Google Desktop Search equivalent
I need an application which will allow me to search a monster index of academic word documents for Slackware GNU Linux the same way Google Desktop Search does for win32.
Beagle seems to be the most popular for linux, but it requires extensive gnome/gtk deps which slackware doesn't have. So what do you use? And what should I use? rworkman on FreeNode suggests I use "slocate(1)". Wtf is that? |
rob0 explains that in Slackware, slocate / updatedb is installed and working by default if you leave the machine running overnight.
|
|
Quote:
|
Drone4four is asking for an application capable of reading and indexing file contents, not file names.
|
recoll is very nice ( http://www.lesbonscomptes.com/recoll/ ). I needed to install three other packages (antiword, unrtf and xapian) but it indexes rtf, doc, odt, txt, pdf ... and the search is very fast. You can even get a preview. And it handles umlauts ok if you need them.
|
Apart from recoll that it is already suggested, you can also try
- beagle: http://beagle-project.org/Main_Page and - Kat: http://sourceforge.net/projects/kat/ I've not tried kat so I can not make any comments. Several months ago I tried Beagle which is quite ok. GS |
Pat could include http://en.opensuse.org/Kerry into his kde release to facilitate searching.
|
beagle has a lot of dependencies, so i think the chances are pretty low:
http://beagle-project.org/Installing_prerequisites http://pinot.berlios.de/ looks promising, but still has the occasional bugs. recoll looks nice. but one really has to take a closer look at its source code. don't hold your breath for one of these to be included anytime soon. ;) EDIT: |
pinot and recoll look very similar, at least in its dependencies, but pinot adds some new dependencies like gtkmm, libtextcat, Google SOAP API ... I'm glad I found recoll first or second after beagle, else I would probably have given up frustrated from all the extra stuff they need.
It was really easy to build on Slackware with only one minor hazzle (antiword insisting to install to /usr/local, until I manually edited the Makefile). |
Quote:
Besides, the last release was a year and a half ago. I'm not sure this is a live project. Ugh. |
Desktop search
Recoll works nicely for me. Light and fast, easy to compile, indexes when I tell it to rather than adding to the clutter running in the background. Helped me find stuff since the latter half of my uni course when I started using it.
I tried Kat initially, but had bad experiences with it, and as someone says above the project is dead as far as I can tell too. I had a quick butcher's at Novell's desktop when it came out with Beagle to see how Beagle works, and while the search part was alright I personally don't like things running in the background eating up resources. I'm not sure if Beagle has got lighter since me trying it out back then. For me it felt like a huge gorilla that wasn't particularly quick on it's feet compared to Recoll. YMMV. :) |
Quote:
I want to say thanks for mentioning recoll. I hadn't run across it before and it was EXACTLY what I was looking for. Thanks! |
recoll acts even nicer if you don't want to index your whole harddisk but just have some folders with documents. You have to fiddle around with the recoll.conf file but this way the indexing goes even faster, especially if your documents change often this will save some time.
|
Quote:
Running MAKE gives this: Code:
bash-3.2$ make |
yes:
Quote:
if you don't have experience with c++, you best wait until the authors release a version that compiles cleanly. |
Quote:
Quote:
|
Try changing line 80 from:
Code:
(char **)&rp, Code:
(const char **)&rp, |
Quote:
Only difference: I installed all four apps to /usr instead of /usr/local like you seem you did. Had to change the Makefile for antiword (changed line to "GLOBAL_INSTALL_DIR = /usr/bin" where /usr/local/bin was hardcoded) and configured the others with --prefix=/usr. But honestly I don't know if that makes a big difference. EDIT: This program has by far the shortest configure output I've seen. If you want to compare ... aspell and xapian-config are the only ones that seem to be checked here: Code:
configure: WARNING: you should use --build, --host, --target |
OK, so here's the output from ./configure:
Code:
checking for aspell... /usr/bin/aspell Quote:
Don't see those lines there anywhere, except for lines 22-24: Code:
common/rclversion.h: VERSION |
Quote:
Quote:
|
Thanks, but I'm afraid it doesn't help much:
Code:
bash-3.2$ make |
It's not a real solution to the error you get, but I could send you my Slackbuild scripts I used or even the packages of all four via email. Don't know how to solve the error other way ... :(
|
Titopoquito, you could send them at linuxpackages so others could use them too.
|
Generally a good (or very good) idea. The problem is I don't have the time now to clean them up from highly idiosyncratic stuff and get them tidy so that they would get accepted at linuxpackages.net or slackbuild.org. Maybe next week, but definitely not this weekend :( And I guess it would take some more time for them to be reviewed and accepted.
|
I am interested in recoll myself, so the review time at SlackBuilds.org would be fairly short :-) In fact I was already prepared to build the packages for my own repository (http://www.slackware.com/~alien/slackbuilds/) but to have your SlackBuilds on SBo is a good addition.
Eric |
Quote:
Thanks. |
If it absolutely won't compile, you could try 'rpm2tgz' then 'installpkg' on an rpm build of it.
|
Ok, I will try to prepare the scripts for submission on slackbuild.org, so that others could get them if they want. kite, I will send you the packages, then maybe in some days the build scripts will be available anyway ...
|
Quote:
But, a little later... no such luck. I tried src2pkg as well and that didn't work either. Quote:
I'll wait for the script at SlackBuilds.org, there's no hurry. And I have just a hunch that it won't last very long before we see it there. ;) By the way: I tried Kat, but that's horrible. Eats up all your memory and Slack is sloooooooooooooooooooooow with Kat on the background (not to mention the two crashes of Kat). So I got rid of it. |
I've just submitted the scripts. Let's hope they are well done so that they are accepted by Eric or the other guys from slackbuild.org :)
|
They're in the SBo repository now.
Eric |
Got it installed now without a hitch. :)
Thanks! UPDATE: It works great! Fast and very little RAM-consuming. Nice work. |
Great thread and slick work!
I have been following this with great interest and now all is installed and running using the Slackbuild scripts. This will fulfil a wish I have had for indexing a data archive. SlackMagic! :-) Thanks to all concerned. |
I installed Recoll like everyone else just fine with SBo. Special thanks to the SBo team! After reading the dox on the Recoll website I learned how to set index certain directories. Now my problem is: how do I click on a searched item on the results page? It's like I can't click the results. wtf??
|
There are preview and edit links in the first line of every result (well, edit only shows up if recoll knows what program to use to edit it). The whole result isn't clickable, but the links should be there.
|
Quote:
|
Now the index is horribly incomplete. Reading the official documentation here, http://www.lesbonscomptes.com/recoll...ual/index.html I learned that this is the command to start the index:
Quote:
edit:sp |
adriv if you do decide to try converting a binary rpm package, you can use src2pkg for that instead of using rpm2tgz and installpkg. src2pkg incorporated binary rpm and deb conversion a while back. It should do a better job because the conents get checked and corrected like other src2pkg output and it will insert the slack-desc for you which is missing for the package otherwise.
Still, getting the sources to compile correctly would the best solution. Maybe you are missing some overlooked package or configure option. I don't use KDE and don't normally install any of the qt or kde stuff so I can't easily help. |
Quote:
Bad to hear that your index isn't complete. I've got nearly no document in my index that is unique enough to test that. |
Quote:
|
Quote:
By the way, the Slackbuild worked perfectly. ;) |
EDIT: The following patch enables OGG and FLAC audio file indexing in recoll when used with the recoll build script from http://www.slackbuild.org
--- What I still missed from recoll is the ability to index flac and ogg audio files. I created a patch for that, which I want to give you below if you want flac and ogg support. These changes are not on Slackbuild.org because the submission part is disabled at the moment. I testet it and it seems okay, but please report if you encounter any problems with that. Oh, and of course it requires that you the Slackware's flac and vorbis-tools package installed :) If you have untarred the Slackbuild "recoll.tar.gz" from Slackbuild, go to that folder. Save the code below as a file and apply it: Code:
patch -p0 < /path/to/where/you/saved/this/file Code:
--- recoll.SlackBuild.withoutflac 2007-06-28 14:54:27.491524250 +0200 |
By the way, I noticed now that Google has released Google Desktop for Linux. http://desktop.google.com/linux/
|
All times are GMT -5. The time now is 08:21 AM. |