The Russian community has an agreement on the format of house numbers, but it is slightly different from how addresses are usually written. Example: "5 литера Б" is "5 литБ" in OSM. This can be fixed by client-side preprocessing, but I think it should be done by the search server.
Most common values (dots are optional):
Full
Abbreviations
OSM
Example
дом
д.
-
42
корпус
корп. / кор. / к.
к
42 к1
литера
литер / лит.
лит
42 литА
строение
стр. / с.
с
42 с1
сооружение
соор.
соор
42 соор1
флигель
флиг. / фл.
фл
42 фл1
In addition, the search will fail if input string contains additional data such as room number. Perhaps, we should create a list of tokens that can be deleted?
In addition, the search will fail if input string contains additional data such as room number. Perhaps, we should create a list of tokens that can be deleted?