Dog API Prodigy recipe

dog_image_recipe.py

import prodigy
import requests
import random
from typing import List
from prodigy.components.preprocess import fetch_media
from prodigy.util import split_string
from prodigy import set_hashes

def get_stream(labels):
    while True:
        images = []
        for label in labels:
            res = requests.get("https://dog.ceo/api/breed/" + label + "/images/random/10").json()
            for i in res["message"]:
                images.append({"image": i, "meta": {"breed": label}})

        random.shuffle(images)

        for i in images:
            yield {"image": i["image"], "meta": {"breed": i["meta"]["breed"]}} 

@prodigy.recipe(
    "dog-image-recipe",
    dataset=("Dataset to save answers to", "positional", None, str),
    label=("One or more comma-separated labels", "positional", "l",  split_string),
)
def dog_image_recipe(
        dataset: str, 
        label: List[str],
    ):

    stream = get_stream(label)
    stream = (set_hashes(eg) for eg in stream)

    return {
        "dataset": dataset,
        "view_id": "image_manual",
        "stream": fetch_media(stream, ["image"], skip=True),
        "config": { 
            "labels": label, 
        },
    }