DisasterTweets

DVC

pip install -r requirements_dvc.txt

dvc pull

dvc repro

FastAPI

Run FastApi

pip install -r requirements.txt

uvicorn src.start_api:app

Build & Run docker with fastapi

dvc pull; dvc repro && sudo docker build -t fastapi_tweets -f Dockerfile .

sudo docker run -p 8080:5011 -v /{full path to project}/logs/:/app/logs/ fastapi_tweets

Update to actual model & Run docker with fastapi

dvc pull resources/model.pkl && sudo docker run -p 8080:5011 -v /{full path to project}/logs/:/app/logs/ fastapi_tweets

Build & Run Train docker

sudo docker build -t repro -f Dockerfile.repro . && sudo docker run -v /{full path to project}/:/app/ repro

---

DVC base-command

dvc init

dvc remote add -d storage gdrive://YOUR_FOLDER-NAME

dvc add resources/data.txt

dvc run -n split \
    -p split \
    -d resources/data.txt \
    -d requirements.txt \
    -d src/split.py \
    -o resources/train.txt \
    -o resources/test.txt \
    python src/split.py

dvc run -n split -p split -d resources/data.txt -d requirements.txt -d src/split.py -o resources/train.txt -o resources/test.txt python src/split.py

dvc run -n train \
    -p train \
    -d resources/train.txt \
    -d requirements.txt \
    -d src/train.py \
    -o resources/model.txt \
    python src/train.py

dvc run -n train -p train -d resources/train.txt -d requirements.txt -d src/train.py -o resources/model.txt python src/train.py

dvc run -n eval \
    -p eval \
    -d resources/test.txt \
    -d resources/model.txt \
    -d requirements.txt \
    -d src/eval.py \
    -M metrics.yaml \
    python src/eval.py

dvc run -n eval -p eval -d resources/test.txt -d resources/model.txt -d requirements.txt -d src/eval.py -M metrics.yaml python src/eval.py

dvc dag  # Visualize a DVC Pipeline