image-dataset

README

image-dataset

face images

collect images from Twitter

export CONSUMER_KEY='*********************'
export CONSUMER_SECRET='****************************************'
go run cmd/collect_twitter_images/main.go -screen_name <Twitter ScreenName> > python/data.tsv

detect faces and save metadata

cd python
pip install -r requirements.txt
# download a trained facial shape predictor for dlib
python create_data.py data.tsv

upload data

go run cmd/upload_images/*.go -datadir python/data -projectID <Project ID>

development

gcloud beta emulators firestore --project <Project ID> start --host-port :8081

export FIRESTORE_EMULATOR_HOST=localhost:8081
export GOOGLE_CLOUD_PROJECT=<Project ID>
export GOOGLE_APPLICATION_CREDENTIALS=<path/to/serviceAccountKey.json>
export SESSION_KEY=<secret key>
go run web/main.go

cd frontend
npm start

deployment

(
    cd frontend
    npm run build
)
go run cmd/generate_index/main.go > firestore.indexes.json
firebase deploy --only firestore:indexes
gcloud app deploy web

dump images

go run cmd/dump_data/main.go -projectID <Project ID> -size 500 -num 10000 -status OK