gagamil/unolims

Base LIMS app for sample testing (two-stage hierarchical pool testing).

unolims

Basic Laboratory Information Management System web app for sample testing using two-stage hierarchical pool testing techinque. Built with Django, signals for main logic operation. This app is developped as a monolithic app. However the Signals framework does offer a possibility to develop a loosely coupled Django project. This way decoupling one of the apps should be a fairly easy job.

licence

MIT

motivation

LIMS might be expensive to build. My experience shows that multiple design flaws and over engineerig can geopardize the LIMS adoption and basically screw the whole workflow. The basics of hierarchical pooling is more or less the same and differs in the details. These detail can be added later on. Provided that this software would be relatively easy to deploy and modify this LIMS would be a good tool for the labs.

Two stage hierarchical pool testing

1. All samples are placed in their respective tubes.
2. These samples are then amalgamated into a number of pools using pooling tubes.
3. Pooling tubes are then being tested.
4. If pool tested negative, the respective individual samples are flagged negative.
5. If pool tested positive, the respective individual samples are being tested individually.

App design

Structural separation by apps.

common

Constants, signals and other names and objects used throughout the apps.

data_importing

Data import services and logic. For Tubes and Runs.

run

All Run, Run Template, Run Results data models and logic.

tubes

All Tube and Batch data models and logic.

web

The web app UI. A hybrid Django with templates configuration. The ecnhanced UX is built using React Js. Whenever possible basic Template based Django Views are used. The React Js app is being used when UX improvement is evident.

Tubes and Batches

In order to be a bit more universal the word Batch is used. Words Goup or pool might be used as well. Tubes are grouped into batches for different reasons. An example is when they are grouped on a rack for scanning and the data is entered into the system. Or when the samples are run and tested.

Batch

Tags

To run the logic and display the data to the user Batch may hold any number of tags. These could be: 'INDIVIDUAL POOLS AND POOL TUBE', 'POOLING SCAN', 'TEST RUN 123', 'CONFIRMED', 'COMPLETE', 'CANCELLED'. Business logic might interpret tags in one way or another with the subsequent outcome.

Extra data

There might be some extra data that needs to be stored and retreived later on. For this reason the Batch contains a JSON field 'xtra_data'.

Events and handlers

Each main model and step triggers a predefined Django signal:

• batch_pending_confirmation
• batch_just_confirmed

to be continued...

unolims's Issues

API endpoint to add run result

Log user and system activity on the objects

https://github.com/justquick/django-activity-stream

API endpoint to add tube batch

Add CI/CD (Github Actions)

Add pagination fo batch list

Up to 25 items per page

Add a common app to hold signals for various apps to use

Basically decouple the project apps.

Import Run results file

csv file with the results
Parse and populate the results

Fix model clean method

Uploading the Tube batch file doesn't work.
The clean method can't pass the correct file to the service.

Limit access to only auth users

Mock data creation script adds wrong batch tags

"Pool Batch" from const needed iso "PoolScan" as string literal

Add app that will handle publishing of the results

Create app that will run the publishing.
Basically 2 main ideas here:

1. Show per client basis
2. Send results to the client

Run results display & analyst page

Allow view of each sample.
Allow finalization that will create/update and set results values for the sample (if pooling tube then for that one).

Update batch title

Add batch form to edit main data including title.

Create TubeBatches from imported data

Data is imported from the frontend app form into the data importing app.
Need to send an event so that the tube app would populate the batch using the provided data.

Don't use tags to tag tube batch type - use choices or bool flag

Configure logging

Don't use print. Use logging lib.
Even better - connect to a logging service.

Add batch timestamp to list page

Probably to tests too.

Add basic tube usability statuses

usability statuses:

• OK
• LOW LEVEL
• CONTAMINATED
• ERROR (generic)
  Maybe make these dynamic?

Enable 12 factor principles

Update tag names

Use the one in const module.
Update the tests + batch list page.

Dockerize

Web form to import run result file

Add JS app support

JS apps will improve UX.
ReactJS library is suitable.
Should be per page js app which hooks when the page is being loaded.
Use web pack to build the app and put to Django static folder the output.

Run creation page

• pick the configuration (run method)
• pick the run tube batches
• set custom run name

When validated allow csv template file to be downloaded

Deploy to DigitalOcean

Or GCP CloudRun?