Migrating Anki to Pleco: Help with database format

burkelibbey

Member
I'm in the process of switching from Anki to Pleco for flash cards, and I'm currently dissecting a database backup to merge my Anki history in to save a bunch unnecessary reviews.

Is there any documentation on the meaning of the fields in the pleco_flash_scores_X tables? I'm specfically interested in:

1. reviewed: is this always just correct + incorrect, or am I missing some important nuance?
2. difficulty: Specifically, I'm looking for documentation on how this factors into the SRS algorithm so that I can find a way to convert Anki's "Ease" factor to this.

(My script so far, FWIW).
 

mikelove

皇帝
Staff member
1) Reviewed also counts cases like our 'review only' tests where cards were neither correct nor incorrect. But Anki doesn't have those, so yes, for import purposes you can just set that to correct + incorrect.

2) It's divided by a set number (40 by default) to produce a difficulty score. So the default Anki "Ease" of 2.5 would turn into a Pleco difficulty factor of 100.
 

Shun

状元
@mikelove This looks like some fine Ruby code to convert data, but of course, for you it‘s everyday code. So I wonder, which code in Pleco or which kinds of code in general do you find the most impressive still? The sort that runs extremely efficiently in C, handling a lot of data quickly, or code that performs highly complex calculations for Machine Learning or Voice Recognition? I don‘t know if you can readily answer this question, I just wondered where you stand on what kinds of code you find most impressive after so much programming. Thanks! :)
 
Last edited:

mikelove

皇帝
Staff member
I still write mostly in C - I prefer my own object-oriented mechanisms over relying on standard C++ ones, and C/C++ are still the only languages one can reliably / officially develop an app for iOS+Android in.

A lot of ML/AI code is written in C++, so we use quite a bit of it in Pleco to connect to that.

“Impressive” I don’t really know, I honestly don’t spend that much time looking at other people’s code outside of stuff that we’re integrating with Pleco.
 

Shun

状元
I still write mostly in C - I prefer my own object-oriented mechanisms over relying on standard C++ ones, and C/C++ are still the only languages one can reliably / officially develop an app for iOS+Android in.

A lot of ML/AI code is written in C++, so we use quite a bit of it in Pleco to connect to that.

“Impressive” I don’t really know, I honestly don’t spend that much time looking at other people’s code outside of stuff that we’re integrating with Pleco.

Thanks a lot! That customized OOP is rather impressive. I see you’re being very pragmatic about programming. I tend to analyze a lot from the outside.
 
Top