Automatically editing wiki pages (for private wiki) in command line?

Here's what I want to do on a wiki (not Wikipedia, a private one). Each person has a page and each day they add something to their own page. I want to have a script which computes the difference between yesterday's page and today's page (like the history thing wiki's have) and then writes the new stuff to a single central page, basically collating what they've written.

Of course I could do this manually but a script would be preferable. I have admin access to the server the wiki is on so there shouldn't be any permission issues. I'm familiar with copying bits from standard text files and concatinating them into a single file but I don't know how to work with the format wiki uses. Is this possible and if so can it be done with someone with just basic scripting knowledge?

I suppose if the 'compute difference' bit is unpleasant I could just get everyone to write the text into a .txt and then have the script copy that into the personal pages and also the central page. That'd just need me to know how to edit wiki pages in the command line.

 

Log in or Sign up to hide all adverts.

I would guess that you could likely just write a parser in whatever CGI language you have available that takes the information recently stored directly from the database and format's it onto a separate page. You could either have it parse everytime the script runs (which wouldn't be a good scripting practice) or have it create a static page once a day with the information you want in it. (this lessens the database accesses).

 

Log in or Sign up to hide all adverts.

The typical tools used for this task; patch, diff, diff3, merge, will not work (easily) due to your Wiki syntax. Whatever diff tools you use will have to be smart enough to properly reformat the diff to Wiki, if I understand what you're trying to do.

However, I will mention GitHub hosts its own wiki's (Using markdown) and git seems to adequately merge branches.

You would probably have a much easier time using Markdown, as patches and diffs don't need to be fully aware of Wiki-syntax closures. If this is something you're interested, I'd be happy to whip up some patching examples. I would need to know first... are you trying to concatenate all NEW data into a single page...or are you trying trying to concatenate all data into a single page?

Code:

#!/bin/bash 

function cat_diff_of_page 
{
    pushd $1 &> /dev/null 

    OLD="page.old"
    NEW="page"

    echo "Diff Page for $1"
    diff $NEW $OLD 

    cp $NEW $OLD 
    popd &> /dev/null 
}

# BASE WORKING DIRECTORY 
WD="/tmp"

# PAGE TO AMALGAMATE DIFFS TO 
DIFFP="$WD/amalgam_dif"

# PAGES TO PROCESS
PTP=( "john" "james" "ralph" )


#!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
# This part is only here to set up the example # 
#!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
for PAGE in "${PTP[@]}"
do 
    echo "$PAGE" 
    if [ ! -e $PAGE ];
    then 
        mkdir $PAGE 
    fi 

    pushd $PAGE 

    echo "Some words" > page 
    echo "Some words" > page.old 
    echo "and more words." >> page.old 
    popd 
done 
#! END SETUP 


# BEGIN EXECUTION 
cd $WD 

echo "#===================================================" >> $DIFFP 
echo "# `date`" >> $DIFFP 

for PAGE in "${PTP[@]}"
do
    echo "--------------------------------------------------" >> $DIFFP
    echo "`cat_diff_of_page $PAGE`" >> $DIFFP 
    echo "" >> $DIFFP 
done 
# END EXECUTION 

This assumes you have structure like...

Code:

.
├── amalgam_dif
├── james
│** ├── page
│** └── page.old
├── john
│** ├── page
│** └── page.old
└── ralph
    ├── page
    └── page.old

And in amalgam_dif It will output...

Code:

#===================================================
# Mon May 14 21:47:46 PDT 2012
--------------------------------------------------
Diff Page for john
1a2
> and more words.

--------------------------------------------------
Diff Page for james
1a2
> and more words.

--------------------------------------------------
Diff Page for ralph
1a2
> and more words.

(Because the diff for each one of them was the same)

 

Log in or Sign up to hide all adverts.

MediaWiki (what Wikipedia uses, roughly) and Twiki are just 2 data standards. Both allow one to mine history and perform diffs, but the details vary greatly. The perl library http://search.cpan.org/dist/Algorithm-Diff/lib/Algorithm/Diff.pm I have found helpful in the past. The problem is, ANY part of a wiki page can change -- unless you structure the content and make updates un-wiki-like -- So you need to decide a spec for what counts as "new" material.
1) Any added/changed paragraph
2) Any added/changed section
3) Any added section
4) Any added block of text at the end.

A question of presentation arises -- do the copied sections link back to their source pages? How are the different contributions delimited?

The last step is merging in or replacing the contents of the old page with the new material.

 

Well, rpenner, let's hope his situation isn't so dire as to require Perl as a solution

Please Register or Log in to view the hidden image!

.

 

Have you ever wanted to start your own wiki website that anyone can edit, like wikiHow, Wikipedia or Wikitravel? Here's how to do it.