App description

App allows creating shopping lists and changing list item status. When user provides name of shop which is added to database products, app displays products sorted in order which minimizes the distance.

In order to select the order providing the shortest route app predicts item type from item name. Predicting item type is performed using model generated from dataset containing pairs of 'product type - product name'. Predicted item types are matched with waypoints present in database (x,y coordinates mapping to location in specific shop where product type can be found).

After selecting locations where products can be found items are sorted, app is searching for order where total distance is the shortest. Selecting the shortest route is performed with use of "Google.OrTools"

ML model predicting item type is generated with use of dataset containing product names in Polish. To use product names in different languages, different dataset should be used.

Running the app

requirements: docker engine, available port 443 and 80. To run the app in local environment, in directory containing docker-compose.yaml, enter in terminal: docker-compose build, and then docker-compose up. After displaying info: Microsoft.Hosting.Lifetime[0] Now listening on: https://[::]:443 info: Microsoft.Hosting.Lifetime[0] Now listening on: http://[::]:80 app should be available in the browser at localhost.

The format of JSON string inserted into ShopWaypointsReadDtoJSON column of ShopWaypointsEntities should match string obtained by using JsonSerializer.Serialize(ShoppingList.Data.Waypoints.WaypointsRepoHardcoded.HardcodedWaypoints).

To use different dataset for generating ML model, put training and test dataset into ./GroceryClassification/Data directory and adjust file names in the Main method of ./GroceryClassification/Program.cs.

Setting up certificate in Powershell ./shoppingList>

dotnet dev-certs https -ep $env:USERPROFILE\.aspnet\https\ShoppingList.pfx -p password_here

Should return The HTTPS developer certificate was generated successfully.

dotnet user-secrets set "Kestrel:Certificates:Development:Password" "password_here"

Should return Successfully saved Kestrel:Certificates:Development:Password = password_here to the secret store.

Example data

SqlShoppingListRepoExampleData.cs injected in ConfigureServices method of ./ShoppingList/Startup.cs inserts example shop's waypoints and shopping list when ShoppingListEntities table is empty. To run the app without example data, replace services.AddTransient<IShoppingListRepo, SqlShoppingListRepoExampleData>();

line in ./ShoppingList/Startup.cs with

services.AddTransient<IShoppingListRepo, SqlShoppingListRepo>();

SqlShoppingListRepoExampleData.cs inserts shop named big-market and example shopping list with id of 1 and password: password

TODOs

• Replace plain text passwords
• Remove dataset labels with too few entries

Opis aplikacji

Aplikacja pozwala na tworzenie listy zakupowej i oznaczania/zmiany statusu przedmiotów z listy. Jeżeli do listy zostanie dodana nazwa sklepu, którego mapa/działy sklepowe znajdują się w bazie danych aplikacja wyznacza optymalną kolejność zbierania produktów z listy, tak aby przebyta droga była jak najkrótsza (kolejność wyznaczana z użyciem biblioteki Google.OrTools). Jeżeli w bazie danych znajduje się mapa sklepu, aplikacja prognozuje w jakim dziale sklepu prawdopodobnie znajduje się produkt.

Prognoza odbywa się na podstawie modelu wygenerowanego z użyciem bazy nazw produktu powiązanych z nazwą produktu (model wygenerowany z użyciem biblioteki Microsoft.ML). Baza danych produktów użyta do wygenerowania modelu: https://www.kaggle.com/agatii/total-sale-2018-yearly-data-of-grocery-shop