API Pipeline

A REST API to pilot the Ecotaxa pipelines

requirement

pip install fastapi
pip install pydantic
pip install "uvicorn[standard]"
<!-- pip3 install opencv-python -->
pip install opencv-python
pip install requests
pip install numpy

run it to dev

uvicorn main:app --reload

run it to VENV

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
#pip install uvicorn  # if issue run manually uviicorn installation
# uvicorn main:app --reload
python -m uvicorn main:app --reload

online docs

http://127.0.0.1:8000/docs http://127.0.0.1:8000/redoc

The gateway API need other containeurs

run DEEP-OC-multi_plankton_separation container

ssh niko
cd ~/complex/DEEP-OC-multi_plankton_separation

build the container

docker build -t deephdc/uc-ecotaxa-deep-oc-multi_plankton_separation .

run it to have interactive shell

~/complex/DEEP-OC-multi_plankton_separation$ 
docker run -ti -p 5000:5000 -p 6006:6006 -p 8888:8888 deephdc/uc-ecotaxa-deep-oc-multi_plankton_separation 

run it in detach mode

~/complex/DEEP-OC-multi_plankton_separation$ 
docker run -d -ti -p 5000:5000 -p 6006:6006 -p 8888:8888 deephdc/uc-ecotaxa-deep-oc-multi_plankton_separation 

change local port 8888 to 8889, because already use by another container, which ????

~/complex/DEEP-OC-multi_plankton_separation$ 
docker run -d -ti -p 5000:5000 -p 6006:6006 -p 8889:8888 deephdc/uc-ecotaxa-deep-oc-multi_plankton_separation 

Test with Thunder Client (obsolette)

In VSCode Thumder plugin, use the collection named Happy Pipeline

test happy pipeline

• to see if server alive test DEEP plankton separation - niko alive ? : to see if multi_plankton_separation is alive on Niko server

Run the gateway API in Docker

Built it

docker build -t gateway_api .

Run it

docker run -p 8000:8000 -v /Users/sebastiengalvagno/piqv/plankton/:/app/data --name happy_pipeline gateway_api

when use Docker use .env.production else .env.development

at this moment, I can't run in docker because the path are linked to other apps then it can't find the file

a possibility, to use docker, is to make the very long local path in the container and use .env.development

use the Swagger UI

open http://localhost:5000/ui

through the VPN (niko run)

make a tunnel before open on with local address, port do not pass throught the VPN

ssh  -f niko -L 5001:localhost:5000 -N
open http://localhost:5001/ui

new container

cd complex
git clone https://github.com/ai4os-hub/zooprocess-multiple-separator.git
cd zooprocess-multiple-separator/
docker build -t zooprocess-multiple-separator:1 .
docker build --no-cache -t zooprocess-multiple-separator:lastest .

docker run -d -ti -p 5000:5000 -p 6006:6006 -p 8888:8888  zooprocess-multiple-separator:1

no detach mode to see error

docker run -ti -p 5000:5000 -p 6006:6006 -p 8888:8888  zooprocess-multiple-separator:1

to generate a py client based on openapi.json

openapigenerator generate -i http://localhost:5000/openapi.json -g python -o ./src

run unit test

python3 -m venv test_venv
source test_venv/bin/activate
pip install --upgrade pip 
pip install -r requirements.txt

python3 -m unittest discover tests

or one test

python -m unittest tests/*.py 

add -v for verbose mode, and ahave list of test functions

Run only on test function

python -m unittest tests.test_server.Test_server.test_dbserver_withconfig