A Tumblr-scraping text post bot
Switch branches/tags
Nothing to show
Clone or download
Latest commit e5bf697 Sep 17, 2017
Type Name Latest commit message Commit time
Failed to load latest commit information.
.gitignore Initial commit. Apr 20, 2017
README.md Amend example Apr 20, 2017
genevabot.py Add tag generation Sep 17, 2017
requirements.txt Add requirements for distribution. Apr 20, 2017
tscrape.py Add tag generation Sep 17, 2017


Geneva Bot

Scrape Tumblr blogs for a corpus, convert it into a Markov probablility matrix, and generate text posts in the style of the original blog.

This process consists of two scripts, tscrape.py and genevabot.py. The first scrapes a Tumblr blog for text and tags (because lots of great content is included in Tumblr tags!) and creates a PyMarkovChain probability database from that data. The second simply reconsititues the probability database into memory and generates some sentances from it.

Here is a sample usage:

13:03:39: leo [~/Projects/genevabot]
$ ./tscrape.py clientsfromhell 1 10 --notags
Scraping page 1
Scraping page 10
Generating database...
Dumping database to clientsfromhell.posts.markov

13:03:59: leo [~/Projects/genevabot]
$ ./genevabot.py clientsfromhell.posts.markov 10
Perfect for everything from invitations to logos, the 35+ unique fonts – for example, an idea
logo, and brand image for my sister’s cake
It’s a Samsung iPhone
Me: I removed all the details
do it for free, and still
By making it easy to make bold and professional designs,
Thanks for the time to actually review said contract, I was
After this phone call, I received some revisions from said customer that came in to work with us can pay our full fees and is standard across
I sent it over – along with a
What will we do