What would be the best way to address multiple language utterances?

[broomva]

I'm wondering if something standardized like i18n or similar could be implemented such that a meta routing layer directs to the correct languange, and then, utterances are defined in a given languange

[bruvduroiu]

Interesting, could this be added as a pre-processing step outside of semantic router?

from langdetect import detect
from collections import defaultdict

# Your list of mixed language strings
mixed_lang_strings = [
    "This is an English sentence.",
    "這是一個中文（繁體）句子。",
    "Das ist ein deutscher Satz.",
    "Another English sentence.",
    "又一个中文句子。",
    "Noch ein deutscher Satz."
]

# Create a dictionary to hold lists of strings by language
lang_dict = defaultdict(list)

# Detect the language of each string and append to the respective list
for string in mixed_lang_strings:
    try:
        lang = detect(string)
        if lang == "zh-cn" or lang == "zh-tw":  # Combining Simplified and Traditional Chinese
            lang = "zh"
        lang_dict[lang].append(string)
    except Exception as e:
        print(f"Error detecting language for '{string}': {e}")

We can then programatically create Routes:

from semantic_router import Route

routes = [Route(name=lang, utterances=lang_dict[lang]) for lang in lang_dict]

Would this solve your issue?