Copy cs to sk for prototyping #132
Conversation
src/rules/sk.toml
Outdated
| @@ -0,0 +1,17 @@ | |||
| allowed_symbols_regex="[A-Za-zěščřžýáíéóďťňúůĚŠČŘŽÝÁÍÉÓĎŤŇäöüÚ‚–\\. \"„“]" | |||
There was a problem hiding this comment.
allowed_symbols_regex="[A-Za-zěščŕřžýáíéóôďťňúůĺľÁÄĚŠČŔŘŽÝÁÍÉÓÔĎŤŇĹĽäöüÚ‚–\. "„“]"
src/rules/sk.toml
Outdated
| needs_uppercase_start = true | ||
| even_symbols = ["\""] | ||
| broken_whitespace = [" ", " ,", " .", " ?", " !", " ;"] | ||
| abbreviation_patterns = ["[A-ZĚŠČŘŽÝÁÍÉĎŤŇÓÚ]+\\.*[a-z]*[A-ZĚŠČŘŽÝÁÍÉĎŤŇÓÚ]+", "atd\\.", "\\baj\\.", "tj\\.", "\\brec\\.", "[nN]apř\\.", |
There was a problem hiding this comment.
abbreviation_patterns = ["[A-ZĹĽĚŠČŔŘŽÝÁÍÉĎŤŇÓÔÚ]+\.[a-z][A-ZĹĽĚŠČŔŘŽÝÁÍÉĎŤŇÓÔÚ]+", "a i\.", "a pod\.", "atď\.", "\baj\.", "tj\
.", "\brec\.", "[nN]apr\.",
""."", "\s[^aikosuvzáó]\s", "zkr\.", "[Tt]zv\.", "[dD]r\.", "\b[aAeE]d\.", "\b[sS]?[tT]r\.", "[aA]rch\.", "Inc\.", "Ltd\.", "[pP]opr\.",
"\b[fF]r\.", "\b[A-Z]+DR\b", "[pP]ozn\.", "[sS]rov\.", "\b[eE][a-z]\.", "[zZ]ejm\.", "[JS]r\.", "\b[lL][lL]",
"Mgr\.", "[mM]j\.", "\b[sS]tol\.", "\b[pP]ol\.", "Ing\.", "[cCkK]pt\.", "\b[lL]t\.", "Mr?s?\.", "\s[^\\s]{1,2}\.", "\bviz\.", "\b[sS]at\."]
|
Blocklist generated from words of frequency 60 and lower |
|
Downloaded and sent for review to five native speakers. |
|
Sorry, I missed that comment.
No. We can't accept corrected sentences, because we need to run a new, fresh export once the rules are added. This is needed to make sure that we fulfil all legal requirements. As sentences are picked at random, any changes to them would be lost. |
Mainly to get the extraction running and to get an idea how much more work will need to be done.