Google Summer of Code 2020 Weekly Feedback

What progress have you made this week?

This week I found CMS Data of 3000 Hospitals both inpatient & outpatient , I am searching for more datasets available by states & I have found it for California, New York, Indiana.

I have given my mentors a detailed project plan with milestones & I will discuss it maybe tomorrow via call.

What do I plan to do next week?

I will start making App UI with dummy data in next two weeks as suggested by @muarachmann , I will also scrap Hospitals CDM of California State next week.

Have you had any blockers or issues that are impeding your project?

No

What progress have you made this week?

In the fourth week of the community bonding (May 23 - May 29), I tried to communicate the sample app with the test server. I added the code to scan a QR code and make a connection with the test server and it was successful. I also tried looking into UI/UX design guidelines, to prepare a good UI for the current project.

The successful connection with the test server was done using native Android app written in Java. I am still trying to get that working in Flutter and have spent quite a bit of time working on that and is not successful so far.

What do I plan to do next week?

  • I will go through the spreadsheet, that @Maria-Freytsis has shared.
  • Come up with a plan with milestones and discuss with the mentors regarding that.
  • Look into app store for some inspiration for the UI and propose a UI for the app.

Have you had any blockers or issues that are impeding your project?

The linking of the Evernym SDK with the mobile app is time consuming as of now, with not much resource. So that’s the only issues which I’m facing.

1 Like

Week 4 : Community Bonding

What progress have you made this week?

  1. Finalized datasets, models, compression methods and system architecture.
  2. Created my first MR in the repo.

What do I plan to do next week?

  1. Perform EDA on the dataset and begin training.
  2. Look into Ultrasound datasets and papers.

Have you had any blockers or issues that are impeding your project?

The process of integration of the Flask app inside the Raspberry Pi is challenging and needs to be more defined. Either it will be possible or the system architecture would have to be modified.

"23-may-2020 to 29-may-2020"

[ Progress ]

  1. Completed the bounding box component part of the UI.
  2. Integrated Reaact MUI library with my CRA.
  3. Decided the palette.

[ Plan before June Starts ]

  1. Clean my frontend code.
  2. Start with training the model.
  3. Read about model versioning and minibatching with tensorflow-serving.

[Blockers]

N/A

This week’s progress (05/25 - 05/29)

  1. Refreshed my skills on components creation with polymer 3.
  2. Set up the environment for the start of the coding session. Obtained access to the previous lh-toolkit GitLab projects.
  3. Posted a blog.

Next week’s Plan (06/01 - 06/05)

  1. Create at least three web components (fhir-observation-status, fhir-observation-category, and fhir-observation-period )
  2. Perform unit testing on the components.

Shouldn’t you be integrating with the mHBS app and use the same style for the UI? @iainbarclay @shbucher

Is your plan still to build a separate UI? Please look at the OHIF Viewer and try to integrate your work into that, as that is part of the lh-radiology. Please don’t start from scratch. That is not how FOSS should be done.

1 Like

There are two Dockerfiles, one for dev and one that makes a production bundle (It’s a meteor.js app)

I should be integrating with the mHBS app, but I was told that it would be a standalone application for now, as the main work is to integrate with the evernym SDK and get that part of the project going. @iainbarclay correct me if I’m wrong. However, @shbucher mentioned in the video call recently that I still have some pending work on the mHBS app and that I would be contacted by @sunbiz.

As far as UI goes, the colour schema would still be followed similar to the mHBS app.

1 Like

@sunbiz @r0bby , I have looked into OHIF viewer demo app. It is great, but what I understand from the chat I had with @judywawira and @pri2si17 the other day, there is a lack of annotated medical images on which object detection can be performed. So we use heatmaps and other tricks to resolve that issue. We propose to make a simple webapp for AI guided annotations. The doctors who will be doing annotations will get bounding boxes from our AI as a starting reference and they can add more bounding boxes, delete, reshape or change positions if they like. Now, for Machine learning we need the bounding Boxes and nothing else. The OHIF viewer on the other hand is for detailed analysis and data recording of radiology images for medical research. I think our app is solely for data collection for ML research so we should have a separate app.

Also, I have some very specific features in my mind for the above application to make the process more efficient and effective.

Yes @prajwal27 , you are correct - the app you are working on at the moment will be used directly on the smartphones of mothers/guardians, so its good for you and @sradha to look into the app stores to get a sense of the “state of the art” for UI design and flow for this type of “baby management” app to inform our direction.

Ultimately we hope that this Guardian app will become part of the mHBS suite, so we’ll definitely be very mindful of the UI work done there already, and propose a direction for this UI that fits well into the mHBS family, and brings best practise from the UI of well regarded apps aimed at use directly by mothers too. [cc @Maria-Freytsis @sunbiz @shbucher @shbucher ]

2 Likes

Thanks Iain. @prajwal27 I am preparing some prototype screens which we can use as a starting point for discussion. I will have the page descriptions and prototype screens soon and will setup a call with you and Iain to discuss.

1 Like

Keep chatter on this thread to a minimum, please. If @mentors have a concern with reports here, please bring it up your project threads.

No need to reply.

1 Like

What progress have you made this week?

This week I made UI screens of app with dummy data, I made Home, Search Screens & other screens , I added Navigation Drawer & bottom navigation UI.

I made a merge request of it & it was successfully merged. I am also scraping CDM of Alaska State , I have already created two issues in Repo (Link), I will make its PR by Sunday

What do I plan to do next week?

I will make UI of rest of the features of app & will scrap CDM of California State

Have you had any blockers or issues that are impeding your project?

No

1 Like

What progress have you made this week?

In the first week of coding period (May 30 - June 5), I have gone through the User Stories which @Maria-Freytsis had shared and also had discussions regarding that. I pushed my code to gitlab with few commits.

I have started working on accpeting the credential which the issuer will provide. I worked on getting messages from the cloud agent, as we will not be using Push Notification (work under progress from Evernym side). I have been constantly interacting with the Evernym team, I also had a video call with one of the Mobile Developers to understand the working of the SDK (very useful).

I also started to maintain documents ( WIKI ) of the responses I got from the SDK. However, the plan is to maintain a wiki consisting of all the discussions, important points, decisions taken to help the developers in future.

What do I plan to do next week?

  • @sradha has shared the UI design and I will start developing those screens.
  • Work on credentials ( receiving and accepting ) given from the issuer.

Have you had any blockers or issues that are impeding your project?

I was initially stuck with issues and doubts in the SDK, but luckily received help from the Evernym team and I believe I have made decent progress.

2 Likes

Week 1 : Coding Period

What progress have you made this week?

  1. Finished EDA on RSNA dataset.
  2. Completed extraction of useful data items and began training.

What do I plan to do next week?

Complete training of models.

Have you had any blockers or issues that are impeding your project?

Ran out of memory errors on my local machine as well as Google Colab due to the masssive dataset size. I am currently implementing mini-batching during training to overcome this error.

2 Likes

"30-may-2020 to 5-june-2020"

[ Progress ]

  1. Made the APIs for the django app.
  2. Generated API documentation from Postman. “Link

[Next Week]

  1. Start the Training.
  2. Setup tf server.
  3. Connect Django and tf Serving.

[Blockers]

N/A

2 Likes

This week’s Plan (06/01 - 06/05)

  1. I created the fhir-observation-status using polymer/lit-element v0.7.1
  2. My mentor suggested that I upgrade all the previous elements from polymer/lit-element version 0.5.1 to lit-element v2.3.1. I started upgrading the elements as instructed.

Next week’s Plan (06/08 - 06/12)

  1. Finish upgrading the elements to lit-element v2.3.1
  2. Create 3 new web components with lit-element v2.3.1 if I complete the upgrade early enough

Blockers

Components not rendering as desired after upgrade. Currently trying to figure out a solution.

2 Likes

"6-june-2020 to 12-june-2020"

[ Progress ]

  1. Setup Tensorflow serving for a trained model on Chest Xrays.

  2. Wrote a python script for communication between TF-Serving and Django Rest Framework.

[Next Week]

  1. Make the python middleware Asynchronous.
  2. Write a CLI for the python middleware.

[Blockers]

N/A

2 Likes

What progress have you made this week?

In the second week of coding period (June 6 - June 12), I have developed few of the UI screens of the app and am yet to push the commits. I’m still working on accepting the credential-offer and the credential from the issuer.

What do I plan to do next week?

  • Complete the UI development of the app.
  • Complete the work on credentials given from the issuer.

Have you had any blockers or issues that are impeding your project?

None as of now.

2 Likes