Fix wiki links about mediawiki_to_git_md HOT 10 CLOSED

See also github/pages-gem#69

from mediawiki_to_git_md.

See also jgm/pandoc#1982 - we or pandoc must also turn spaces into underscores in the relative URLs, and deal with special characters like percent signs with encoding, etc.

This also applies to the filename for each wiki page, where we probably need to follow the MediaWiki "canonical database form". Quoting http://www.mediawiki.org/wiki/Manual:ImportImages.php which focused on files (see our issue #1):

Note: The "canonical database form" ... is obtained from the file name by capitalizing the first letter, replacing all spaces with underscores, and then replacing multiple consecutive underscores with one underscore.

Quoting http://www.mediawiki.org/wiki/Manual:Title.php

In the article name spaces and underscores are treated as equivalent and each is converted to the other in the appropriate context (underscore in URL and database keys, spaces in plain text). "Extended" characters in the 0x80..0xFF range are allowed in all places, and are valid characters. They are encoded in URLs. Extended characters are not urlencoded when used as text or database keys. Other characters may be ASCII letters, digits, hyphen, comma, period, apostrophe, parentheses and colon. No other ASCII characters are allowed, and will be deleted if found (they will probably cause a browser to misinterpret the URL).

from mediawiki_to_git_md.

Having tested this now, GitHub/Jekyll serves pages with and without the .html extension UNLESS there is a folder present with the same base name (e.g. BioSQL.html works, but BioSQL does not - perhaps this is fixable?).

This means we can (probably?) continue to use the MediaWiki style links (without extension) and work when rendered via GitHub/Jekyll. However, the links won't work when viewing the markdown directly in GitHub where the links would need a .md extension.

However, we still need to fix spaces vs underscores, capitalisation, and rare special characters in links.

from mediawiki_to_git_md.

With the biopython.org wiki we were lucky that one URL was a problem - wiki/BioSQL - minor enough we can probably just move the single page child page wiki/BioSQL/Windows to wiki/BioSQL_on_Windows (or similar) and remove the problematic folder wiki/BioSQL/.

For www.open-bio.org there are currently ten folders like wiki/BOSC/ and wiki/BOSC_2014/ which make direct access to the parent page require adding .html to the URL. See OBF/OBF.github.io#1

from mediawiki_to_git_md.

The pandoc issue was fixed, the MediaWiki writer will do spaces to underscores in wikilink URLs.

from mediawiki_to_git_md.

Still issues (probably worth filing with pandoc), e.g. on the OBF main page,

``the scope and role of our [[membership]].inMain_Page.mediawiki`

Pandoc v1.15.2 made this the scope and role of our [membership](membership "wikilink"). in Main_Page.md.

It should have been made into a wikimedia style page name with leading capital, i.e. the scope and role of our [membership](Membership "wikilink"). instead.

from mediawiki_to_git_md.

Another problem, again on the OBF main page,

`[[:File:OBF-Bylaws.pdf| Our bylaws]] lay out how the Board is elected, ...``

with pandoc v1.15.2 gives:

[ Our bylaws](:File:OBF-Bylaws.pdf "wikilink") lay out how the Board is elected, ...

Interestingly :File:OBF-Bylaws.pdf as the URL works in mediawiki, redirecting to File:OBF-Bylaws.pdf which seems more sensible.

I've edited the wiki to link directly to the file with:

[[media:OBF-Bylaws.pdf|Our bylaws]] lay out how the Board is elected, ...

from mediawiki_to_git_md.

I've created another issue in pandoc re the files .
jgm/pandoc#3052

from mediawiki_to_git_md.

Have you found any solution for fixing links to other .md files, adding the extension? I wonder if anybody else wrote a script to post-process it, parse the generated MD, check links, and try to apply fixes to links that don't work according to which way they're broken.

from mediawiki_to_git_md.

@ec1oud We've just used extension-less URLs in all the MediaWiki conversions thus far.

from mediawiki_to_git_md.