xml file formats and search by content

[bezvardis]

Since yesterday I was trying to get search by content to find particular files containing particular words. I noticed that it just would not find some particular files. Some other files would be found, though. I got desparate and searched all possible topics on the net. But then I copied the text that finder could not find into a MS word document and saved it. Finder found it immediately. The same text saved by NeoOffice coul not be found if searched by content. Now I stumbled across the help in NeoOffice which talks about xml file formats and says that it actually uses some kind of compression like that of zip files. So I thought - maybe the content in this format is so deeply concealed that finder just cannot get to it and terefore does not find the file? If that is so - is that a bug or it is a feature of OpenOffice?

[bezvardis]

Now I experimented a bit more with this and found out that the file can be found if I enter the search words in the properties of the file. when I remove them and reset the properties, finder again cannot find the file.

It seems also that file saved by NeoOffice in .doc format becomes searchable to Finder

[pluby]

The reason that Finder won't search the contents of Neo/J files is because Finder does not support searching of the OOo file format. Apple did put in searching of MS Office formats, but Apple has not bothered to make any of their tools handle the OOo file format.

Patrick

[ovvldc]

[pluby]

Yes, the OOo files are merely plain text XML files zipped up. In fact, you can manually unzip them using the following terminal command:

jar xvf <OOo file>

Patrick

[bezvardis]

[sardisson]

Some time ago someone mentioned trying to get support for search OOo files in certain search applications (other than the Finder). The end result was that the author of at least one of the programs agreed to look in to supporting OOo files. If you do a search here, you'll probably find the thread and the names of the apps.

Ed has at least given some thought to writing a plugin for the forthcoming 10.4 Spotlight engine and the OpenDocument (OOo 2.0) format...whether he'll have the time to do so only he can answer.

Smokey

[fabrizio venerandi]

if you have not problem with hd space or formatting, you can use the old staroffice format file. I uses staroffice format for big document for example, 'cause opening calc file in neooffice format is too slow.
And osx search content look into staroffice files.


f.

[bezvardis]

Thank you all for useful advices. Eventually I managed to find some opensource program called docsearch which can do the search of NeoOffice and other kinds. The interface is not very friendly to the eye but at least it works.

[sardisson]

I can't find the old posts I was looking for; I think they probably got lost when Ed restored trinity after he came back from vacation one month. There's one thread where SOB noted he had requested the authors of EasyFind and SpeedSearch to add OOo/Neo doc support. But I remember a post where someone reported that some search author had agreed to add support to the next version of his product, and I can't find it now.

Here are some other relevant OOo links, just to have them in one place:
http://www.danielnaber.de/loook/
http://oootools.free.fr/fooox/
http://www.openoffice.org/issues/show_bug.cgi?id=14468

Smokey

[bezvardis]

I had some interesting discussion on this topic at the Apple Panther support page and here are some work-arounds that I found:
1) docSearcher that I mentioned (requires indexing and gives back search results that to me were not very comprehensible but others might think different) docSearch finds all documents in all formats that contain particular thread. the download page is here http://www.brownsite.net/docsearch.htm
2) someone posted a terminal search code (if that's how it is called) ( http://discussions.info.apple.com/webx?13@116.yy3YaI37RO4.927450@.68a8d1b9/6 ) which works quite fast, uses no indexing and gives a simple list of files containing the thread. But it looks only for xml file formats so all the other files have to be searched separately. that someone also said that it might be quite easy to write similar code for Apple Script and make it run from NeoOffice or finder or whatever - maybe there are poeple who want to take this challenge and write a script? I have no knowledge in the Script and thus cannot do that myself
3) at this link http://www.darwinwars.com/lunatic/bugs/oo_macros.html someone posted a set of macros among which there is supposed to be one that does the search inside Openoffice. Other macros worked fine for me but the search made NeoOffice quit. I don't know if that was my mistake or the fault of the script. If someone else tries that and makes it work - let me know.
4) Easyfind did not work for me when I wanted to find the NeoOffice documents, which might mean that the promises from them to look at this problem remained just promises

[sardisson]

[Max_Barel]

[OPENSTEP]

FWIW I've had the idea to do a Spotlight filter for OOo docs for some time but have not had the time to do it. Most of my relevant research is in the NSFilter proposal on dashboardbuddha, and the bulk of the concept would apply for an OSS Spotlight plugin framework. For better or for worse, I tend to put bug fixing (or narrowing down bugs) on a higher priority level then Spotlight plugins as, well, the OS isn't even fully released to the public yet

ed

[bezvardis]

Gib Henry has posted a script (written by biovizer) that resulted from discussion on this topic on one of Apple discussion boards http://trinity.neooffice.org/modules.php?name=Forums&file=viewtopic&t=1186
The script though has some limitations (e.g, one cannot quit it easily when it runs, it only searches for one word or a phrase but not keywords, it also searches only Openoffice files and no other - things like that). Otherwise it works pretty well and quite fast and looks nice. If anyone would like to improve it - that would be even more fantastic than the fact that biovizer wrote this piece