Query preprocessing🔗
The query preprocessing step allows to transform the incoming query before it is processed by the tokenizer. This includes removing unrelated parts, rephrasing and phrase tagging.
Configuration🔗
Query preprocessing is defined in the 'query_preprocessing.yaml' configuration
file. The file must contain a list of steps. Each step has a mandatory
parameter step which defines the type of preprocessor. Additional step
configuration may then be set with additional parameters.
The steps are executed in the order that they are defined in the configuration file. Order matters here: each preprocessor works with the output of the previous step.
Pre-defined preprocessors🔗
The following is a list of preprocessors that are shipped with Nominatim. To learn how to define custom preprocessors, see the section on query processing modules.
split-japanese-phrases🔗
This preprocessor divides Japanese addresses into three categories: prefecture, municipality, and other. It adds potential phrase boundaries between the categories.