[Bing Spell Check] Azure spell check not working on Hungarian

[makaralaszlo]

Hello,

I want to use the Bing spellcheck API to correct the customer input data. I followed the documentation while making the implementation and checked that the Hungarian language was listed on the support list. Supported languages; in this given link under the "Bing supported language codes", the Hungarian language is listed as "hu".

I made the call using the Rest API by the following code:

example_text = "hejes íras lopas eseten mi a teendom" 
endpoint = 'https://api.bing.microsoft.com/v7.0/SpellCheck'
data = {'text': example_text}

params = {
    'setLang': 'hu',
     'mode': 'spell'
}

headers = {
    'Content-Type': 'application/x-www-form-urlencoded',
    'Ocp-Apim-Subscription-Key': api_key,
}

response = requests.post(endpoint, headers=headers, params=params, data=data)
json_response = response.json()
print(json.dumps(json_response, indent=4))

The output of the call was the following, where is empty data, but the input data was incorrect; it contained misspellings, and if I put English example text in it, it runs correctly for English.

{
    "_type": "SpellCheck",
    "flaggedTokens": []
}

In the case of using the SDK, the same applies; here is the Python code snippet of the call:

endpoint = 'https://api.bing.microsoft.com/v7.0/SpellCheck?mkt=hu&setLang=hu&mode=spell'
client = SpellCheckClient(endpoint=endpoint, credentials=CognitiveServicesCredentials(subscription_key))
result = client.spell_checker('hejes íras')

print(result)
    for flagged_token in result.flagged_tokens:
    print(flagged_token)
    for suggestion in flagged_token.suggestions:
        print(suggestion)

The output of the suggested words that are clearly not Hungarian words listed:

{'additional_properties': {'_type': 'Spelling/FlaggedToken'}, 'offset': 0, 'token': 'hejes', 'type': 'UnknownToken', 'suggestions': [<azure.cognitiveservices.language.spellcheck.models._models_py3.SpellingTokenSuggestion object at 0x1024ceac0>], 'ping_url_suffix': None}
{'additional_properties': {'_type': 'Spelling/TokenSuggestion'}, 'suggestion': 'hanes', 'score': 0.723271218679888, 'ping_url_suffix': None}

{'additional_properties': {'_type': 'Spelling/FlaggedToken'}, 'offset': 6, 'token': 'íras', 'type': 'UnknownToken', 'suggestions': [<azure.cognitiveservices.language.spellcheck.models._models_py3.SpellingTokenSuggestion object at 0x1024ce6a0>], 'ping_url_suffix': None}
{'additional_properties': {'_type': 'Spelling/TokenSuggestion'}, 'suggestion': 'bras', 'score': 0.723271218679888, 'ping_url_suffix': None}

Any suggestions or help is highly appreciated.
Thanks!

[azure-sdk]

Label prediction was below confidence level 0.6 for Model:ServiceLabels: 'Cognitive - Bing:0.384932,Cognitive Services:0.13450806,azure-spring:0.027664522'

[mccoyp]

Hi @makaralaszlo, thank you for opening an issue! It looks like the Query parameters documentation suggests using 4-letter codes instead of 2-letter codes for the setLang parameter. Have you tried using "hu-hu" for the setLang parameter?

I'll also tag the service team so that you can get a response from them as soon as possible if the above suggestion doesn't work.

Thank you for your feedback. This has been routed to the support team for assistance.

[SaurabhSharma-MSFT]

@makaralaszlo I believe this is the correct table to follow for spell check api v7 which does not include Hungarian. I am though checking internally and get back to you.

[makaralaszlo]

Thanks for the suggestion @mccoyp; I checked it, and the output is the same if I use 'hu' or 'hu-HU'.

@SaurabhSharma-MSFT is there any chance it will be supported soon if it is not supported?

[SaurabhSharma-MSFT]

@makaralaszlo Unfortunately none to my knowledge. If the feature is implemented that will be updated in the supported market list. We will now proceed to close this thread. If there are further questions regarding this matter, please reopen it and we will gladly continue the discussion.