mophahr / wikipediagame Goto Github PK

Is it allowed or not? And how would we count it?

We probably need at least to exclude links leading to a member of the following categories (in German) to make it more fun:

• Mitgliedstaat der Vereinten Nationen
• Bundesland (Deutschland)
• Tag
• Jahr
• ...

I dont know if this fucks up the small-wold properties though.

So far its hard coded in wikigame/link_extraction.py

Start a path -> klick a link -> klick the back-button -> klick another link

You get the 'not reachable from your path' error

Maybe only check first name?

The list of person related categories is ~9GB for german. This crashes on most machines (memory error).

Given the current issues with link filtering based on categories, I propose doing it by names.

Some list of names I've found so far:

• .gov baby names
• USA census of 1990: 80k unique names
• http://stackoverflow.com/questions/1803628/raw-list-of-person-names

Given a set of links, I propose we match them against a list of names, and consider it if the link name has at least two names in the list.

in link_extraction.py

How can we fix that?

'ascii' codec can't encode character '\xf6' in position 67: ordinal not in range(128)

Request Method: GET
Request URL: http://127.0.0.1:8000/article/Morgan%20Freeman
Django Version: 1.6.4
Exception Type: UnicodeEncodeError
Exception Value:

'ascii' codec can't encode character '\xf6' in position 67: ordinal not in range(128)

Exception Location: /usr/local/lib/python3.3/http/client.py in putrequest, line 953
Python Executable: /home/msch/Programme/wikipediaGame/venv/bin/python
Python Version: 3.3.2
Python Path:

['/home/msch/Programme/wikipediaGame',
'/usr/local/lib/python33.zip',
'/usr/local/lib/python3.3',
'/usr/local/lib/python3.3/plat-linux',
'/usr/local/lib/python3.3/lib-dynload',
'/home/msch/Programme/wikipediaGame/venv/lib/python3.3/site-packages']

Server time: Wed, 18 Jun 2014 14:39:15 +0000

so far it's only time

Is it a lot of effort to have a language selector button somewhere on the start page?

I don't know how the networks-longo in this case is, but for example 2 clicks should be a path length of 2, not 3.

So I propose we don't count the start article.

one of them returns an 'no links found error'

Not that important for now, since there aren't that many around

For example on http://127.0.0.1:8000/article/Frank%20Wilczek there's a link to http://en.wikipedia.org/wiki/Black_holes which should have been removed..

So you could stay in the leaderboard...

example: "Hertha BSC" is not a name

even though https://en.wikipedia.org/wiki/David_McAllister contains a bunch...