Prev: 5.3. The RVA system[ home ]Next: 6. Metadata support

5.4. Music Store Builder

5.4.1. Data sources
5.4.2. Regular expressions
5.4.3. Predefined transformations
5.4.4. Capitalization

Aqualung's Music Store Builder allows you to easily create a store from your existing audio files. This is essential if you already have a large music collection on your harddrive.

The Builder supports two approaches for automatically adding audio files to the Music Store. The directory driven mode requires the collection to be organized in a strict filesystem structure, where the directory structure itself specifies which files belong to a record, and which records belong to an artist. The structure should conform to one of the following patterns:

a) root / record / track
b) root / artist / record / track
c) root / artist / artist / record / track
d) root / artist / artist / artist / record / track

Pattern a) is useful when you do not have separate subdirectories for artists, only for records. The b) means that each artist and record has its own subdirectory. It corresponds to the structure introduced above along with the Music Store. The patterns c) and d) are for those who have too many artists to keep them at the same level; you may thus have a subdirectory for every artist starting with letter a, b, etc., or you can even dedicate two levels to them, perhaps as b/be/beatles, c/cr/crimson, etc.

The independent mode does not follow any directory structure, only searches for audio files recursively starting from root. This doesn't allow the Builder to perform CDDB lookups, so only metadata and filename transformations are available with this mode. If your audio files are basically organized in a directory structure that allows a directory driven build to be performed, but you also happen to have audio files on higher (record or artist) levels, run the Builder in directory driven mode first, then use the independent mode on the same root directory to add the outstanding files too.

The Builder can gain information about your music from three sources: metadata (if your audio files are correctly tagged), CDDB lookups, and the filesystem names of the audio files and directories themselves.

If you are up to building a new store, first create an empty store (right-click in Music Store, then choose the Create empty store item), then right-click on the new store and choose Build / Update store from filesystem. After selecting the desired building mode, the builder dialog appears with several options broken down into five notebook pages. Later, if you only want to update the content of a store, you can start a build process on a non-empty store as well.

On the General page you have to select the root directory of your music collection. With the directory driven mode, the appropriate structure should also be selected.

You can exclude files from the build process that match a wildcard pattern (or more precisely, at least one element in a list of wildcard patterns). The list is comma-separated, e.g. `*.jpg,*.png,*.gif' will cause all files having jpg, png, or gif extension to be left out. Note that hidden files (whose name starts with a period) are always excluded. Also note that no spaces should be used around the commas, because they would be counted in the wildcard patterns, causing a lot of headache.

Similarly, you may include only those files that match a wildcard pattern. The syntax is the same as with the exclusion. This easily allows you to build a store containing music e.g. encoded with FLAC, and another store for Ogg, while the audio files can be placed together even in the same directory. Exclude and include patterns are case-insensitive.

If you choose to reset existing data, only track name, length, manual RVA and empty comment fields will be overwritten, and the latter two are only if you have enabled importing the Replaygain and Comment tags on the Track tab. Removing non-existing files will remove all tracks from the store that no longer exist on the filesystem. Empty artists and records will also be removed.

The builder dialog has separate notebook pages for artists, records and tracks. The settings made on a page only concern the appropriate category (either artist, record or track). The operations and transformations are performed in the same order as they are placed on the page, from top to bottom.

5.4.1. Data sources

With each mode you can set the priority of data sources for artist names, record titles and tracks independently (highest comes first). Simply drag & drop the lines of the list to achieve the desired order. For example, if you have proper tags in you audio files for track names only, but none for artists and records, you may use metadata for tracks at first place, and use CDDB for artists and records instead. The data sources can also be disabled by clicking on the check boxes.

The Builder looks for usable data in the given priority, and will use the first suitable one.

A short note on the Filesystem source: when using the directory driven mode with structure a), the artists and records will have the same filesystem name, which is the name of the record directory. Using pattern b), c) or d) will result in all artists, records and tracks having the name of their respective directory (or file, for tracks) as the filesystem name. With c) and d), the name of the rightmost artist directory is used. When using the independent mode, artists, records and tracks will have the same filesystem name, which is the filename of the audio file being imported.

Regardless of the data source that won the contest for being the source of your artist, record or track name, you can perform several transformations on this string before accepting it as final.

5.4.2. Regular expressions

The main goal of the regexp module is to perform sed-like substitutions, allowing you to extract the artist, record or track name from the string. It has the effect of a sed s/regexp/replacement/g command. The regexp must not match the empty string (the only exception is leaving the regexp entry empty, which disables the whole regexp module). The replacement is allowed to be empty in which case the matched parts are deleted. The replacement may also contain the special escapes 1 through 9 to refer to the corresponding matching sub-expressions in the regexp. See sed(1) and grep(1) for more information on sed substitution and regexp syntax. There is a sandbox where you can enter a string and see how it would be transformed, providing a safe environment for experiments.

5.4.3. Predefined transformations

These provide some basic fixes, such as removing the extension or leading numbers, converting underscores to spaces, and trimming leading, tailing or multiple spaces. You can use them together with the regexp module as well, in which case only the replacement will be transformed, as expected.

5.4.4. Capitalization

Finally, with this module you can force a capitalization pattern on the string. The force case entry is a comma-separated list of words or expressions that will be left in the form you entered them. An obvious example is the case of Roman numbers.

After the standard modules present for artist, record and track names, there are some category dependent settings. For artists and records you can set the source of the sort name field, for tracks you can decide whether to import the Replaygain and Comment tags, if available.

Once you are satisfied with the settings click OK. A new window showing the progress is displayed, and the store is filled up. You can abort the process if you encounter some error or weirdness, or just want to refine some settings.

Prev: 5.3. The RVA system[ home ]Next: 6. Metadata support