• Features Overview
• Installation and Usage
• Test Data
• Limitations

The parts extraction tool facilitates the creation of parts from a full orchestral score encoded in MEI format, provided that the data structure adheres to the specific encoding guidelines. For instance, you can extract the first violin part from a symphonic score.

The major benefit of the tool is the ability to extract parts that share a staff. Consider the following example, where the first and second oboes are notated on the same staff. Using Extract Parts, you can create separate parts for the first and second oboes.

In addition to the standard single-voice parts, it is possible to create custom score reductions. For example, the user may wish to extract only the accompanying parts.

The tool is written in XSLT. If you are using the oXygen XML Editor, the repository provides a project file called extract-parts.xpr that contains the corresponding transformation scenario and the required settings. This can be opened by double-clicking on it in the file explorer or by opening it from the project view in the oXygen XML editor.

To illustrate, I will describe the process of extracting a first oboe part from a previously shown score. First, open the configuration file (./config/config.xml). Here you can select the parts you need by adding the element <dme:part/> as a child of <dme:parts/>. The @dme:layer and @dme:staff attributes allow you to specify the staves and layers you want to extract.

<parameters xmlns="http://www.mozarteum.at/ns/dme">
  <parts>
    <part staff="1" layer="1"/>
  </parts>  
</parameters>

Then, apply the transformation scenario extract-parts to the ./tests/data/K165/01/dmeedtA_165-001.mei file.

After applying the transformation scenario, a new MEI file dmeedtA_165-001_(1-1).mei is created in the same folder. The suffix (1-1) in the filename indicates the part that has been extracted. The following is the rendering of the extracted oboe part.

Some of the default behaviour of the tool can be controlled by setting parameters. In the oXygen XML editor, this can be done by double-clicking on the transformation scenario and selecting the parameter dialogue box. You may also wish to make a copy of the original transformation scenario.

P_LANGUAGE
The default language for the instrument names and abbreviations generated by the tool is Italian. Other languages such as English and German are available. The language can be changed by setting the parameter to 'EN' or 'DE'.

For this feature, Extract Parts uses an XML configuration file located in ./lib/lists/instruments.xml. This file contains a comprehensive list of names and abbreviations and their transpositions for common instruments and voices.

P_OUTPUT_PATH and P_OVERWRITE_FILE

The default output path of the file is in the same folder as the MEI file (./). You can change this by setting the parameter using the relative path syntax. For instance, if you want to output the MEI file to a different folder, you can set the parameter as follows: ./output/. Additionally, if you set P_OVERWRITE_FILE to true() you can use the save as option available in oXygen. and specify a different path or filename.

P_SHRINK_MEASURES
If the extracted part contains more than four measures with a rest, a multirest measure is created. If you have very good reasons, you can override this behaviour by setting the parameter to true() (default) or false(). Note that the Evaluate as XPath checkbox must be enabled. This option is only available for single part extraction.

P_GLOBAL_CONTEXT_ITEM and P_XSPEC_TEST Both parameters are only used in the context of XSpec unit testing.

P_LIB_PATHRelative of absulute path to thelibfolder:/path/to/the/lib. Default: lib`.

P_USER_PARAMETERS
This parameter defines which staves and layers are requested for extraction. This is an XML element defined in /config/config.xml. If all the parts of the score are requested, the file will be copied without processing.

Extract Parts expects a specific data structure. Examples can be found in the ./tests/data folder. However, not all data may be up to date. The following data is currently approved to be compliant with this version of the tool.

• ./tests/data/K165/01/dmeedtA_165-001.mei
• ./tests/data/K165/01/dmeedtA_165-001_m11.mei
• ./tests/data/K165/dmeedtA_165-004.mei
• ./tests/data/K550/03/550-003_m9_(2).mei
• ./tests/data/K621/621-23_m187_flutes.mei
• ./tests/data/K621/621-23_m25_fagotti.mei
• ./tests/data/290-001_m5_(3).mei
• ./tests/data/K543/01/543-01_mm109-111_viola.mei
• ./tests/data/K527/527-07_mm8-9_fagotto12.mei
• ./tests/data/K550/01/550-001_m139-140_clarinetti.mei
• ./tests/data/K621/621-23_m181_flutes.mei

• If the MEI file contains multiple <scoreDef>s with a different number of staves, the tool may not work properly. The workaround is to temporarily comment out the score portions.
• The handling of variants encoded with <mei:choice/> and <mei:app/> elements is a challenging area of development, as many combinations are possible. The current version of the tool partially supports variants, but it can make mistakes.
• To use the extracted parts for performance, you may need to make some modifications, such as adjusting line and page beginnings, adding cue notes, etc.

Current version: 5.1.1. For details see the CHANGELOG.

Author: [email protected]