Skip to main content

Reactive Multi-language Gradio App with minimal effort

Project description

gradio-i18n

GitHub License PyPI - Version Huggingface

Reactive Multi-language Gradio App with minimal effort. Enables Gradio app displayiing localized UI responding to the browser language settings.

Installation

pip install gradio-i18n

Basic Example

If you want localized UI based on user's browser language, simply wrap your block definition with gradio_i18n.Translate().

  • Wrap all your texts to localize with gradio_i18n.gettext(), refer to the example.
  • Translate() returns the user's language to pass to function inputs.
  • Translate() accepts a filepath or dict as the translation dict.
  • Optional argument placeholder_langs will fill key not translated as placeholder back to the disk file.
import gradio as gr

from gradio_i18n import Translate, gettext as _

def greet(name, gender, lang):
    return _("Greeting") + f"! Hello {name} {gender} in {lang}!"

with gr.Blocks() as demo:
    with Translate("translation.yaml", placeholder_langs=["en", "zh", "ja", "ko", "es", "fr", "de"]) as lang:
        name = gr.Textbox(
            label=_("Name"), placeholder=_("Input your name here.")
        )
        gender = gr.Radio(choices=[_("Male"), _("Female"), _("Unknown")])
        output = gr.Textbox(label=_("Greeting"))
        submit_btn = gr.Button(value=_("Submit"))

    submit_btn.click(greet, inputs=[name, gender, lang], outputs=output)

demo.launch()

Change UI language Example

You may want user to choose their expected language, pass the component controlling the language value to gradion_i18n.Translate().

def greet(name, gender, lang):
    return _("Greeting") + _("Hello {name} {gender} in {lang}").format(name=name, gender=gender, lang=lang)


with gr.Blocks() as demo:
    lang = gr.Radio(choices=[("English", "en"), ("中文", "zh"), ("日本語", "ja"), ("한국인", "ko"), ("español", "es"), ("française", "fr"), ("Deutsch", "de")], label=_("Language"))
    with Translate("translation.yaml", lang, placeholder_langs=["en", "zh", "ja", "ko", "es", "fr", "de"]):
        name = gr.Textbox(label=_("Name"), placeholder=_("Input your name here."))
        gender = gr.Radio(choices=[_("Male"), _("Female"), _("Unknown")])
        output = gr.Textbox(label=_("Greeting"))
        submit_btn = gr.Button(value=_("Submit"))

    submit_btn.click(greet, inputs=[name, gender, lang], outputs=output)

demo.launch()

Advanced Usage for better control

  1. Prepare a translation dict like examples below.
  2. Wrap text intended to be localized with gradio_i18n.gettext()
  3. Invoke gradio_i18n.translate_blocks(), within the context of gradio blocks.
import gradio as gr
from gradio_i18n import gettext, translate_blocks


def greet(name):
    return f"Hello {name}!"


lang_store = {
    "en": {
        "Submit": "Submit✅",
        "Name": "Name 📛",
        "Greeting": "Greeting 🎉",
        "Input your name here.": "Input your name here. 📝",
    },
    "zh": {
        "Submit": "提交",
        "Name": "名字",
        "Greeting": "问候",
        "Input your name here.": "在这里输入你的名字。",
    },
}

with gr.Interface(
    fn=greet,
    inputs=gr.Textbox(
        label=gettext("Name"), placeholder=gettext("Input your name here.")
    ),
    outputs=gr.Textbox(label=gettext("Greeting")),
    submit_btn=gettext("Submit"),
) as demo:
    translate_blocks(translation=lang_store)

demo.launch()

[!NOTE] Keep in mind that the translate_blocks() function MUST BE called in the gradio block context (with)

Get/Set current language

Except for automatically translated text value in Gradio components, user may expect to get the current language of the user for localizing contents. To get the current language, simply pass in any component into argument lang of translate_blocks(), and you will get the language value in the object.

Also, if you change the value of the passed in object, the overall UI will update accordingly.

Example:

import gradio as gr
import gradio_i18n

def get_lang(lang):
    return f"You are using language: {lang}"


with gr.Blocks() as block:
    lang = gr.Dropdown(choices=["zh", "en", "ja"])
    display_lang = gr.Text()
    btn = gr.Button()
    btn.click(get_lang, inputs=[lang], outputs=[display_lang])
    gradio_i18n.translate_blocks(lang=lang)

block.launch()

[!NOTE] Only dict type state is supported at this moment.

Generate a translation dictionary placeholder

To build the transtion dictionary, we provide function dump_blocks() to dump all the i18n texts from the gradio blocks object.

This is an example of using yaml/json to persist the translation.

JSON

import gradio as gr
import gradio_i18n
import json

trans_file = "translations.json"
if not os.path.exists(trans_file):
    lang_store = {}
else:
    lang_store = json.load(open(trans_file))

# define your gradio block here....
# with gr.Blocks() as block:
#     ....
#     gradio_i18n.translate_blocks(block, lang_store)

collected_texts = gradio_i18n.dump_blocks(block, langs=["zh", "en"], include_translations=lang_store)
json.dump(collected_texts, open(trans_file, "w"), indent=2, ensure_ascii=False)

YAML

import gradio as gr
import gradio_i18n
import yaml

trans_file = "translations.yaml"
if not os.path.exists(trans_file):
    lang_store = {}
else:
    lang_store = yaml.safe_load(open(trans_file))

# define your gradio block here....
# with gr.Blocks() as block:
#     ....
#     gradio_i18n.translate_blocks(block, lang_store)

collected_texts = gradio_i18n.dump_blocks(block, langs=["zh", "en"], include_translations=lang_store)
yaml.safe_dump(collected_texts, open(trans_file, "w"), allow_unicode=True)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

gradio_i18n-0.0.10.tar.gz (12.3 kB view details)

Uploaded Source

Built Distribution

gradio_i18n-0.0.10-py3-none-any.whl (10.0 kB view details)

Uploaded Python 3

File details

Details for the file gradio_i18n-0.0.10.tar.gz.

File metadata

  • Download URL: gradio_i18n-0.0.10.tar.gz
  • Upload date:
  • Size: 12.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.19

File hashes

Hashes for gradio_i18n-0.0.10.tar.gz
Algorithm Hash digest
SHA256 1b4a542aeeb605944b5b5a60cf826c256ca48fcd17910fa72ddb97b0a39555cc
MD5 fb0d20b989c9b6536ca8d7d91de88203
BLAKE2b-256 8affc3983b6309acb5eec53a775566132d678d9c4fac9b137ab88531a8d49248

See more details on using hashes here.

File details

Details for the file gradio_i18n-0.0.10-py3-none-any.whl.

File metadata

  • Download URL: gradio_i18n-0.0.10-py3-none-any.whl
  • Upload date:
  • Size: 10.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.19

File hashes

Hashes for gradio_i18n-0.0.10-py3-none-any.whl
Algorithm Hash digest
SHA256 87254a4cd1647c2522942803109b4184bbfb7f4e160716742b0b275c903e8301
MD5 aa56372f2a11421a8db08be4bd9f02ae
BLAKE2b-256 82bde12c7ac376f324eb9316e5c508ace3af826558c79c38de616ce89e54e1d8

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page