passiolife/flutter-nutrition-ai-module

Plug-and-Play

flutter-nutrition-ai-module's Issues

Adding Ingredient to existing food - should ask to name recipe

Scan foods flow

• Onboarding view
• Result view with Edit and Log buttons
• Item Added to Diary view
• Barcode not recognised view
• Nutrition facts scanner (not yet implemented in the Flutter SDK

Flutter Edit Screen Fixes

• For any food, the heart symbol is highlighted, as if the user has made that food a favourite
• We display open food facts link on all foods (e.g. search "Lemonade Target") - that food is not from open food facts. or Coffee, brewed
• The Open Food Facts link is not clickable. please see iOS for reference implementation
• We are not displaying Brand name
• The More Details link does not lead anywhere

Search screen

• Currently implemented with just the food name and icon

Edit food screen

Diary screen

Refactor the v3 FoodRecord

iOS Food Record
Android Food Record

Issues that are blocking uploading to pub.dev

Package validation found the following errors:

* line 1, column 1 of lib/src/common/util/date_time_utility.dart: This package does not have intl in the `dependencies` section of `pubspec.yaml`.
    ╷
  1 │ import 'package:intl/intl.dart';
    │ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ╵
* line 3, column 1 of lib/src/common/util/double_extensions.dart: This package does not have intl in the `dependencies` section of `pubspec.yaml`.
    ╷
  3 │ import 'package:intl/intl.dart';
    │ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ╵
* line 3, column 1 of lib/src/common/connectors/local_db_connector.dart: This package does not have intl in the `dependencies` section of `pubspec.yaml`.
    ╷
  3 │ import 'package:intl/intl.dart';
    │ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    ╵

* Your package is 188.7 MB. Hosted packages must be smaller than 100 MB.
The size issue is due to the 3 videos in the readme file. Please create lightweight gifs from these videos.

The resources are linked in the example app

The strings and images from the package are not visible in the example app

Create the Flutter Nutrition AI Module plugin

First steps:

• Create an empty plugin project that will represent the module
• The plugin will have to depend on the nutrition_ai plugin
• Copy over to the plugin the Splash screen functionality from the Demo App, that has the ability to configure the SDK
• In the example app of the plugin project create a simple screen which has one button. Clicking the button should open the Splash screen from the plugin

@sahilmind Let's discuss here how to pass the SDK key from the example app to the plugin, to enable configuring the SDK from the app side. Please write a suggestion how to solve this issue

Camera recognition flow

Camera flow:

• Show tutorial the first time a user enters the food scanner screen
• The recognition results from the sdk are divided into 4 cases:

1. If no results are returned from the recognition callback, show the scanning view below the viewfinder

2. If the results is a visual candidate, show the food name and the icon in the bottom sheet (the figma design has 2 names, but the SDK produces only one so we show the one). If the user pulls the bottom sheet up, the alternative list is displayed (again, the alternatives only have one name and the figma design doesn't take this into account).

   To populate the list of alternatives we will use two different types of results: the list of detected candidates and the alternatives of the first detected candidate. The first detected candidate (the one with the highest confidence) is the main result that is show to the user. The alternatives are a list constructed as: other items in the list of detected candidates + alternatives of the first result.

   If the user clicks on an alternative results we navigate to the Edit Food screen.

3. If the results is a product candidate, we fetch the full PassioFoodItem before displaying the item on the screen.

4. If the SDK produces a barcode or packaged food code, but we don't get a result for the nutritional data, we display the barcode not recognized screen.

• Keep in mind that for visual scanning we don't have the full PassioFoodItem object, but for product codes we do. This means that when we navigate to the Edit Food screen or Log a visual food, we then do the networking call and complete the action.