API input

The payload parameter type in predict(self, payload) can vary based on the content type of the request and can be parsed according to the Content-Type header in the request. The parsing rules are as follows:
For header Content-Type: application/json, thepayload parameter will be the parsed JSON body.
For Content-Type: multipart/form-data / Content-Type: application/x-www-form-urlencoded, payload will be starlette.datastructures.FormData (key-value pairs where the values are strings for text data, or starlette.datastructures.UploadFile for file uploads; see Starlette's documentation).
For all other Content-Type values, payload will be the raw bytes of the request body.
​
JSON data
Request
$ curl https://***.syndic.ai \
    -X POST \
    -H "Content-Type: application/json" \
    -d '{"key": "value"}'
Inside Model
class PythonPredictor:
    def __init__(self, config):
        pass
​
    def predict(self, payload):
        print(payload["key"])  # prints "value"
Sample models
​Yolov5​
​Face blurring​
​Image Classifier​
Form data
Sample request
$ curl https://***.c1.syndic.ai \
    -X POST \
    -F "[email protected]" \
    -F "[email protected]" \
    -F "[email protected]"
Inside Model
from PIL import Image
import pickle
​
class PythonPredictor:
    def __init__(self, config):
        pass
​
    def predict(self, payload):
        text = payload["text"].file.read()
        print(text.decode("utf-8"))  # prints the contents of text.txt
​
        obj = pickle.load(payload["object"].file)
        print(obj["key"])  # prints "value" assuming `object.pkl` is a pickled dictionary {"key": "value"}
​
        img = Image.open(payload["image"].file)
        print(img.size)  # prints the dimensions of image.png
Sample models
​Deepface​
​
Last modified 1yr ago