Implement FileForms - Githubissues

rji-futures-lab / django-fcc-opif

For extracting and analyzing data from the FCC Public Inspection files

0 stars 0 forks source link

Implement FileForms #47

Open gordonje opened 4 years ago

gordonje commented 4 years ago

[ ] Define a FileForm model
- [ ] Add name as CharField
- [ ] Add created_at as a DateTimeField
- [ ] Add last_updated_at as a DateTimeField
- [ ] Add boxes as a JSONField
- [ ] Add ocr_data as JSONField
- [ ] Add proto_file_id as CharField (which holds a FacilityFile.file_id or CableSystemFile.file_id)
[ ] Define list view to see all FileForm instances
[ ] Define create and edit view for FileForm instances
- [ ] In create view, user only needs to fill in a unique name (created_at should auto-populate)
- [ ] In edit view, user needs to load a file image and draw boxes around text to be extracted.
  - [ ] File loading could be implemented as a simple text field to drop in the URL. Slightly fancier UI could be something a auto-complete picker field.
  - [ ] On save, we check if proto_file_id is already populated. If so, we create a new FileForm instance.
  - [ ] Also set boxes
  - [ ] We can try calling [Textract] via boto3, but if we don't get response within 30 seconds, we suspect API Gateway will throw a timeout.
    - [ ] If so, then maybe this can be an async call. Not sure if the lambda will stay alive after the view returns an HTTP response.
    - [ ] If it doesn't, then we will need to deal with getting the OCR data separately, possibly in a management command running on a schedule.