KatieHYT/cradle-django

cradle-django

Data

Download from https://drive.google.com/drive/folders/1K6-OSrJXWz5ekSmmepAuEYi5iAVy24wa

Settings

• Step-1:
  Enter a docker container using the following docker image:
  nvcr.io/nvidia/pytorch:21.03-py3

  # example command to start a docker container named "my_container"
  # mapping port "200:200" (local machine : inside container)
  # mapping volumn "{your_project_folder_path}:/TOP" (local machine : inside container)
  # using image(nvcr.io/nvidia/pytorch:21.03-py3)
  
  docker run -d -it --name my_container -p 200:200 -v /my_project_folder_path:/TOP nvcr.io/nvidia/pytorch:21.03-py3
  

  # example command to enter the container
  docker exec -it my_container bash
  

• Step-2:
  Go to your working directory and clone this repo.
  Go into the repo.

• Step-3:

  Edit the following information in cradle/settings.py to the paths on your machine.

  • OPENAI_API_KEY
  • GOOGLE_REVIEW_DIR
  • PUBLIC_IP

• Step-4:

  pip install -r requirments.txt
  

• Step-5: Get submodules.

  git submodule update --init --recursive
  

• Step-6: Start the service.

  python manage.py migrate
  # assuming you use port 200
  python manage.py runserver 0.0.0.0:200
  

POST example (python)

import io
import requests

url = "XXX.XXX.XXX.XXX/petlover/callback"

data_dict = {
    "txt": "%petfriendly%  peak design",
}
def get_stream():
    s = requests.Session()

    #with s.post(url, json=data_dict, stream=True, data=io.BytesIO(b'many many bytes')) as resp:
    with s.post(url, json=data_dict, stream=True) as resp:
        for line in resp.iter_lines():
            if line:
                print(line)

get_stream()

POST example (bash)

curl -X POST XXX.XXX.XXX.XXX/petlover/callback \
--header "Content-Type: application/json" \
--data '{"txt": "%petfriendly%  peak design"}'