Did you know ... | Search Documentation: |
Packs (add-ons) for SWI-Prolog |
Title: | Logic-free text (HTML) templates |
---|---|
Rating: | Not rated. Create the first rating! |
Latest version: | 1.3.1 |
SHA1 sum: | 847e491ad7a26704944c304ac6b159ca45af9c4e |
Author: | James Lawrence Turner http://jlturner.com/ |
Matthew Carter http://ahungry.com/ | |
Raivo Laanemets http://rlaanemets.com/ | |
Sando George http://sandogeorge.github.io/ | |
Home page: | https://github.com/rla/simple-template |
No reviews. Create the first review!.
Text (HTML) template processor for SWI-Prolog. Works best for cases when you have mainly static HTML.
![Build Status](https://travis-ci.org/rla/simple-template)
Visual Code syntax support: https://marketplace.visualstudio.com/items?itemName=rlaanemets.simple-template-syntax
Input markup (test.html
file):
<h1>{{= title }}</h1> {{ each items, item }} <h2>{{= item.title }}</h2> <div class="content">{{- item.content }}</div> {{ end }}
rendering with data:
use_module(library(st/st_render)). current_output(Out), st_render_file(test, _{ title: 'Hello', items: [ _{ title: 'Item 1', content: 'Abc 1' }, _{ title: 'Item 1', content: 'Abc 2' } ] }, Out, []).
output:
<h1>Hello</h1> <h2>Item 1</h2> <div class="content">Abc 1</div> <h2>Item 1</h2> <div class="content">Abc 2</div>
If you find your tools work better with a template tag syntax that more closely resembles other web templating engines, such as Django (Python) / Djula (Common Lisp) / Twig (PHP), you can also use an alternate frontend syntax called semblance:
Input markup (test.html
file):
<h1>{{ title }}</h1> {% each items, item %} <h2>{{ item.title }}</h2> <div class="content">{% unescape item.content %}</div> {% end %}
rendering with data:
use_module(library(st/st_render)). current_output(Out), st_render_file(test, _{ title: 'Hello', items: [ _{ title: 'Item 1', content: 'Abc 1' }, _{ title: 'Item 1', content: 'Abc 2' } ] }, Out, _{ frontend: semblance }).
output:
<h1>Hello</h1> <h2>Item 1</h2> <div class="content">Abc 1</div> <h2>Item 1</h2> <div class="content">Abc 2</div>
It is not possible to mix two syntaxes through a single render call.
Alternatively, you can specify your own templating delimiter syntax. This is useful to prevent template collision, in similar cases to semblance, but is more flexible.
Input markup (test.html
file):
<h1><%= title %></h1> <% each items, item %> <h2><%= item.title %></h2> <div class="content"><%~ item.content %></div> <% end %>
rendering with data:
use_module(library(st/st_render)). current_output(Out), st_render_file(test, _{ title: 'Hello', items: [ _{ title: 'Item 1', content: 'Abc 1' }, _{ title: 'Item 1', content: 'Abc 2' } ] }, Out, _{ frontend: syntax_tokens( comment("<%#", "%>"), out("<%=", "%>"), out_unescaped("<%~", "%>"), statement("<%", "%>") ) }).
output:
<h1>Hello</h1> <h2>Item 1</h2> <div class="content">Abc 1</div> <h2>Item 1</h2> <div class="content">Abc 2</div>
If you'd like to use an unescape keyword, you can use the helper
keyword_unescape_start(Token)
in place of a string, where Token
is the starting Token to unify:
use_module(library(st/st_render)). current_output(Out), st_render_file(test, _{ title: 'Hello', items: [ _{ title: 'Item 1', content: 'Abc 1' }, _{ title: 'Item 1', content: 'Abc 2' } ] }, Out, _{ frontend: syntax_tokens( comment("<%#", "%>"), out("<%=", "%>"), out_unescaped(keyword_unescape_start("<%="), "%>"), statement("<%", "%>") ) }).
With input markup (test.html
file):
<h1><%= title %></h1> <% each items, item %> <h2><%= item.title %></h2> <div class="content"><%= unescape item.content %></div> <% end %>
output:
<h1>Hello</h1> <h2>Item 1</h2> <div class="content">Abc 1</div> <h2>Item 1</h2> <div class="content">Abc 2</div>
Render template from string:
st_render_string(String, Data, Stream, File, Options).
Render template from file:
st_render_file(File, Data, Stream, Options).
Render template from codes:
st_render_codes(Codes, Data, Stream, File, Options).
Options accept the following:
utf8
.html
.false
.false
.simple
(default) or semblance
.error
(default) or false
(evaluates to false).Processing instructions start with double opening curly braces (`{{`) and end with double closing curly braces (`}}`). Semblance syntax variations are described below.
There are 4 types of processing instructions.
There are two output instructions. `{{= expression }}` outputs the expression
value and escapes the special HTML characters. {{- expression }}
outputs the
expression value but does not escape the output. There must be no space
between `{{ and
= or
-`.
Includes use the syntax `{{ include path/to/file }}`. The path should be relative. The relative path is resolved against the current file location and the currently set file extension is added. All values from the current scope are available to the included file. There is also `{{ include path/to/file, expression }}` that sets the scope of the included file to the value of the expression (must be a dict).
Dynamic includes use the syntax `{{ dynamic_include expression }}`. The value of the expression must be a file path specifier. There is also a variant with the scope expression.
Includes are currently processed at runtime. You should enable caching if you include inside a loop otherwise the included file is parsed over and over again.
Conditional instructions have the following syntax:
{{ if cond_expression1 }} ... {{ else if cond_expression2 }} ... {{ else }} ... {{ end }}
`else if and
else` parts are optional.
The each loop allows to process lists. The syntax for each loop:
{{ each expression, item_var, index_var, len_var }} ... {{ end }}
The value of expression must be a list. item_var
, index_var
and len_var
refer to current item, current item index and the length of the list.
index_var
and len_var
are optional.
Blocks allow to compose templates while passing around the contained template as a value. This allows to wrap templated content with another template.
General syntax (block usage):
{{ block path/to/file }} Block content. {{ end }}
And to render the distributed block content:
{{ slot }}
Example 1: wrapping content into a panel element.
Input file:
<div class="other_content1"></div> {{ block panel }} <span>This will be wrapped in panel.</span> {{ end }} <div class="other_content2"></div>
Panel file:
<div class="panel">{{ slot }}</div>
Rendering result:
<div class="other_content1"></div> <div class="panel"><span>This will be wrapped in panel.</span></div> <div class="other_content2"></div>
Variable scoping inside block content is lexical. Variable scope inside block is either the current scope or selectable with an expression similar to the `{{ include file, expr }}` instruction.
Example 2: layouts.
Layout file (layout.html
):
<!DOCTYPE html> <html lang="en"> <head> <meta charset="UTF-8"> <title>{{= title }}</title> </head> <body> <h1>{{= title }}</h1> {{ slot }} </body> </html>
Concrete page file:
{{ block layout }} <span>This is the page file.</span> {{ end }}
Data:
_{ title: "A page title" }
Rendering result:
<!DOCTYPE html> <html lang="en"> <head> <meta charset="UTF-8"> <title>A page title</title> </head> <body> <h1>A page title</h1> <span>This is the page file.</span> </body> </html>
The semblance syntax has the following differences.
Output instruction: {{ variable }}
.
Unescaped output: `{% unescape rawvar %}`.
Conditional instructions:
{% if cond_expression1 %} ... {% else if cond_expression2 %} ... {% else %} ... {% end %}
Each loop:
{% each expression, item_var, index_var, len_var %} ... {% end %}
Comments: `{# comment #}`.
Expressions are ground Prolog terms with the interpretation described below.
In most contexts, atoms are interpreted as scope entries ("variables"). An error is thrown when there is no suitable entry.
The following arithmetic operators and functions are supported:
-, +, *, /, mod, rem, //, div, abs, sign, max, min, random, round, truncate, floor, ceiling, **, ^
.
Boolean expressions include modified unification using the =
operator. The
operator provides coercion from atom to string when either of operands is an atom
and the other is a string. For other types, no coercion is applied. Similarily works
the \=
operator. Supported boolean operators are: `,` (logical and), ;
(logical or)
and \+
(logical negation). Supported comparison operator are: `>, <, >=, =<,`.
Conditional expression if(expression, true_expr, false_expr)
value is the value of
true_expr
when the expression
evaluates to anything but 0 or false
. Otherwise the
value is the value of false_expr
.
The +
operator is overloaded to provide string or atom concatenation when
the left side of the operator is an atom or string. The result is string.
The `.` operator is used for accessing values from dicts. Internally the ./3 predicate is used which means that user-defined functions on dicts will work as intended.
Compound terms that match none of the operators and constructs described above are treated as user-defined (see below) function calls. An error is thrown when there is no such function defined.
A special expression is atom(something)
. Its value is atom something
. This
is added to differentiate them from other atoms which are otherwise interpreted
as scope entries.
Literal lists work as expected with the elements having the interpretation as the expressions above.
The default behaviour when a variable that is not in the scope (template data dict) is encountered is to throw a no_entry
error. This behaviour can be modified to set the variable value to false
(default is error
) by setting the undefined
option to false
when rendering templates. For example:
st_render_file(test, _{ title: 'Hello', items: [ _{ title: 'Item 1', content: 'Abc 1' }, _{ title: 'Item 1', content: 'Abc 2' } ] }, Out, _{ undefined: false }).
This is useful in situations where a developer may want to include a block in the template based on the existence of a variable. For example:
{% if page_header %} <h2>{{page_header}}</h2> {% end %}
There are three single-argument functions to encode URI parts: encode_path
,
encode_query_value
and encode_fragment
. These call
uri_encoded/3 with
the respective first argument.
Global constants can be set with the st_set_global/2 predicate by importing
the st_expr
module. When a scope contains an
entry with the same name as a global then the local value is preferred.
User-defined functions can be set with the st_set_function(Name, Arity, Goal)
predicate in the st_expr
module. Arity refers to the function arity. Goal must
have arity Arity + 1.
Comment blocks start with `{{%` and end with `}}`. Comments do not appear in the output.
Template caching is enabled by setting option cache: true
during rendering. This makes
the system cache parsed templates. This is particulary useful when using includes in loops.
To purge the current cache contents, use the st_cache_invalidate
predicate.
Some (but not all) whitespace is removed by setting option
strip: true
during rendering. Line indents and some duplicate line ends
are removed from the output. Whitespace removal is parse-time and does not
incur any runtime penalty.
Encoding for template files can be specified by setting option encoding
during rendering. Accepted
values are all that are accepted by the encoding
option of the read_file_to_codes/3 predicate.
This package requires SWI-Prolog 7.x.
pack_install(simple_template).
See http://packs.rlaanemets.com/simple-template/doc/.
In the package root, insert into swipl:
[tests/tests]. run_tests.
Or if you cloned the repo:
make test
Please send bug reports/feature request through the GitHub project page.
The MIT License. See LICENSE file.
Pack contains 25 files holding a total of 68.4K bytes.