How to produce a list of all words of a given Language from Wiktionary?

Question

Say I want to produce a list of all Latin words from the English Wiktionary (because Wiktionary in English contains the most complete dictionary for Latin, and perhaps for every language).

More precisely, I want to produce a list of every entry that corresponds to a Latin word. These entries have one thing in common, namely, that the website address has the following type:

https://en.wiktionary.org/wiki/<word>#Latin

For instance, sensorum. The key here is that 1) the domain is the English wiktionary (en.wiktionary) and 2) the website ends in #Latin.

Based on other answers (e.g. here), I found out there is a directory containing all pages of the English wiktionary here. But this does not, as far as I can see, separates by language. The "main" file (nwiktionary-latest-all-titles.gz) contains a list of around 7 million objects (words?). This might be the full list of pages in English. I want only those pertaining to a certain language.

There seems to be related posts here and here but I cannot make sense of the answers for my own case. There is also this resource to extract items from Wiktionary, but haven't figure out how to use it for my purposes yet (I used it for extracting lists of categories here but not all words in Wiktionary are indexed into categories).

Anyone can provide some hints on how to do it?

PS: if this is off-topic, please suggest where this might fit better.

Answer 1

Working with the full Wiktionary dump is hard, so I recommend using data sources which extract the parts you are interested in and present them in an easier to digest format.

wiktextract looks like a great fit for your request, especially this list.

If you are willing to accept data from the Latin Wiktionary, dbnary is also an excellent data source. If you prefer relational data to RDF triples, I provide downloadable sqlite databases as part of WikDict.