Re: [Tracker] Spreadsheets support: Gnumeric

[Jamie McCracken <jamiemcc blueyonder co uk>]

Luca Ferretti wrote:
> If we want add support for different spreadsheet formats, we can add
> some filters using Gnumeric support tools:
>
> SSCONVERT
>         "It is  a  command  line  utility  to  convert spreadsheet files
>         between various spreadsheet file formats." (from man ssconvert)
>         
>         Supported formats:
>         .as    Appliz
>         .oleo  Oles
>         .wb?   Quattro Pro
>         .dbf   Xbase
>                Paradox Database or Index
>                SC/xspread
>         .wk?   Lotus 123
>         .dif   data interchange
>         .mps   linear and integer programming
>                MultiPlan (SYLK)
>         .csv   comma separated values
>         .tsv   tab separated values
>         Plus, of course the .gnumeric format and MS Excel.
>         
>
> SSINDEX
>         "It is a command line utility to generate index data for various
>         spreadsheet file formats." (from man ssindex)
>         
>         This is really really good tool for our purpose. It's used by
>         Beagle too. While ssconvert is used to convert the format of a
>         spreadsheet (so we should export it in a know format, extract
>         relevant stuff from here, then delete temporary stuff), ssindex
>         do the hard job for us. Running `ssindex -i` on attached
>         Gnumeric spreadsheet you have:
>         
>         <?xml version="1.0" encoding="UTF-8"?>
>         <gnumeric>
>           <data>January</data>
>           <data>January</data>
>           <data>Travelling:</data>
>           <data>Total:</data>
>           <data>Buy new music:</data>
>           <data>Drinking:</data>
>           <data>Buy new RAM:</data>
>           <data>Eating</data>
>           <data>February</data>
>           <data>February</data>
>           <data>Travelling:</data>
>           <data>Total:</data>
>           <data>Buy new movies:</data>
>           <data>Drinking:</data>
>           <data>Buy printer ink:</data>
>           <data>Eating</data>
>         </gnumeric>
>         
>         All relevant data in XML, ready to be indexed. Numbers are
>         formulas are purged. Maybe it could be able to extract metatada,
>         but I wasn't able to save them in Gnumeric...
>         
>         The only limitation seems the supported formats: running
>         `ssindex -i` you have this list:
>         
>         application/x-gnumeric
>         application/csv
>         application/tab-separated-values
>         text/comma-separated-values
>         text/csv
>         text/spreadsheet
>         text/tab-separated-values
>         text/x-comma-separated-values
>         
>         Not so much, but honestly I was able to successfully run it on
>         Excel too, so maybe it could work on old formats like Lotus123
>         and Quattro Pro. I've to find some of those files on the Net.
>         
> So, anyone ready to work on it?

yeah this comes with the gnumeric-gtk package in Edgy as does ssconvert

looks simple - just create the shell script for it and shove the output 
to plain text using XSLTProc (see tracker/filters/text/xml_filter)

btw someone was complaining we dont index OASIS spreadsheet - I think 
that needs investigating too although the ssconvert might help there too.

--
Mr Jamie McCracken
http://jamiemcc.livejournal.com/