I have been trying to implement the below shell code with python. I am about to make use of deepaffects speaker identification api . So before its's use i need to enroll the audio file with user id , in their docs there is no python example given instead of below shell commands.
curl -X POST "https://proxy.api.deepaffects.com/audio/generic/api/v1 /sync/diarization/enroll?apikey=<ACCESS_TOKEN>" -H 'content-type: application/json' -d @data.json
# contents of data.json
{"content": "bytesEncodedAudioString", "sampleRate": 8000, "encoding": "FLAC", "languageCode": "en-US", "speakerId": "user1" }
So far now I had written the below code.
import requests
url = 'https://proxy.api.deepaffects.com/audio/generic/api/v1 /sync/diarization/enroll?apikey=<3XY9aG7AbXZ4AuKyAip7SXfNNdc4mwq3>'
data = {
"content": "bytesEncodedAudioString",
"sampleRate": 8000,
"encoding": "FLAC",
"languageCode": "en-US",
"speakerId": "Pranshu Ranjan",
}
headers = {'content-type': 'application/json'}
r = requests.post(url, data=data, headers=headers)
print(r)
But I don't know how to pass the "content": "bytesEncodedAudioString"
. I have audio samples in mp3 format in my local directory.
here is the deepAffects api reference and they support multiple audio formats
According to documentation:
content (String) base64 encoding of the audio file.
Just use built-in base64
module to encode your audio file:
import base64
import requests
filepath = "C:\Audio\...\file.mp3"
with open(filepath, 'rb') as f:
audio_encoded = base64.b64encode(f.read()) # read file into RAM and encode it
data = {
"content": str(audio_encoded), # base64 string
"sampleRate": 8000,
"encoding": "FLAC", # maybe "MP3" should be there?
"languageCode": "en-US",
"speakerId": "My Name",
}
url = ...
r = requests.post(url, json=data) # note json= here. Headers will be set automatically.