Max Gordon is a researcher at the Karolinska Institute who works with developing a deep learning algorithm for interpreting orthopedic X-ray images. It is using radiologist’s reports in order to deduce labels that are of interest such as presence of fracture, osteoarthritis and other features. The previous image viewer’s interface Swagger had issues in terms of managing the stored data. This could potentially be solved by switching to GraphQL according to Max Gordon. On request by Max Gordon studies were conducted in order to conclude if Swagger or GraphQL was more compatible with the medical CE-marking and which was more suited for expanding the system. The studies mainly consisted of measuring interpretability at different target groups and how they faced Swedish law when it came down to handling personal data. Based on the qualitative and quantitative studies, the conclusion was drawn that Swagger was more compatible with the medical CE-marking and expansion of the image viewer. / Max Gordon är en forskare på Karolinska Institutet som arbetar med att utveckla en algoritm och tillämpning vilket kan tolka röntgenbilder. Denna används tillsammans med rapporter från radiologer för att kunna utreda röntgenbilders egenskaper. Den föregående bildvisarens gränssnittsspecifikation Swagger hade problem med att hantera den lagrade datan, detta kunde potentiellt lösas genom att använda GraphQL, enligt Max Gordon. På begäran av Max Gordon skulle kompabiliteten mellan GraphQL och CE-märket utredas. För att kunna dra slutsatsen om GraphQL eller Swagger var mest lämplig för CE-märket och vilken var mest anpassad för vidareutveckling av bildvisaren i framtiden. Detta genom att jämföra resultatet av fallstudier på Swagger och GraphQL som huvudsakligen undersökte hur god språkens tolkbarhet var hos olika målgrupper, samt hur hanteringen av persondata förhåller sig till den svenska lagstiftningen. Baserat på de kvantitativa samt kvalitativa undersökningarna ansågs Swagger mest kompatibelt med CE-märkets krav och vidareutveckling av bildvisaren.

This thesis deals with the design and implementation of a tool for generating parts of client applications in TypeScript language from the description of a REST interface. The goal of the tool is an automatic generation of an application skeleton which uses the Angular framework including data structures and services enabling access to server endpoints. The thesis describes the interfaces based on the REST architectural style, as well as the technologies used for their description. The tool was implemented in TypeScript language and uses the Node.js runtime. The output file generation is based on the Mustache.js template system. The result is a tool which allows the generation of source code based on the provided WADL or OpenAPI interface description, which can be simply used as a part of an Angular application.

Application Programming Interface (API) development is growing in popularity and because of that, it is more important than ever to develop high-quality APIs. There is often a rush to build APIs and therefore their design can suffer consequently. This may have a negative impact on the companies and consequently a poor user experience. Different “first” methodologies decide the approach when it comes to developing APIs. The purpose of this study is to investigate what API-first methodology has for value when developing APIs. This is done by developing two APIs where one is using the API-first approach while the other makes use of Code-first. This thesis is using a qualitative method of semi-structured interviews to collect data, with the focus to highlight the differences between the APIs developed. Specifically, the differences in test efficiency, user experience, and API correctness.  The conclusion from the study points out that by using API-first the user experience gets a higher quality in terms of documentation, test efficiency, and accuracy in the business requirements. Also, it provides a suggestion for when and why a company should use it. The findings suggest that when the project is big and complex, then API-first is a good option to consider. The continuous dialog between developer and client is why it is such an approving method to use. The hope is that the findings of this thesis are an indicator that API-first is a significant method to use and that more companies consider it when developing APIs.

RESTful services are becoming increasingly popular. This work, that was carried out at Scania IT, investigates how a RESTful service should be designed and specified so that it meets the demands of an enterprise environment. In particular, this report will focus on documentation and validation of RESTful services, i.e. specifications. These are important aspects in an enterprise environment. The investigation consists of a comparative case study of four specification formats and Hypermedia As The Engine Of Application State (HATEOAS). The results show that the most commonly used specification formats have flaws. They also suggest that Swagger and RAML (RESTful API Modeling Language) are the two most mature formats. The results also show that HATEOAS, which is a more dynamic approach, can be useful in an enterprise environment but that it requires careful design. / REST-tjänster blir alltmer populära. Detta arbete, som utfördes på Scania IT, undersöker hur en REST-tjänst bör designas och specifieras för att möta de krav som finns i en enterprise-miljö. Specifikt kommer denna rapport att fokusera på dokumentation och validering av REST-tjänster, med andra ord specifikationer. Dessa är viktiga aspekter i en enterprise-miljö. Undersökningen består av en jämförande fallstudie med fyra specifikationsformat samt Hypermedia As The Engine Of Application State (HATEOAS). Resultaten visar att de mest vanligt förekommande specifikationsformaten har brister. De antyder också att Swagger och RAML (RESTful API Modeling Language) är de två mest mogna formaten. Resultaten visar också att HATEOAS, som är ett mer dynamiskt angreppssätt, kan vara användbart i en enterprise-miljö men att det kräver omsorgsfull design.

No description available.

DH Anticounterfeit, a company working against counterfeit products, has a mobile application that lacks a proper back-end system to handle and store DH Anticounterfeits data. There are plenty of ways to go about to x this problem, but this thesis is focused on the development of a REST API that will be replacing the non-existent system. The main goal is to create the API and run it within a containerized environment using Docker. The whole system should run on Amazon Web Services, that will be working as a server to keep the system going at all times. Nowadays, security is important due to plenty of ways to attack a system, which is why security and authorization will be a high priority. The result will be a well-functioning, high security system that uses MongoDB database to store and process data in a secure way.

Tato diplomová práce popisuje testování REST API rozhraní aplikací. Výsledkem práce je aplikace Restty, založená na použití nástroje Swagger, která umožňuje testovat jednotlivé části API aplikací, i vytvářet a spouštět komplexní testovací scénáře nad daným rozhraním. Teoretická část práce vysvětluje principy webových služeb, vzdálených rozhraní a představuje nástroj Swagger i technologie použité k implementaci. V následující kapitole jsou v práci prezentovány designové návrhy aplikace, na které plynule navazuje kapitola o implementaci a testování, pro které je zvoleno rozhraní nástroje Red Hat JBoss BPM Suite. V závěru práce jsou vyhodnoceny přínosy aplikace Restty a navrženy případné budoucí rozšíření.

Táto práca sa zaoberá fuzz testovaním REST API. Po prezentovaní prehľadu techník používaných pri fuzz testovaní a posúdení aktuálnych nástrojov a výskumu zameraného na REST API fuzz testovanie, sme pristúpili k návrhu a implementácii nášho REST API fuzzeru. Základom nášho riešenia je odvodzovanie závislostí z OpenAPI formátu popisu REST API, umožňujúce stavové testovanie aplikácie. Náš fuzzer minimalizuje počet po sebe nasledujúcich 404 odpovedí od aplikácie a testuje aplikáciu viac do hĺbky. Problém prehľadávania dostupných stavov aplikácie je riešený pomocou usporiadania závislostí tak, aby sa maximalizovala pravdepodobnosť získania potrebných vstupných dát pre povinné parametre, v kombinácii s rozhodovaním, ktoré povinné parametre môžu využívať aj náhodne generované hodnoty. Implementácia je rozšírením Schemathesis projektu, ktorý generuje vstupy za pomoci Hypothesis knižnice. Implementovaný fuzzer je použitý na testovanie Red Hat Insights aplikácie, kde našiel 32 chýb, z čoho jednu chybu je možné reprodukovať len za pomoci stavového testovania.