Comments on: Announcing TwitterEngine: Your own twitter aggregation site on AppEngine

By: Steffen Tiedemann Christensen

Steffen Tiedemann Christensen — Fri, 25 Jun 2010 19:04:26 +0000

@BiNaRiO: I think the two parts are related — at least from the error message you’re getting:

# Fatal error when loading application configuration:
# mapping values are not allowed here
# in “mediosdecomunicacion/app.yaml”, line 33, column 13

Can you paste your app.xml here and I’ll see if I can spot the issue?

By: BiNaRiO

BiNaRiO — Fri, 25 Jun 2010 04:33:47 +0000

I get this message:
http://pastebin.com/V2ARV5Wz
Python error or configuration error?

I have another question. In your example have change twitterengine to 23tweets from first line of app.yaml, however the first seven lines of this file does not have any code or text. So I do not know to change. I suspect this may be causing the failure. I hope you can help me. Thank you.

By: jim

jim — Sun, 30 May 2010 11:14:52 +0000

no worrys got iti think. have to register your site with google app engine

By: Steffen Tiedemann Christensen

Steffen Tiedemann Christensen — Sun, 30 May 2010 10:57:20 +0000

@jim: The guide in the blog post takes you through the steps of deploying the Twittengine code to your own *.appspot.com domain -- basically, while following the steps you'll be able to choose the subdomain (and you can even point you own domain to it if you want to). The key to all of this is The Google AppEngine SDK.

By: jim

jim — Sun, 30 May 2010 10:53:26 +0000

how can i get this on my website instead on on http://www.____.appspot.com?

By: Steffen Tiedemann Christensen

Steffen Tiedemann Christensen — Mon, 28 Dec 2009 09:45:04 +0000

Hey Ian,

To save on requests to the twitter API the Twitterengine is bundling imports for multiple users into one request to Twitter. Since the import is done for new messages, you’ll find that active users will crowd out non-active users in the first imports. Over time, this is not an issue since you”ll be importing new stuff every five minutes and there won’t be time to for the crowding out to occur.

A simple fix is to change the “140″ in line 113 of main.py to a lower value. Possibly “20″ or “30″ — and then run the import again manually. This will give you more old tweets. Then change the value back to “140″ for the auto-imports when you have the service up and running. Otherwise you’ll run out of requests to the API.

By: ian

ian — Mon, 28 Dec 2009 02:37:19 +0000

@Steffen

Thanks! Figure it out after posting :(

The only issue I have right now is that it doesn’t seem to import tweets from accounts whose last tweet is like a week or so old.

I have more that 100 users that I’m monitoring. Could that be part of the reason?

By: Steffen Tiedemann Christensen

Steffen Tiedemann Christensen — Wed, 23 Dec 2009 23:09:27 +0000

@ian, You cannot configure this through config.py alone, instead you'll need to remove a line from main.py. The line reads: t.append('@'+str(u)) Just delete this line and you should see the only tweets from the specified users; not replies back at them.

By: ian

ian — Wed, 23 Dec 2009 22:43:31 +0000

How can we make this to return only tweets “from” the users specified under config.py and not include the re-tweets “for/to” those users?

By: Steffen Tiedemann Christensen

Steffen Tiedemann Christensen — Mon, 11 May 2009 06:19:24 +0000

@James:

git is a stand-alone application — not a part of Python. So to get the code by following the instructions above, you’ll need to install git on your computer. See http://git-scm.com/.

It you want to take the easier way, Github lets your download the code as a zip file as well: http://github.com/steffentchr/twitterengine/zipball/master