Contents |
Search returns incomplete results
Searching for files with paths
If you are searching for a file by typing in a full or partial path, Beagle won't find a match. That is because Beagle does not store a full path in its index for files, so that when files and directories are moved, it can be handled efficiently. Search for just the filename instead.
Searching for text
If you are searching for text you know should match a document that isn't found, one of four things is likely the cause: the item hasn't yet been indexed, the item's type was misclassified, the item's type doesn't have a corresponding filter, or there was an error trying to index the item. We'll explore these in detail.
Item isn't indexed yet
The first time you run Beagle, it has to crawl your home directory and index all of your data. If you have a lot of files, emails, or other documents, or your system is under heavy load, this can take up to several hours, and you may have to wait for all of your data to be indexed. The next time you run Beagle, it will still have to crawl your system, but it will only index files that have changed since the last time it was run.
Beginning with Beagle 0.2.14, the Beagle GUI will display an informational message box if the initial indexing process is running and that results may be incomplete. This box will disappear when the initial crawl has finished.
In all versions of Beagle, you can also use the beagle-index-info tool to see the status of your indexes. If the Indexing or Crawling line says true, then it is still indexing new data. You should also be able to see the Count increasing as documents are indexed.
Currently in SVN, you can use the beagle-dump-index tool to see if a specific document is indexed. Run:
$ beagle-dump-index --properties <file or URI>
to see if an item is found in one of the indexes. This works best for files and web pages, which have common URI formats. Emails, IM conversations, notes, etc. are harder to find. You might want to simply run:
$ beagle-dump-index --properties
and grep for well-known chunks of information like document title, email addresses, etc.
If the document is indexed but you are not getting the expected results or the indexing process has finished but missed certain documents, please file a bug and include your search terms and (if possible) the document that fails to match.
Item type is misclassified
Beagle uses freedesktop.org's MIME type specification and reference implementation to detect MIME types of files on the file system. If you are searching for content in a file and it isn't matching, but you do match on things like filename, this is likely the cause. If you run the beagle-extract-content tool on a file, it should tell you what MIME type was detected and which filter was used.
If the MIME type is misdetected, please file a bug against shared-mime-info and attach the doucment if possible.
If Beagle chose the wrong filter for a file, please file a bug against Beagle and attach the output of beagle-extract-content and (if possible) the document.
Item doesn't have a filter
Beagle indexes some information about all files it comes across, even if it is just the filename. To exact information from the file's data itself, Beagle uses pluggable filters. There are filters for OpenOffice documents, emails, HTML files, and many more. If Beagle doesn't have a filter for a certain file type, it won't be able to index the data inside the file.
To see if this is the case, check the log files, and look for the file. You will see a few lines like this:
+file:///home/john/subdir/foo.rpm No filter for /home/john/subdir/foo.rpm (application/x-rpm)
Filters are a great way to become familiar with the Beagle code and the easiest way to make a code contribution. See the Development page for more info.
Error trying to index the item
Finally, you may have hit a bug within Beagle. The best way to diagnose this is to check the log files for exceptions while indexing. Please include the log files for both the main Beagle daemon process and the index helper process when you file a bug.
