Nepomuk Strigi File Indexing and strigiclient not using the same data?

Question

I'm using KUbuntu 10.04. For quite a while now, I have configured and enabled the Nepomuk/Strigi server to index files on my desktop. Strigi is stil indexing after over a week, and the dialog when I click on the Strigi icon in the system tray tells me there are currently 17,062 files in index, and the Nepomuk store size is 2.8 GiB.

I just installed strigiclient to actually be able to search that data, and upon launching strigiclient, it tells me there are zero documents indexed, and that the index size is 0MB. When I click the "start indexing" button in strigiclient, the numbers begin to grow, independent of the Nepomuk/Strigi file indexing data.

Are Strigi and strigiclient supposed to be using the same data? What has the Nepomuk Strigi File Indexing been doing this whole time, and how can I seach the data it has indexed? I would like to use Strigi to index my data, but the documentation on that is nearly non-existent, and I'm left experimenting and guessing, and finding that strigiclient and Nepomuk/Strigi don't seem to care for each other.

How can I conduct simple searches with Strigi? (i.e. "find a file containing the string 'foobar'" or "find a file containing 'foobar' in its name").

Answer 1

I belive you are correct. Kubuntu 11.04 has much the same issue. strigiclient shows -1 files indexed. My understanding is that newer KDE versions use Nepomuk to track file information and Nepomuk invokes a nepomukstrigiservice, not strigidaemon, for full-text indexing. Use System Settings > Desktop Search > Desktop Query > File Indexing to control how it works; the configuration is stored in ~/.kde/share/config/nepomukstrigirc , not ~/.strigi/daemon.conf. It may be possible to use strigiclient/strigidaemon alone with their separate config and database, but you should probably then disable Nepomuk.

The GUI way to do full-text search using Nepomuk/strigi in KDE is to use the Dolphin file manager, which in KDE 4.6 has a Find... panel with a [Content] option. There are deepgrep and deepfind command-line tools that also seems to use its index. Of course there's also grep -ri sometext some/path ☺