Flowchart.js – Simple SVG flow chart diagrams from textual representation
adrai.github.io
adrai.github.io
Both could use better node layout engines and support for DOT syntax or might be easier to adapt graphviz with Emscripten.
And apologies for another "is there something like this for X" comment!
<grumpyOldGuy>But I guess everything have to be rewritten and reinvented in js nowadays... I kind of getting why all the COBOL geezers where tired of all the shit going around again, just solving the same old shit in another way instead of truly getting somewhere, or why Douglas Engelbart got depressed from spending his last 50 years looking at people reinventing his stuff...</grumpyOldGuy>
Cool though.
They have an interactive shell for it - experiment with http://interactive.blockdiag.com/nwdiag/.
var svg = document.getElementsByTagName("svg");
var code = svg[0].outerHTML;
Of course, that assumes you're just getting the first SVG, etc and so on. Anyhow, this does NOT work on IE because IE hates innerHTML / outerHTML for SVG elements.One tiny thing that stands out is that RaphaelJS is built to support even IE6, and yet this flowchart program seems to bork under IE.
Edit #2: forgot to mention, Raphael was specifically designed to fallback to VML if SVG is unavailable, so the export functionality would need to support both or pick and choose, or just reconstruct programmatically. A dilemma indeed.
https://github.com/gabelerner/canvg
Still needs a very newish browser however. No IE 6-8 (& 9?) support.
Then it gets weirder because RaphaelJS isn't exactly responsive-ready. Takes some work to make graphics just resizable (you can scale with the viewbox, but that's about it.) Recently, I switched to pure SVG (even when programmatically generated) because nested SVG elements provide basically a mechanism for relative/responsive scaling, positioning and such.
The guy who came up with RaphaelJS was ahead of his time, and I believe he got hired by Adobe in part thanks to the project's exposure. The development stopped awhile back, and it shows, which is really sad because I liked how the programming style for Raphael is done.
One takes the SVG from the browser, encodes it and sends the <svg> as the body of an email (inlined, not as an attachment). Here's the relevant part:
function render() {
var link = document.createElement('a'),
time = Date.now();
link.href = 'mailto:? subject=exported%20sketch&body=%3C%3Fxml%20version%3D%221.0%22%20standalone%3D%22no%22%3F%3E%3C!DOCTYPE%20svg%20PUBLIC%20%22-%2F%2FW3C%2F%2FDTD%20SVG%201.1%2F%2FEN%22%20%22http%3A%2F%2Fwww.w3.org%2FGraphics%2FSVG%2F1.1%2FDTD%2Fsvg11.dtd%22%3E'+encodeURIComponent(new XMLSerializer().serializeToString(board));
link.id = time;
document.body.appendChild(link);
document.getElementById(time).dispatchEvent(new MouseEvent('click'));
document.getElementById(time).parentNode.removeChild(document.getElementById(time));
}
The other prepares it as a file and creates a download link (not seen on screen) and programmatically clicks that link to download it: function download() {
var link = document.createElement('a'),
time = Date.now(),
svg = '<?xml version="1.0" standalone="no"?><!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">'+new XMLSerializer().serializeToString(board);
link.id = time;
link.setAttribute('download','sketch.svg');
link.href = 'data:text/html;charset=utf-8,' + encodeURIComponent(svg);
document.body.appendChild(link);
document.getElementById(time).dispatchEvent(new MouseEvent('click'));
document.getElementById(time).parentNode.removeChild(document.getElementById(time));
}
But be forewarned, since SVG uses XML namespaces instead of HTML, using things like '.innerHTML' won't work in certain browsers (iOS Mobile Safari for example) while innerHTML does tend to work as you would expect on most desktop browsers!Canvas would work for exporting a bitmapped snapshot of the vector, but lacks the following:
- filesize: SVG is vector, but Canvas renders a bitmap
- non-destructive: SVG is human-editable after the fact, bitmaps only 'remember' the value of the top-most pixel and 'forgets' everything that has overlapped
- human-editable: Though not pretty, SVG's can be read and edited by hand, bitmaps require a graphics tool to edit easily
Nomnoml, a tool for drawing UML diagrams based on a simple syntax: http://www.nomnoml.com/
I recently worked on a project where we started off looking to use something similar to what op posted. Despite being a very old library, nothing really provided the level of stability and options that Graphviz had. If you can get past the somewhat odd syntax, I would highly recommend it for complicated directed graph output.
Lots of directed graph options. I'd argue the model data is even simpler to create and comprehend, but there's a LOT more to GoJS than there is to this project. Also it isn't free.
(Disclaimer: I'm a GoJS developer)
What I've been doing is I just have a dir of .dot files and then I have a gen.sh script which writes the SVG/PNG files.
Dot is a bit more low level but more powerful. Great for documenting protocols or state machines.