[PYTHON] [Voice analysis] Collect bird song data with xeno-canto

What is xeno-canto?

Bird bark database. There is API that can be handled by json. This time, I used Python to collect all the calls of the genus Fulica (Coots) in Japan.

Source code

import requests
import json
import urllib

url = "https://www.xeno-canto.org/api/2/recordings?query=gen:Fulica+cnt:japan" 

#get json(requests.get)

response = requests.get(url)
jsonData = response.json()["recordings"]

#Download section(The file name is{id}.mp3)

for data in jsonData:
    url = ("http:"+data["file"]) 
    title = data["id"]
    urllib.request.urlretrieve(url,"{0}.mp3".format(title))

About url

According to the Xeno-canto API documentation, the query is

• gen: common name
• sp : specific name
• ssp : subsupecific name
• en : english name
• cnt : country Etc. can be specified.

Of this time

url = "https://www.xeno-canto.org/api/2/recordings?query=gen:Fulica+cnt:japan" 

Then, it is specified by gen: Fulica, cut: Japan.

About the download section

url = ("http:"+data["file"])

In, the download url is in the key [" file "]. (Since http: is not attached, add it.)

environment

python 3.7.4 (Anaconda) MacOS Catalina 10.15.5