Base-Art - Utils

SubveRSSed 1.0

2005-04-26T12:52:22+02:00

I previously introduced a Subversion post-commit hook to generate RSS feeds based on commits. But the generated feeds were not that smart. SubveRSSed 1.0 now handle RSS 1.0 feeds ! The next step will probably be Atom.

update (05/08/10) : SubveRSSed 1.1 has a great speed improvement since i dropped the “full RSS feed” support.

Cheetah templates overriding

2005-03-22T09:01:25+01:00

Cheetah is a very good templating language, especially because it’s the one (correct me if i’m wrong) to apply the Object-Oriented model to Documents. It can seem a bit weird at the beginning but when you get used to it, you can’t live without it :-)

So, i was looking for a good way to overload template parts without pain, here is how (couldn’t find the trick on Cheetah doc):

Assume you have a template Package/Page.tmpl like that:
```
#def footer
  This is my foot !
#end def
```

Create a Page_custom.tmpl template:

#extends Package.Page

#def footer
  $Page.footer($self) Where's my mind ? (dixit the Pixies)
#end def

You need to put a self variable in the template’s searchList which is the Cheetah template instance itself :-)

I don’t think that trick will work if you don’t compile your templates to Python code.

EFL tools in Python ?

2005-03-10T11:17:16+01:00

Just noticed an edje compiler in Ruby. That’d be fun to design graphical E17 stuff in Python too :-)

Inserting legend in images using PIL

2005-02-09T10:37:00+01:00

I wanted to insert legends directly in images instead of displaying them below respective images. ImageMagick can do that (Text directive) but it’s not smart enough to make the legend always visible, even in dark images (if the legend color is black). So i used this script to accomplish that task. Legend is inserted at the bottom of the image, in a little black bordered box. Advice: use pilfonts because default font is a bit small :-)

Here is a sample output:

progressCopy

2004-12-19T18:07:54+01:00

I use to deal with large files under the console (copying iso images, videos) but cp doesn’t display much informations. Few days ago i read a post from Richard Jones who wrote a very nice iterator aimed to display ETA informations about tasks on the console. So, here comes progressCopy.py.

Though it’s not as powerfull as GNU CP, it performs everyday tasks like copying, moving files or directories around with ETA informations :) Enjoy the console !

Database replicata

2004-10-06T12:40:29+02:00

I’m having few issues with the db (sqlite powered) behind this blog. So it’s necessary to migrate it to another DBMS (like PostgreSQL). The problem is that SQLite is a bit too light :) Fortunately, Alinea is using SQLObject to manage its database. So migration is trivial and can be done using a little script which takes two db URIs and handle the data replication from a db to another.

Bringing RSS to SVN

2004-09-17T09:05:01+02:00

About one week ago, Aaron Brady posted a svnstatus.py script which displays an SVN repository version history in an HTML document. I liked it and wanted to use it as a post-commit hook script for the XMLObject SVN repository. But svnstatus is not optimal at all. If you want to periodically build a static HTML file it’s ok. The problem, if you want to use it in post-commit hook is that it rebuilds the file for each commit, which turns to be really CPU hungry.

That’s why i wrote SubveRSSed.py. SubveRSSed incrementally builds an RSS feed containing the commit history of a SVN Repository. Then an HTML document is built based upon the RSS feed. This approach has many advantages:

post-commit hook friendly :)
the RSS can be made available for aggregators and people can monitor the activity/vitality of a repository
the history can still be consulted via the HTML document, if you don’t like aggregators nor RSS ;)

But there is one issue thought, the RSS accumulates all revisions, so it can grow enormously. I think i’ll make 2 feeds, one complete and one with the 20 latest revisions .. So SubverRSSed will evolve, may be in the PythonFR SVN repository :-)

BTW, many thanks to Aaron for svnstatus from which SubveRSSed used some functions and embedded CSS.

update: the script now handles 2 RSS feeds as described above :)

update2 : check out SubveRSSed at this url

update3(2005/05/05) : There’s a more recent post about this script.

When IMAP lags with big folders

2004-04-17T20:50:12+02:00

I’m using mutt to read my mail through IMAP4_SSL… But currently it lacks of a cache system, so dealing with big folders (~2000) is quite hard for it.

That’s why I hacked an archive utility script to better fit with my needs. The original one is here. The purpose is to move old mails on subtrees of the folder we want to archive, sorted by month.

I mainly added an option relating mail dates comparisons. So now there are two possibilities :

believe IMAP dates (appearing as timestamps in beginning of mail filenames) (default behavior)
believe Date: Mail header (-m option)

Use-case :)

$ ./mdirarchive.py -m respyre
$ tree ~/Maildir/.respyre* -L 1
/home/phil/Maildir/.respyre
|-- courierimapacl
|-- courierimapkeywords
|-- courierimapuiddb
|-- cur
|-- new
`-- tmp
/home/phil/Maildir/.respyre.2003-06
|-- cur
|-- new
`-- tmp
/home/phil/Maildir/.respyre.2003-09
|-- cur
|-- new
`-- tmp
/home/phil/Maildir/.respyre.2003-10
|-- cur
|-- new
`-- tmp

There is now a subfolder per month. Check script online help for more options.

Download