tinypedia

A very bare bones, text only, local Wikipedia server. It works directly on the
.bz2 compressed dumps without creating any additional files.

Current State

Retrieving articles by title using the path #<URL-encoded-article-name>works
and the text can be viewed as extracted by wtf_wikipedia.js + some formatting
for sections. Sadly this fails to extract the text from special markup such as
IPA pronounciations. The raw mediawiki markdown can also be extracted using
/wiki/<URL-encoded-article-name>

Since we currently use URL encoding directly this is not compatible with the
title encoding used by Wikipedia (e.g. Ada%20Lovelace instead of Ada_Lovelace).

Building and Installing

First make sure you have Go and the go command installed and that
$GOTPATH/bin is in your path. Then install with a simple go get

go get github.com/ad-freiburg/tinypedia

Running

Change to the directory containing the dump files

• enwiki-latest-pages-articles-multistream-index.txt.bz2
• enwiki-latest-pages-articles-multistream.xml.bz2

And simply run the tinypedia exectuable

tinypedia

If you have named the files differenty use the -i and -d command line
switches to point tinypedia to the index and data files respectively.