Investigate monolingual cleaning

[eu9ene]

Based on @marco-c's feedback we should investigate how the HPLT project cleans monolingual data and whether we should adjust our cleaning procedure.

https://hplt-project.org/HPLT_D3_1___Software_for_cleaning_data_sets.pdf

[marco-c]

Worth looking at: https://hplt-project.org/HPLT_D3_1___Software_for_cleaning_data_sets.pdf.

[marco-c]

See also #247.

[ZJaume]

We are now using https://github.com/pablop16n/web-docs-scorer which gives us much better results when inspecting cleaned data. But this requires cleaning at document level.