There are quite a few issues where blanks are reported. For example:
#275
This is NOT a template processing problem. This is a functionality that is not in this tool, namely, the ability to process Wikipedia Lua into text.
In one specific case, the existing extractor has a hand-rolled Lua processor, for convert, which fills in blanks in EN pages such as "The lowest recorded temperature was ," for Buffalo -> "The lowest recorded temperature was -20F, It should be pointed out that this is already "wrong", in that if there is another language whose Wikipedia has a differentconvert` module, it will produce the wrong text.
An alternative would be to add a dependency such as lupa and start interpreting some of these modules. It wouldn't cover all modules, it'd be the first external dependency for WikiExtractor (probably start off making it optional), and it'd slow things down. It would fill in a lot of the existing blanks, though.
Any thoughts on such a project?
There are quite a few issues where blanks are reported. For example:
#275
This is NOT a template processing problem. This is a functionality that is not in this tool, namely, the ability to process Wikipedia Lua into text.
In one specific case, the existing extractor has a hand-rolled Lua processor, for
convert, which fills in blanks in EN pages such as "The lowest recorded temperature was ," for Buffalo -> "The lowest recorded temperature was -20F,It should be pointed out that this is already "wrong", in that if there is another language whose Wikipedia has a differentconvert` module, it will produce the wrong text.An alternative would be to add a dependency such as
lupaand start interpreting some of these modules. It wouldn't cover all modules, it'd be the first external dependency for WikiExtractor (probably start off making it optional), and it'd slow things down. It would fill in a lot of the existing blanks, though.Any thoughts on such a project?