so much effort into this issue and taking the time to write it all down :)

Luckily, this should be quite easy to fix: I will release an update to make this work on windows in the next 24 hours. (We can just check if hiredis is available at runtime and fall back to the pure-ruby implementation if it is not.)

The pure ruby version works fine for medium sized data sets (up to a few hundred thousand recommendable items). If you have a large data set (more than 1 million recommendable items) I would highly recommend running recommendify with the included C extension on one or more Linux/BSD VMs if that is possible for you.

I'll update this issue tomorrow when the fix is released.

cheers 🍻
~ paul

[rocksboxB]

Wow, thanks, Paul!

This is awesome - and thanks for being so responsive/looking into this. I
(and the other windows-devs out there) really appreciate it!

-Boris

On Fri, Feb 22, 2013 at 11:08 AM, Paul Asmuth notifications@github.comwrote:

Hey,

thanks a lot for putting so much effort into this issue and taking the
time to write it all down :)

Luckily, this should be quite easy to fix: I will release an update to
make this work on windows in the next 24 hours. (We can just check if
hiredis is available at runtime and fall back to the pure-ruby
implementation if it is not.)

The pure ruby version works fine for medium sized data sets (up to a few
hundred thousand recommendable items). If you have a large data set (more
than 1 million recommendable items) I would highly recommend running
recommendify with the included C extension on one or more Linux/BSD VMs if
that is possible for you.

I'll update this issue tomorrow when the fix is released.

cheers [image: 🍻]
~ paul

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/23#issuecomment-13965588.

[fmoudoute]

Hi, did any body get a chance to find a tutorial on how to use this gem on Mac ? I have been playing around with it, but do not seem to get anywhere. I am at stage zero after installing the gem. I found a few articles where people seem to have figured out how to use it, but I am not that advanced on rails yet. Would appreciate some help . Where to put the class myRecommander <recommendify::base? Where to add the recommender = Myrecommender.ner? where to add the set creation please?
Thank you