Re: Creating New Filter
In a sense, yes. What I want to be able to do is use a query like 'LIKE image1.jpg' and Beagle will return all images within a certain distance of image1.jpg. pHash uses its own database organized as a Then you want to implement an IQueryable that parses the query, looks for LIKE:xxx query clauses and creates results using its own database. Just one caveat: beagle query syntax allows boolean AND, OR, grouping and several others things - and your IQueryable will not be able to deal with these queries. That is, your IQueryable will not be able to meaningfully support abcd LIKE:1234 (meaning: files with abcd in it and similar to 1234); it can handle LIKE:1234 queries safely. Of course, the IQueryable can flat refuse to serve any query not of the form LIKE:xyz. There would be another issue of getting results from multiple backends - its not clearly defined what happens in this case. So if you have the files backend enabled, then you will have results from both the files backend and your iqueryable and results from your iqueryable might get replaced by the files backend results. There might have been some way you could override this ... but I don't remember any details =( - dBera ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Creating New Filter
With that said, it seems like we could create our own backend that is just an IQueryable that merely traverses a directory and creates our own index. Then anytime a file is being searched for with the correct query syntax, we can consult our index and return the correct results. However, our index is specific to the media file type, e.g. all audio files go into one index, images in another but it seems like the backend should not care about the file type, so I'm not sure if this is the correct approach or not. I am still not sure what you want. Do you want to augment the search functionality so that users can *also* get similar files ? Or do you want to return the list of similar files when a user search matches a particular media file e.g. if I search joe satriani, it gives me media which are tagged with joe-s. and similar ones. Think about the different components in this way: - IQueryable - only for querying - if you want to do extra things only during a query - Backend - the full hammer - if you want to do everything your way - Filter - only for data to store in the index - if you want to add extra things in the index (both queryable information or non-queryable information) Keep in mind the fact that, you can store additional information in the index for each file which you can use during querying to do extra things. But I am not sure of your exact intention about pHash and beagle. ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Re: Creating New Filter
I am having trouble getting my Beagle filter to work. I have placed it in the correct location but when running beagle-extract-content it reports that 0 filters were loaded from my extension. Possibly (1) the localtion was incorrect (2) the mimetype or extension are not correctly reported You can figure what is going wrong by carefully looking at (1) external-filter.xml file (2) output of beagle-extract-content (3) output of beagle-info --list-filters (or whatever is the right command to list all the filters) I also have several additional questions: 1) Can custom indexing methods be created? I want to use my own database/tree to store a perceptual hash of a media file to allow similarity searching of media files. In order to efficiently search for similar files, we are using a specially built tree. No. If you want to store your own indexing data, then why would you need beagle ? 1) Are filters able to be chained? I am creating an image filter and only want to create the perceptual hash from it but would like any other image filters installed to also run on the file. Yes only if you are willing to write C# code. E.g. lot of our image filters are extended from other simpler image formats by subclassing from the generic image classes. You can do the same; but there is no way to do this outside of C#. This is by design. - dBera -- - Debajyoti Bera @ http://dtecht.blogspot.com beagle / KDE fan Mandriva / Inspiron-1100 user ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers
Creating New Filter
Hello all, I am having trouble getting my Beagle filter to work. I have placed it in the correct location but when running beagle-extract-content it reports that 0 filters were loaded from my extension. What could be the cause of this? The code is very simple and I can post it if necessary. I also have several additional questions: 1) Can custom indexing methods be created? I want to use my own database/tree to store a perceptual hash of a media file to allow similarity searching of media files. In order to efficiently search for similar files, we are using a specially built tree. 1) Are filters able to be chained? I am creating an image filter and only want to create the perceptual hash from it but would like any other image filters installed to also run on the file. Thank you Evan ___ Dashboard-hackers mailing list Dashboard-hackers@gnome.org http://mail.gnome.org/mailman/listinfo/dashboard-hackers