These are chat archives for ushahidi/Community

6th
Dec 2015
Inian
@inian
Dec 06 2015 03:37
Hey I am getting the following error when following the vagrant installation method
==> default: Notice: /Stage[main]/Main/Exec[bin-update]/returns: !!! Migration tool phinx not installed, unable to run migrations
==> default: Error: /var/www/bin/update --no-interaction returned 1 instead of one of [0]
==> default: Error: /Stage[main]/Main/Exec[bin-update]/returns: change from notrun to 0 failed: /var/www/bin/update --no-interaction returned 1 instead of one of [0]
any idea why?
Inian
@inian
Dec 06 2015 04:46
I managed to finish the install by sshing into the server
But now, I find that a request to /smssync does not return a JSON object as shown here http://smssync.ushahidi.com/developers/
it just returns an empty array
shouldnt the object with payload key be returned for a simple get request to the sync url?
Daniel Hnyk
@hnykda
Dec 06 2015 08:09
Hey, sorry, had to get some sleep here in Europe... Glad you solved the phinx thing, it took us some time too. I am afraid I can't help you with smssync, we haven't been using it.
himukin
@himukin
Dec 06 2015 09:40
Thanks @hnykda. I am also working with Inian & Arvind. could you please share the info how to make twitter data source working in v3?
Daniel Hnyk
@hnykda
Dec 06 2015 11:18
You need to allow that in Settings -> Data sources. There you use your tokens which you can generate from creating new Twitter app (there you specify the aggregator hashtag). Then you need to run bin/ushahidi dataprovider incoming and it should fetch the corresponding feeds.
I recommend to add this as a cronjob.
bvrprasad
@bvrprasad
Dec 06 2015 11:29
Hi Daniel ... Thanks for the hint on the cronjob. What would be the recommended setting? And will we be missing any tweets if we run it as cronjob?
Daniel Hnyk
@hnykda
Dec 06 2015 11:33
Very welcome. You mean the time for a cronjob? Ushahidi developer set that to 5 minutes, but it is only four "playground" environment. Personally I would say that it is reasonable choice even for production. But if you have a good machine, you could probably set it even lower interval (~2min?). And you are not going to miss any tweets - it will fetch all new tweets. Caution: it doesn't fetch already posted tweets. Only those which have been posted from the running the script for the last time.
bvrprasad
@bvrprasad
Dec 06 2015 11:36
great!
is there a way to import older tweets?
if we fetch it through a different twitter app API, can we import using some tool?
Daniel Hnyk
@hnykda
Dec 06 2015 12:10
API definitely allows there. But V3 doesn't have full documentation yet. But some guys reverse-engineered that API for posting, so it shouldn't be hard to write some little script for that.
Daniel Hnyk
@hnykda
Dec 06 2015 12:44
@bvrprasad They have built simple twitter processing app for ushahidi v3 in node-red (get tweet -> submit as a post). In case you could use node-red, it should be piece of cake to edit it for processing some batch work of tweets. I could ask them to share the project, it's not out publicly yet.
Daniel Hnyk
@hnykda
Dec 06 2015 17:30
@bvrprasad @himukin @inian One more thing. On current OpenHack 2015 we were working on some preprocessing and structuring data from twitter, email and others. It's far from being perfect and mergeable to the master, but it can handle structuring for example some specified format or sending text for analysis to arbitrary server providing parsing/analysing/ML HTTP API. It could work as a dirty and quick temporary solution, if you don't have any yet.
himukin
@himukin
Dec 06 2015 20:11
@hnykda , that is very interesting. could you please share the github/open source url for this preprocessing?
Daniel Hnyk
@hnykda
Dec 06 2015 22:58
We moved our conversation to the different channel called openhackexp (public) so I don't pollute it here. If anyone is interested, just join there.