Often times, it can be helpful to look at the frequency of a word as you're looking it up.
The goal of yomi-freq is to make this efficient workflow more widely accessible by allowing the user to generate their own Yomichan frequency list dictionaries.
After ensuring that you have Python and Pip installed, clone the repository and install the necessary dependencies:
git clone https://github.com/kamui-fin/yomi-audio && cd yomi-audio
pip install -r requirements.txt
The basic usage of yomi-freq is by supplying an input list file and output directory:
python main.py -i [paths...] -name [name]
Additional options include:
--output: directory where the output dictionary is stored (default:.)--revision: custom revision name for metadata (default: freq)--limit: limit number of entries in dictionary (default: no limit)--chunksize: custom size for each chunk during processing: (default 10,000)
Multiple lists can be combined into a single yomichan frequency list by passing in multiple paths to -i. Frequency values are calculated using mean and rounded.
The script expects some sort of list or csv-like file where each line starts with a single word, followed by a delimiter. For example, the following structure is accepted:
[word][delimiter][frequency]
So is:
[word]
The delimiter character is up to you but optimally it's either a comma or tab. Only the word column will be read in the existing order to determine the resulting frequency "ranking", in the sense that the smaller the value, the more frequenct the word is.
freq.zip is the output Yomichan dictionary file and can be directly imported now.
All contributions in the form of pull requests, bug reports, etc. are gladly welcomed. If you would like another frequency list format to be supported, let me know.
