Show HN: A tiny landing page for all my Finnish language learning software(finbug.xyz)
finbug.xyz
Show HN: A tiny landing page for all my Finnish language learning software
https://finbug.xyz/
I've had this domain sitting around for some time, and finally decided to turn it into a minimalistic "one stop shop" for fellow learners of this language. Everything on here is freely given to the public and was something that spawned directly out of my own needs and desires.
2 comments
This is neat. I notice that you do similar reverse lemmatisation as my Wiktionary API wrapper:
https://dictionary.nuenki.app/get_definition?language=Finnis...
Feel free to use it if you like, it's practically free, and open source if you'd like to host it yourself: https://github.com/Alex-Programs/nuenki-dictionary
Looking at your code, you're processing the actual word in a Finnish-specific way. If you'd like to generalise it, you can also just see if the definitions contains `x of y`, where y is a link, and automatically dereference to `y`.
Feel free to use it if you like, it's practically free, and open source if you'd like to host it yourself: https://github.com/Alex-Programs/nuenki-dictionary
Looking at your code, you're processing the actual word in a Finnish-specific way. If you'd like to generalise it, you can also just see if the definitions contains `x of y`, where y is a link, and automatically dereference to `y`.
Ha, what an honor to finally be noticed by Nuenki guy. Big fan of the project.
I did consider doing something Wiktionary-centric, and in fact have a Wiktionary JSONL scrape lying around courtesy of https://kaikki.org/ (from the same guy who started OpenSSH!). `tsk` does something similar to your defereferencing when it hits a "go deeper" phrase.
I decided against that approach in favor of the libvoikko spell checker because Finnish lies in this interesting zone of being an agglutinative language with a really, really regularized orthography. People love their neologisms here, and unfortunately most of them aren't catalogued in Wiktionary quite yet. I've found the mechanistic approach covers a lot of those edge cases well.
Take the word junttihenkiseni - the root form is junttihenkinen, but as of 05/06/2025 https://en.wiktionary.org/wiki/junttihenkinen does not actually exist. So `tsk` will have no data for it, but `finstem` with its mechanical approach works just fine. The word was originally coined to refer to Finland's unique spin on rock music in the 1970s and 80s.
On a broader level, if I can avoid hitting the network with small personal projects like this, I do try to. For example, `tsk` comes bundled every Finnish word with an English dictionary entry from Wiktionary, in a ~25 MB JSONL embed, and that allows us to build the randomly pruning trie that lets us get instantaneous prefix search across such a large space of things. I have met a lot of people who want to move to Finland from places where Internet is a sparse and valuable commodity, and I think their lives are much improved by having a tool they can just download one time and then use any place their laptop can be powered on.
I did consider doing something Wiktionary-centric, and in fact have a Wiktionary JSONL scrape lying around courtesy of https://kaikki.org/ (from the same guy who started OpenSSH!). `tsk` does something similar to your defereferencing when it hits a "go deeper" phrase.
I decided against that approach in favor of the libvoikko spell checker because Finnish lies in this interesting zone of being an agglutinative language with a really, really regularized orthography. People love their neologisms here, and unfortunately most of them aren't catalogued in Wiktionary quite yet. I've found the mechanistic approach covers a lot of those edge cases well.
Take the word junttihenkiseni - the root form is junttihenkinen, but as of 05/06/2025 https://en.wiktionary.org/wiki/junttihenkinen does not actually exist. So `tsk` will have no data for it, but `finstem` with its mechanical approach works just fine. The word was originally coined to refer to Finland's unique spin on rock music in the 1970s and 80s.
On a broader level, if I can avoid hitting the network with small personal projects like this, I do try to. For example, `tsk` comes bundled every Finnish word with an English dictionary entry from Wiktionary, in a ~25 MB JSONL embed, and that allows us to build the randomly pruning trie that lets us get instantaneous prefix search across such a large space of things. I have met a lot of people who want to move to Finland from places where Internet is a sparse and valuable commodity, and I think their lives are much improved by having a tool they can just download one time and then use any place their laptop can be powered on.
I'm the "Nuenki guy"!?!? Thanks!
And yeah, I used the same data source, just heavily processed. It's a great project!
I like your approach. I wish English used neologisms more; I use them occasionally, and it's just quite fun to create a new, lexically valid word to describe something novel.
The Nuenki dictionary database is a bit over a gigabyte, albeit with ~30 languages in it, so yeah that's definitely a bonus! JSONL probably compresses quite well, too. I added compression to Nuenki's (serialised-struct) dictionary entries a while back and it reduced the size by about 30%, iirc.
And yeah, I used the same data source, just heavily processed. It's a great project!
I like your approach. I wish English used neologisms more; I use them occasionally, and it's just quite fun to create a new, lexically valid word to describe something novel.
The Nuenki dictionary database is a bit over a gigabyte, albeit with ~30 languages in it, so yeah that's definitely a bonus! JSONL probably compresses quite well, too. I added compression to Nuenki's (serialised-struct) dictionary entries a while back and it reduced the size by about 30%, iirc.
I think everyone in the world knows, at least: "kolme, kaksi, yksi, tuli!"