Add optional preprocessing pipelines

[hrshdhgd]

Reference: https://github.com/monarch-initiative/mondo-ingest/issues/112#issuecomment-1329314132

• roman numerals: A function call to convert arabic to Roman numerals (or maybe vice versa? )
• stop word removal: similar to --exclude-tokens implemented in PR here
• plural: I think the Lemmatization pipeline already take scare of this?

[cmungall]

roman numerals can be covered by synonymizer?

[matentzn]

Yes, but not easily: you will need one rule per numeral? Or are you thinking an extension to synonymiser?