Avtomatiziran sistem za priporočanje receptov iz spletišč

Pečnik, Matej (avtor), Smrdel, Aleš (mentor)

V diplomskem delu predstavimo razvoj in delovanje avtomatiziranega sistema, katerega glavna naloga je priporočanje receptov na podlagi izdelkov in sestavin, ki jih imamo v shrambi. V ta namen smo razvili komponente strežniškega dela ter spletno in mobilno aplikacijo. Ena izmed komponent je iskalnik sestavin, ki s pomočjo drevesa sestavin za določen vhod ugotovi, za katero sestavino gre. Za pridobitev podatkov o izdelkih in receptih sistem uporablja luščilnik, ki iz nekega spletnega vira pridobi podatke ter jih strukturirano shrani v podatkovno zbirko. Poleg luščilnika vsebuje sistem tudi iskalnik receptov in iskalnik izdelkov na računu. Prvi skrbi za izpis vseh receptov, ki jih lahko naredimo iz sestavin in izdelkov, ki jih imamo. Drugi pa poišče vse izdelke, ki smo jih pridobili s pomočjo optičnega prepoznavanja znakov skeniranega računa. Razvili smo tudi aplikacijski programski vmesnik, preko katerega strežniški del komunicira z odjemalci. Spletna aplikacija omogoča različne funkcionalnosti kot so: pregled izdelkov in sestavin shrambe, pregled receptov, pregled priljubljenih receptov. Poleg tega omogoča tudi dodajanja izdelkov in sestavin na različne načine. Mobilna aplikacija poleg vseh funkcionalnosti, ki jih omogoča spletna aplikacija, omogoča še dodajanje izdelkov po črtni kodi in po računu. Obe funkcionalnosti uporabljata vgrajeno kamero mobilne naprave.

• Jezik: Slovenski jezik
• Ključne besede: spletna aplikacija, mobilna aplikacija, OCR, priporočanje receptov, luščilnik podatkov, API, drevo sestavin
• Vrsta gradiva: Diplomsko delo/naloga
• Organizacija: FRI - Fakulteta za računalništvo in informatiko
• Leto izida: 2016
• PID: 20.500.12556/RUL-84531
• Datum objave v RUL: 29.08.2016

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Automated system to recommend recipes from websites

Izvleček

In our diploma thesis we present the development and functioning of automated system, which main task is to recommend the recipes on the basis of the products and ingredients, which we have in our pantry. For this purpose we developed the components of the server part as well as web and mobile application. One of the components is the search engine of the ingredients, which, by means of the tree of ingredients, for a specific input determines which is the ingredient. In order to obtain the data the system uses the scraper which acquires data from a web source and saves them in the database in a structured way. In addition to the scraper, the system also contains the search engine of the recipes and the search engine of the products within the bill. The first one is responsible for the display of all the recipes which can be made of the ingredients and products which we have. The other one determines all the products which we acquired by means of optical character recognition of the characters of the scanned bill. We also developed application programming interface through which the server part communicates with the clients. Web application enables different functionalities, such as: review of the products and ingredients of the pantry, review of the recipes, review of the favorite recipes. In addition, also enables adding the products and ingredients in different ways. Mobile application in addtion to all the functionalities enabled by web application, it also enables adding products by means of a bar code and by means of a bill. Both functionalities use the built-in camera of the mobile device.

• Ključne besede: web application, mobile application, OCR, recommending recipes, data scraper, API, tree of ingredients