README.Docker.md

Building and running your application for Docker

When you're ready, start your application by running: docker build -t hanzigraph-server .

Start a container with: docker compose up

Your application will be available at http://localhost:8000.

Customise Wordlists and Sentences

You might want to run HanziGraph on your own filtered wordlists instead of the full dictionary provided.

For this, create a wordlist.json file containing the array of your words:

[
  "礼物",
  "一起",
  "能",
  "..."
]

Then, in compose.yaml add a new bind mount to your custom wordlist by adapting YOUR-PATH-TO and TARGET-CHARSET.
TARGET-CHARSET should be one of simplified, traditional, hsk or cantonese.

    ...
    volumes:
      - "<YOUR-PATH-TO>/wordlist.json\
        :/usr/src/app/public/data/<TARGET-CHARSET>/wordlist.json\
        :ro"
    ...

Similarly, you can customise the available sentences too. Create a sentences.json file containing an array with elements as the one below. The Chinese translations are split into meaningful words to guide segmentation.

    {
        "en": "Guilin's sceneries are the most beautiful ones under the heaven.",
        "zh": [
            "桂林山水",
            "甲天下",
            "。"
        ],
        "pinyin": "Gui4lin2 shan1shui3 jia3 tian1xia4."
    },

Just like with wordlists, activate your custom sentences:

    ...
    volumes:
      - "<YOUR-PATH-TO>/sentences.json\
        :/usr/src/app/public/data/<TARGET-CHARSET>/sentences.json\
        :ro"
    ...

Finally, start the container with docker compose.