lua-htmlparser/README.md

# LuaRock "htmlparser"

Parse HTML text into a tree of elements with selectors

[1]: https://api.jquery.com/category/selectors/

## Install
Htmlparser is a listed [LuaRock](http://luarocks.org/repositories/rocks/). Install using [LuaRocks](http://www.luarocks.org/): `luarocks install htmlparser`

### Dependencies
Htmlparser depends on [Lua 5.1-5.3](https://www.lua.org/download.html) or [LuaJIT](https://luajit.org/download.html), which provides 5.1-compatible ABI.
To be able to run the tests, [lunitx](https://github.com/dcurrie/lunit) also comes along as a LuaRock

## Usage
Start off with
```lua
local htmlparser = require("htmlparser")
```
Then, parse some html:
```lua
local root = htmlparser.parse(htmlstring)
```
Optionally, you can pass loop-limit value (integer). This value means the deepness of the tree, after which parser will give up. Default value is 1000.
Also, global variable `htmlparser_looplimit` is supported (while this optional argument takes priority over global value)

The input to parse may be the contents of a complete html document, or any valid html snippet, as long as all tags are correctly opened and closed.
Now, find specific contained elements by selecting:
```lua
local elements = root:select(selectorstring)
```
Or in shorthand:
```lua
local elements = root(selectorstring)
```
This wil return a list of elements, all of which are of the same type as the root element, and thus support selecting as well, if ever needed:
```lua
for _,e in ipairs(elements) do
	print(e.name)
	local subs = e(subselectorstring)
	for _,sub in ipairs(subs) do
		print("", sub.name)
	end
end
```
The root element is a container for the top level elements in the parsed text, i.e. the `<html>` element in a parsed html document would be a child of the returned root element.

## Selectors
Supported selectors are a subset of [jQuery's selectors][1]:

- `"*"` all contained elements
- `"element"` elements with the given tagname
- `"#id"` elements with the given id attribute value
- `".class"` elements with the given classname in the class attribute
- `"[attribute]"` elements with an attribute of the given name
- `"[attribute='value']"` equals: elements with the given value for the given attribute
- `"[attribute!='value']"` not equals: elements without the given attribute, or having the attribute, but with a different value
- `"[attribute|='value']"` prefix: attribute's value is given value, or starts with given value, followed by a hyphen (`-`)
- `"[attribute*='value']"` contains: attribute's value contains given value
- `"[attribute~='value']"` word: attribute's value is a space-separated token, where one of the tokens is the given value
- `"[attribute^='value']"` starts with: attribute's value starts with given value
- `"[attribute$='value']"` ends with: attribute's value ends with given value
- `":not(selectorstring)"` elements not selected by given selector string
- `"ancestor descendant"` elements selected by the `descendant` selector string, that are a descendant of any element selected by the `ancestor` selector string
- `"parent > child"` elements selected by the `child` selector string, that are a child element of any element selected by the `parent` selector string

Selectors can be combined; e.g. `".class:not([attribute]) element.class"`

## Element type
All tree elements provide, apart from `:select` and `()`, the following accessors:

### Basic
- `.name` the element's tagname
- `.attributes` a table with keys and values for the element's attributes; `{}` if none
- `.id` the value of the element's id attribute; `nil` if not present
- `.classes` an array with the classes listed in element's class attribute; `{}` if none
- `:getcontent()` the raw text between the opening and closing tags of the element; `""` if none
- `.nodes` an array with the element's child elements, `{}` if none
- `.parent` the element that contains this element; `root.parent` is `nil`

### Other
- `.index` sequence number of elements in order of appearance; root index is `0`
- `:gettext()` the complete element text, starting with `"<tagname"` and ending with `"/>"` or `"</tagname>"`
- `.level` how deep the element is in the tree; root level is `0`
- `.root` the root element of the tree; `root.root` is `root`
- `.deepernodes` a [Set][1] containing all elements in the tree beneath this element, including this element's `.nodes`; `{}` if none
- `.deeperelements` a table with a key for each distinct tagname in `.deepernodes`, containing a [Set][1] of all deeper element nodes with that name; `{}` if none
- `.deeperattributes` as `.deeperelements`, but keyed on attribute name
- `.deeperids` as `.deeperelements`, but keyed on id value
- `.deeperclasses` as `.deeperelements`, but keyed on class name

## Limitations
- Attribute values in selector strings cannot contain any spaces
- The spaces before and after the `>` in a `parent > child` relation are mandatory 
- `<!` elements (including doctype, comments, and CDATA) are not parsed; markup within CDATA is *not* escaped
- Textnodes are no separate tree elements; in `local root = htmlparser.parse("<p>line1<br />line2</p>")`, `root.nodes[1]:getcontent()` is `"line1<br />line2"`, while `root.nodes[1].nodes[1].name` is `"br"`
- No start or end tags are implied when [omitted](http://www.w3.org/TR/html5/syntax.html#optional-tags). Only the [void elements](http://www.w3.org/TR/html5/syntax.html#void-elements) should not have an end tag
- No validation is done for tag or attribute names or nesting of element types. The list of void elements is in fact the only part specific to HTML

## Examples
See `./doc/sample.lua`

## Tests
See `./tst/init.lua`

## License
LGPL+; see `./doc/LICENSE`
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`# LuaRock "htmlparser"`
Initial commit 2013-03-19 10:24:17 +00:00
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			`Parse HTML text into a tree of elements with selectors`

Added loop limit, added kludge for template engines, documented loop limit, added error messages, fixes #42 2017-04-08 18:55:25 +00:00			`[1]: https://api.jquery.com/category/selectors/`
added some links and descriptions 2013-03-28 12:22:57 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Install`
Rock published 2013-03-28 22:36:43 +00:00			Htmlparser is a listed [LuaRock](http://luarocks.org/repositories/rocks/). Install using [LuaRocks](http://www.luarocks.org/): `luarocks install htmlparser`

fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`### Dependencies`
Added loop limit, added kludge for template engines, documented loop limit, added error messages, fixes #42 2017-04-08 18:55:25 +00:00			`Htmlparser depends on [Lua 5.1-5.3](https://www.lua.org/download.html) or [LuaJIT](https://luajit.org/download.html), which provides 5.1-compatible ABI.`
			`To be able to run the tests, [lunitx](https://github.com/dcurrie/lunit) also comes along as a LuaRock`
Rock published 2013-03-28 22:36:43 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Usage`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			`Start off with`
			```lua
			`local htmlparser = require("htmlparser")`
			```
			`Then, parse some html:`
			```lua
			`local root = htmlparser.parse(htmlstring)`
			```
Added loop limit, added kludge for template engines, documented loop limit, added error messages, fixes #42 2017-04-08 18:55:25 +00:00			`Optionally, you can pass loop-limit value (integer). This value means the deepness of the tree, after which parser will give up. Default value is 1000.`
			Also, global variable `htmlparser_looplimit` is supported (while this optional argument takes priority over global value)

First draft for a Rock setup 2013-03-28 11:24:10 +00:00			`The input to parse may be the contents of a complete html document, or any valid html snippet, as long as all tags are correctly opened and closed.`
fix typos in README.md 2013-12-01 23:40:13 +00:00			`Now, find specific contained elements by selecting:`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			```lua
			`local elements = root:select(selectorstring)`
			```
			`Or in shorthand:`
			```lua
			`local elements = root(selectorstring)`
			```
conforming to changes following #25 Selector result is now a list, instead of a Set 2013-12-10 12:31:17 +00:00			`This wil return a list of elements, all of which are of the same type as the root element, and thus support selecting as well, if ever needed:`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			```lua
conforming to changes following #25 Selector result is now a list, instead of a Set 2013-12-10 12:31:17 +00:00			`for _,e in ipairs(elements) do`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			`print(e.name)`
			`local subs = e(subselectorstring)`
conforming to changes following #25 Selector result is now a list, instead of a Set 2013-12-10 12:31:17 +00:00			`for _,sub in ipairs(subs) do`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00			`print("", sub.name)`
			`end`
			`end`
			```
added some links and descriptions 2013-03-28 12:22:57 +00:00			The root element is a container for the top level elements in the parsed text, i.e. the `<html>` element in a parsed html document would be a child of the returned root element.
First draft for a Rock setup 2013-03-28 11:24:10 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Selectors`
Added loop limit, added kludge for template engines, documented loop limit, added error messages, fixes #42 2017-04-08 18:55:25 +00:00			`Supported selectors are a subset of [jQuery's selectors][1]:`
added some links and descriptions 2013-03-28 12:22:57 +00:00
			- `"*"` all contained elements
			- `"element"` elements with the given tagname
			- `"#id"` elements with the given id attribute value
			- `".class"` elements with the given classname in the class attribute
			- `"[attribute]"` elements with an attribute of the given name
shortened descriptions 2013-04-08 12:54:00 +00:00			- `"[attribute='value']"` equals: elements with the given value for the given attribute
			- `"[attribute!='value']"` not equals: elements without the given attribute, or having the attribute, but with a different value
added some links and descriptions 2013-03-28 12:22:57 +00:00			- `"[attribute\|='value']"` prefix: attribute's value is given value, or starts with given value, followed by a hyphen (`-`)
			- `"[attribute*='value']"` contains: attribute's value contains given value
			- `"[attribute~='value']"` word: attribute's value is a space-separated token, where one of the tokens is the given value
			- `"[attribute^='value']"` starts with: attribute's value starts with given value
			- `"[attribute$='value']"` ends with: attribute's value ends with given value
last few selectors lacked a description 2013-03-28 12:29:53 +00:00			- `":not(selectorstring)"` elements not selected by given selector string
grammatical 2013-03-28 12:31:08 +00:00			- `"ancestor descendant"` elements selected by the `descendant` selector string, that are a descendant of any element selected by the `ancestor` selector string
			- `"parent > child"` elements selected by the `child` selector string, that are a child element of any element selected by the `parent` selector string
After proofreading on the site 2013-03-28 11:38:13 +00:00
Just a few more `'s 2013-03-28 11:40:47 +00:00			Selectors can be combined; e.g. `".class:not([attribute]) element.class"`
First draft for a Rock setup 2013-03-28 11:24:10 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Element type`
After proofreading on the site 2013-03-28 11:38:13 +00:00			All tree elements provide, apart from `:select` and `()`, the following accessors:
Samples 2013-03-28 11:47:22 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`### Basic`
added some links and descriptions 2013-03-28 12:22:57 +00:00			- `.name` the element's tagname
			- `.attributes` a table with keys and values for the element's attributes; `{}` if none
			- `.id` the value of the element's id attribute; `nil` if not present
			- `.classes` an array with the classes listed in element's class attribute; `{}` if none
			- `:getcontent()` the raw text between the opening and closing tags of the element; `""` if none
			- `.nodes` an array with the element's child elements, `{}` if none
typo 2013-12-11 07:48:10 +00:00			- `.parent` the element that contains this element; `root.parent` is `nil`
Samples 2013-03-28 11:47:22 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`### Other`
conforming to changes following #25 Selector result is now a list, instead of a Set 2013-12-10 12:31:17 +00:00			- `.index` sequence number of elements in order of appearance; root index is `0`
shortened descriptions 2013-04-08 12:54:00 +00:00			- `:gettext()` the complete element text, starting with `"<tagname"` and ending with `"/>"` or `"</tagname>"`
added some links and descriptions 2013-03-28 12:22:57 +00:00			- `.level` how deep the element is in the tree; root level is `0`
After proofreading on the site 2013-03-28 11:38:13 +00:00			- `.root` the root element of the tree; `root.root` is `root`
added some links and descriptions 2013-03-28 12:22:57 +00:00			- `.deepernodes` a [Set][1] containing all elements in the tree beneath this element, including this element's `.nodes`; `{}` if none
typo 2013-12-11 07:50:27 +00:00			- `.deeperelements` a table with a key for each distinct tagname in `.deepernodes`, containing a [Set][1] of all deeper element nodes with that name; `{}` if none
added some links and descriptions 2013-03-28 12:22:57 +00:00			- `.deeperattributes` as `.deeperelements`, but keyed on attribute name
			- `.deeperids` as `.deeperelements`, but keyed on id value
			- `.deeperclasses` as `.deeperelements`, but keyed on class name
reordered paragraphs 2013-04-08 12:17:14 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Limitations`
Removed the dropped limitations 2014-01-10 19:55:10 +00:00			`- Attribute values in selector strings cannot contain any spaces`
shortened descriptions 2013-04-08 12:54:00 +00:00			- The spaces before and after the `>` in a `parent > child` relation are mandatory
			- `<!` elements (including doctype, comments, and CDATA) are not parsed; markup within CDATA is not escaped
fix typos in README.md 2013-12-01 23:40:13 +00:00			- Textnodes are no separate tree elements; in `local root = htmlparser.parse("<p>line1<br />line2</p>")`, `root.nodes[1]:getcontent()` is `"line1<br />line2"`, while `root.nodes[1].nodes[1].name` is `"br"`
shortened descriptions 2013-04-08 12:54:00 +00:00			`- No start or end tags are implied when [omitted](http://www.w3.org/TR/html5/syntax.html#optional-tags). Only the [void elements](http://www.w3.org/TR/html5/syntax.html#void-elements) should not have an end tag`
			`- No validation is done for tag or attribute names or nesting of element types. The list of void elements is in fact the only part specific to HTML`
reordered paragraphs 2013-04-08 12:17:14 +00:00
fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Examples`
reordered paragraphs 2013-04-08 12:17:14 +00:00			See `./doc/sample.lua`

fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## Tests`
reordered paragraphs 2013-04-08 12:17:14 +00:00			See `./tst/init.lua`

fix format README.md (#46) 2017-04-08 03:46:27 +00:00			`## License`
LGPL 2013-12-06 13:31:49 +00:00			LGPL+; see `./doc/LICENSE`