CODEWITHSUNDEEP
pythonazureresthttp-postazure-speech
user28741964
user28741964

Reputation: 1

Azure Speech to Text API using file in blob storage- 415 response Unsupported Media type

I am trying to run the Azure speech to text resource on some audio held in blob storage, so I have a python program that creates the file in blob storage and then attempts to run the speech resource on the file, but I keep getting a 415 response and the response content is just "b''", so Im having trouble figuring out what is wrong below is my request code

resp = requests.post(
        url, 
        data={
            "displayName": filename,
            "locale": "en-GB",
            "contentContainerUrl": container_uri,
            "ContentUrls": [content_uri],
                "properties": {
                "diarizationEnabled": False,
                "wordLevelTimestampsEnabled": False,
                "punctuationMode": "DictatedAndAutomatic",
                "profanityFilterMode": "Masked"
            }
            },
        headers={"Ocp-Apim-Subscription-Key": sub_key, "Content-Type": "application/json"}
        )

and here is the response

displayName: Recording.wav
Content container url:https://RESOURCE.blob.core.windows.net
contents url:  https://RESOURCE.blob.core.windows.net/cnnn-blob/Recording.wav
<Response [415]>
response content: b''

I have searched around for people with similar problems, but theyre accessing different resources, and what worked for them (specifying content type, changing authorization method) isnt working for me. I have also tried uploading the same file converted to different file types, but I dont think the target file is the issue, as the 415 refers to the request itself right?

Upvotes: 0

Views: 58

Answers (1)

JayashankarGS
JayashankarGS

Reputation: 8140

Make sure you are giving supported audio file like .wav, according to this documentation you need to use anyone of the contentContainerUrl or contentUrls property.

Use below code.

body = {
            "displayName": "sample-1.wav",
            "locale": "en-Us",
            "contentUrls": ["https://vjgsblob.blob.core.windows.net/data/audio/sample-1.wav?sp=r&st=2025-....sas_token"],
        }

resp = requests.post(
        "https://eastus.api.cognitive.microsoft.com/speechtotext/v3.2/transcriptions", 
        data=json.dumps(body),
        headers={"Ocp-Apim-Subscription-Key": "key", "Content-Type": "application/json"}
        )

Check the transcription job status with below code.

requests.get(resp.json()['self'],
             headers={"Ocp-Apim-Subscription-Key": "key", "Content-Type": "application/json"}
        ).json()

If it is succeeded, then get the details.

enter image description here

details = requests.get(
        resp.json()['links']['files'],
        headers={"Ocp-Apim-Subscription-Key": "8FphcJAdgP1lQyr3WEUMr8kbRiwlT5N7euy71O7w0NvBbUOLVpq4JQQJ99BAACYeBjFXJ3w3AAAYACOGSAkM", "Content-Type": "application/json"}
        )

details.json()

enter image description here

Next make get request to file contentUrl.

Upvotes: 0

Related Questions