A simple script to parse a file containing Chinese text and generate Anki flashcards from it.
Text is parsed and split into individual words and phrases using the Stanza NLP package. Google translate is then used (via translate-shell) to retrieve translations and definitions for each separate word/phrase.
These translations are post-processed to include pinyin, and the final results stored in a tab-delimited file which can be imported by Anki.
This should be useful for parsing things like news articles, blog posts, or even full books (if you have access to the text for it).
At present, the tool is not hosted on PyPi.
The simplest way to get started is to clone the repo and use poetry to install & run it:
git clone https://github.com/khughitt/anki-chinese
cd anki-chinese
poetry install
You should then be able to run the CLI using poetry run ankichinese from the project directory, as
shown in the "Usage" section below.
usage: ankichinese [-h] [--append] [--convert]
[--format {traditional,simplified}]
INFILE OUTFILE
E.g.:
poetry run ankichinese \
example/harry_potter_wikipedia.txt \
example/harry_potter_wikipedia_vocab.txt
This will generate a tab-delimited file with five columns:
- Chinese
- Pinyin
- English
- Definition
- Sentence(s)
You can then import this file into Anki and assign fieldnames to each of the five columns.
Then, to specify how Anki should display the information on your flash cards, you can create a new note type which uses the five fields.
For example:
Front Template
{{ Chinese }}
Back Template
{{ Chinese }}
<br />
( {{ Pinyin }} )
<br />
<br />
{{ English }}
<br />
<br />
{{ Definition }}
<br />
<br />
{{ Sentences }}
Styling
.card {
font-family: arial;
font-size: 20px;
text-align: center;
color: black;
background-color: white;
}
.tone1 {
color: red;
}
.tone2 {
color: green;
}
.tone3 {
color: blue;
}
.tone4 {
color: purple;
}
.highlight {
color: magenta;
}
This should result in a format similar to what is shown in the example screenshot above.
Note that pinyin phrases are wrapped in html elements including a class indicating the tone mark of the character. This way, Anki can be configured to use a different font color for each different tone, if desired, as shown above.
- Google translate limits users to ~150 API requests every few hours so the script will need to pause after each such batch. Progress is saved periodically to avoid having to re-query many words if the script is stopped mid-way.
- I am currently using the script to translate traditional Chinese (繁體字). The script has generally been written in such a way that it can be easily extended to support simplified Chinese parsing, but a little bit of work would need to be done first.
- The list of "common words" that are filtered out
I probably won't have time to work on this much, but feel free to submit pull requests with bug fixes or improvements.
As mentioned above, steps were taken to begin to support simplified Chinese as well, but I have not thoroughly tested that functionality and it likely needs a little more effort to get it working properly.