How to Alter a Hyphenation Dictionary

The hyphenation dictionaries are stored as XML files in the [CHEMISTRY_INSTALL_DIR]/config/hyph directory.

You can copy the dictionaries you need to change in another directory, then use the -hyph-dir parameter to refer them inside your transformation.

Each file is named with the language code and has the following structure:

<hyphenation-info>

<hyphen-min before="2" after="3"/>

<exceptions>
o-mni-bus
...
</exceptions>

<patterns>
préémi3nent.
proémi3nent.
surémi3nent.
....
</patterns>

</hyphenation-info>

To change the behavior of the hyphenation, you can modify either the patterns or the exceptions sections:

exceptions: Contains the list of words that are not processed using the patterns, each on a single line. Each of the words should indicate the hyphenation points using the hyphen ("-") character. If a word does not contain this character, it will not be hyphenated.
For example, o-mni-bus will match the omnibus word and will indicate two possible hyphenation points.
Note: Compound words (i.e. e-mail) cannot be controlled by exception words.
patterns: Contains the list of patterns, each on a single line. A pattern is a word fragment, not a word. The numbers from the patterns indicate how desirable a hyphen is at that position.
For example, tran3s2act indicates that the possible hyphenation points are "tran-s-act" and the preferable point is the first one, having the higher score of "3".