Migration to Edge Delivery Service using the Importer

Transcript

The general idea of the importer is pretty straightforward. It takes a page DOM and transforms it into a Markdown file which is then converted to a docx file. For now, let’s consider that the Markdown file is a one-to-one equivalent to the docx file thus the next references to Markdown or docx are equivalent to the output of the transformation process. As Markdown is a pretty simple format, the DOM transformation is really basic, ih1 becomes a heading one, a paragraph or text and a span or a div becomes a paragraph, an anchor tag stays a link, an img tag an image. All styling, layout or div nesting disappears in the Markdown output. The only special case is a table tag which becomes a grid table element in the Markdown output and becomes a table in Word, which is the foundation for blocks. The point is really to only extract the content from the original page. And the importer primary objective is to help in digesting a large amount of pages from an existing website. If you have only a few pages on the website, it is easier and faster to manually copy slash paste the content into Word documents. But in the case of a large website with pages that are structurally similar, for example a blog site with thousands of blog articles, it would be fastidious to manually copy slash paste all pages. To summarize, if a large set of pages look the same, this is when you want to use the importer and write a specific import .js transformation file. The transformation file will contain a set of basic transformation rules. Writing those import transformation rules is an iterative process, you will first digest one page with a minimum of transformations. Then you should try to style this page, you will see that you might need some blocks for certain pieces. That’s when you come back to your transformation rules and add more to automatically convert some areas of the DOM into blocks. Once the page is exactly what you expect in terms of content structure, try another one. You might find other blocks that could benefit from automatic transformations. Try with two, three more. Then you can run the import on your whole set of pages. It is worth mentioning that this is a developer tool and requires some understanding of JavaScript and the DOM. But nothing else.