svgdom is an SVG Full 1.1 processing library, which allows you to parse, manipulate, generate and write an SVG content.
This library was an attempt to create a generic SVG DOM which can be used by various applications. But it the end it turned out that it's easier and faster to use roxmltree svgtypes to extract only the data you need.
There are two main problems with svgdom
:
-
You can't make a nice API with a Vec-based tree and you can't have a safe API with an Rc-tree.
The current implementation uses so-called Rc-tree, which provides a nice API, but all the checks are done in the runtime, so you can get a panic quite easily. It's also hard/verbose to make immutable nodes. You essentially need two types of nodes: one for immutable and one for mutable "references". A Vec-based tree would not have such problems, but you can't implement the simplest operations with it, like copying an attribute from one node to another since you have to have a mutable and an immutable references for this. And Rust forbids this. So you need some sort of generational indexes and so on. This solution is complicated in its own way. Performance is also in question, since inserting/removing an object in the middle of a Vec is expensive.
-
The SVG parsing itself is pretty complex too. There are a lot of ways you can implement it.
svgdom
creates a custom Rc-tree where all the attributes are stored as owned data. This requires a lot of allocations (usually unnecessary). The parsing/preprocessing algorithm itself can be found in docs/preprocessor.md The problem with it is that you can't tweak it. And in many cases, it produces results that you do not need or do not expect.svgdom
was originally used by svgcleaner and resvg and both of these projects are no longer using it.
svgdom is designed to simplify generic SVG processing and manipulations. Unfortunately, an SVG is very complex format (PDF spec is 826 pages long), with lots of features and implementing all of them will lead to an enormous library.
That's why svgdom supports only a static subset of an SVG. No scripts, external resources and complex CSS styling. Parser will convert as much as possible data to a simple doc->elements->attributes structure.
For example, the fill
parameter of an element can be set: as an element's attribute,
as part of a style
attribute, inside a style
element as CSS2, inside an ENTITY
,
using a JS code and probably with lots of other methods.
Not to mention, that the fill
attribute supports 4 different types of data.
With svgdom
you can just use node.has_attribute(AttributeId::Fill)
and don't worry where this
attribute was defined in the original file.
Same goes for transforms, paths and other SVG types.
The main downside of this approach is that you can't save an original formatting and some data.
See the preprocessor doc for details.
- The element link(IRI, FuncIRI) is not just a text, but an actual link to another node.
- At any time you can check which elements linked to the specific element.
See
Node
's doc for details. - Support for many SVG specific data types like paths, transforms, IRI's, styles, etc. Thanks to svgtypes.
- A complete support of text nodes: XML escaping,
xml:space
. - Fine-grained control over the SVG output.
- Only SVG elements and attributes will be parsed.
- Attribute values, CDATA with CSS, DOCTYPE, text data and whitespaces will not be preserved.
- UTF-8 only.
- Only most popular attributes are parsed, other stored as strings.
- No compressed SVG (.svgz). You should decompress it by yourself.
- CSS support is minimal.
- SVG 1.1 Full only (no 2.0 Draft, Basic, Tiny subsets).
- Library follows SVG spec in the data parsing, writing, but not in the tree structure.
- Everything is a
Node
. There are no separatedElementNode
,TextNode
, etc. You still have all the data, but not in the specific struct's. You can check the node type viaNode::node_type()
.
Rust >= 1.32
Licensed under either of
- Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.