Hello and welcome to this last week of the Introduction to a Web of Linked Data. This week is about Integration with Other Data Format and Sources. The reason for that is that the Web is evolving toward all form of data toward the integration of all kinds of sources and all kinds of formats. This week is divided into six parts. The first part is about our RDFa: an RDF syntax inside HTML. The second part is about GRDDL, extracting RDF from XML or HTML. The third part is about JSON-LD, a JSON syntax for RDF. Then we will see how we can transform tabular data and metadata like CSV into RDF. A number of sources of data are in fact databases R2RML will be explored in the fifth spot to see how we can export data from a database to RDF and finally we will conclude with LDP which is a REST API to linked data. Let us start with RDFa which is as I said a syntax for RDF inside HTML. When you consider an HTML page, you can find many kinds of data inside. Here you have an example of a page about the book, "The Man Who Mistook His Wife For a Hat," and if you take a closer look, you will see that we have many kinds of data inside that could be interesting for other application, but they don't really have access to that data because it's inside the page. The idea with RDFa is that you can inject markups inside the HTML in order to say that inside this HTML page there are data. For instance here, we have markup seeing that a given part of the text is in fact the title of the book or another part is the creator of the book. To do that, RDFa defines a number of attributes. There are two parts in RDFa: RDFa Light and RDFa Core. The light version actually has the attributes that are the most commonly used such as vocab to define a default vocabulary for section. Prefix to declare other vocabularies. Resource to identify a resource. Typeof to type or resource using common vocabulary and property which is in fact a way to do a link to a value or to a resource. RDFa Core adds additional attributes such as the content attribute to provide a specific value. The datatype attribute to type values. The about attribute to change the subject of a property or the rel attribute to decompose object properties in list. Let us now consider the example of HTML that contains RDFa markup. If I load this HTML in a browser, I will just see the content of the page as usual. In fact, the gray part which is the RDFa content is mostly ignored by the browser. But these part which is now in red can be extracted by an RDFa parser and transformed into a number of triples that you can see here in the total syntax. So, this HTML is both a document that can be seen in a browser, and a data source that can be obtained by parsing using an RDFa parser this document. More precisely if you look at the content, we use the vocab attribute to introduce the schemas, the vocabulary that are used. Here we use two of them. One to describe document, and one to describe a person. Then we can introduce the subject of the triples, the resources that are going to be described. Here we have a resource which is the book and a resource which is the person. The type of the resource person is actually introduced using the typeof attributes. Then, we can add properties to these resources. We show here three properties and their values which are literal values. The title, the date of creation, and the name of the person. We can also have properties with object value pointing to other resource. For instance, the creator or property on the document is here pointing to the person indicating that the person is the creator of this document. The resource is actually indicated using #fg which is then re-used in the next block. Also note that the use of the resource attribute allows us to very precisely specify the object of this property. Here in this last case, we use property on href that is we reuse the href reference in the HTML directly as the value of the property "mbox". Among the schema that are used in RDFa, one of the most widely used schema is schema.org. It is a vocabulary by major search engines and application. It provides usual terms to describe person, products, reviews, and so on. Here is an example of using the vocabulary from schema.org in RDFa to actually show the reviews on a project and the aggregated reviews. These can be used in the pages of a catalog from an ecommerce website for instance to indicate the reviews on the products, and these data in the pages of the catalog can then be reused by search engine to improve the way it displays the search result. Another application is for instance, the "Like" button. How do a "Like" button knows what is the topic of the like you are doing? It's because inside the webpage, you have a notation RDFa data, metadata that explains what is the page about. For instance, a movie or a recipe. There is also a notion in RDFa of a core initial context that here's a number of predefined prefixes that can be used without defining them in the code. So, to summarize, this data that we have in these pages using RDFa are available in fact to everyone. For instance, if you use data for a "Like" button, these data can also be reused by anyone who applies an RDFa parser to the page. So, if you apply an RDFa parser to these pages, you will obtain these data. In other words, any webpage using that technology can be turned into a data source.
