A web interface for pdf computer vision tests.
- Ubuntu >= 14.10
- Postgresql >= 9.4
- Python >= 3.5
First install computer vision script at https://github.com/rich-hart/inspection, then install the webserver:
git clone https://github.com/rich-hart/test_auto_ui.git
cd test_auto_ui
virtualenv -p $(which python3.5) venv
source venv/bin/activate
pip install -U pip
pip install -r requirements/dev.txt
createdb -U postgres qa_automation
./manage.py migrate
./manage.py loaddata fixture.yaml
./manage.py createsuperuser
./manage.py runserver
In another terminal go to the project root directory and start the worker queue:
export DJANGO_SETTINGS_MODULE=test_auto_ui.settings
celery -A inspection worker -l info
Upload pdfs to /api/pdfs/
Create new options at /api/options/
Create computer vision tests at /api/cvtests/
To modify or delete pdfs, options, or tests locate the specify id of the object and add it to the end of the url like so: /api/{object}/{id}
e.g. /api/cvtests/27
.
The tool works best on pdfs that have had minor to moderate alterations to them. Tests on PDFs that are around several hundred pages long can take a few hours to run. PDFs over 1000 pages should be run over night.
Computer vision tests are run by a worker queue so several tests can initiated at the same time.