/ /cmd /cmd.plated /cmd.plated.build /cmd.plated.watch/html /html.plated /html.plated.chunks /html.plated.files /html.plated.macros /html.plated.operators/html.plated_plugin /html.plated_plugin.blog /html.plated_plugin.copy /html.plated_plugin.docs /html.plated_plugin.import /html.plated_plugin.redirect/js /js.plated /js.plated.build /js.plated.plugin /js.plated.setup /js.plated.watch/js.plated_chunks /js.plated_chunks.deepmerge /js.plated_chunks.delimiter_close_str /js.plated_chunks.delimiter_open_str /js.plated_chunks.delimiter_wrap_str /js.plated_chunks.expand_tag /js.plated_chunks.fill_chunks /js.plated_chunks.format_chunks /js.plated_chunks.lookup /js.plated_chunks.markdown /js.plated_chunks.merge_namespace /js.plated_chunks.pop_namespace /js.plated_chunks.prepare /js.plated_chunks.push_namespace /js.plated_chunks.remove_underscorechunks /js.plated_chunks.replace /js.plated_chunks.replace_once /js.plated_chunks.reset_namespace /js.plated_chunks.set_namespace/js.plated_files /js.plated_files.base_files_to_chunks /js.plated_files.build /js.plated_files.build_file /js.plated_files.empty_cache /js.plated_files.empty_folder /js.plated_files.exists /js.plated_files.file_to_chunks /js.plated_files.filename_fixup /js.plated_files.filename_is_basechunk /js.plated_files.filename_is_plated /js.plated_files.filename_to_dirname /js.plated_files.filename_to_output /js.plated_files.find_dirs /js.plated_files.find_files /js.plated_files.mkdir /js.plated_files.prepare_namespace /js.plated_files.source_to_output /js.plated_files.watch /js.plated_files.write/js.plated_output /js.plated_output.remember /js.plated_output.remember_and_write /js.plated_output.write /js.plated_output.write_all/js.plated_plugin /js.plated_plugin.blog /js.plated_plugin.blog.process_dirs /js.plated_plugin.blog.process_file/js.plated_plugin.copy /js.plated_plugin.copy.process_dirs /js.plated_plugin.copy.process_file /js.plated_plugin.copy.process_output/js.plated_plugin.docs /js.plated_plugin.docs.process_dirs /js.plated_plugin.docs.process_file/js.plated_plugin.import /js.plated_plugin.import.process_file/js.plated_plugin.redirect /js.plated_plugin.redirect.process_dirs /js.plated_plugin.redirect.process_file


Plated operates on a directory structure of files and simply copies the files it does not understand into the output. As such you could have an almost 100% generic static site with only a few special files getting special treatment dependent on their name containing a ^. sequence of characters.


Chunk files are parsed one line at a time to break them into a number of named chunks, lines that begin with the special characters #^ trigger special processing all other lines are simple assigned to the current chunk. eg:

here is some text
that will be assigned to chunkname

A chunk ends when we run out of file or we start another chunk.

As well as the chunk name we can also add some flags to change how the chunk is processed. These flags are added by name=value pairs after the chunkname and separated by white space.

#^chunkname flag=value otherflag=othervalue

These values are intended to change how the chunk is processed here are the ones that we currently support. Multiple flags can of course be applied to the same chunk.

#^chunkname trim=ends

This flag will cause all whitespace to be removed from the beginning and end of the chunk. Useful if you intend to include a chunk in a html attributes as newlines would would mess that up.

#^chunkname form=json

Parse the text as json, this allows for structured data or arrays to be defined in chunks. This data can be referenced inside macro expansion using the . operator to reference members. Most plugins have a json chunk containing their settings.

#^chunkname form=markdown

This text will be parsed as markdown and converted into html. Markdown is an easy way to create simple formated content for blog posts or just pages in general. Note that the default format is just to leave the text of the chunk unprocessed which means normal html tags.

Chunk names that begin with an underscore are considered internal plated chunks ans some will be provided by the plated code for each page rendered. Here is a list of chunks that will be provided by the system.

#^chunkname same=replace

This alters how we replace chunks when they are declared twice, normally the chunk in the lowest level will simply replace a chunk declared elsewhere. This is the default setting.

#^chunkname same=append

If we change to append mode then chunks lower down will be appended to the higher up chunks. This is very useful for css chunks where we can append rules as we cascade down and these rules will take precedence over the earlier rules due to the magic of css.

#^chunkname same=merge

This should only be used with a json chunk, it merges the data with its parent chunk so we can change or add values as we cascade down through the directories similar to append but working as objects rather than just plan text. This is why it should only be used with json chunks.


This will be set to a url that gets to the root of this website and should be used in front of all urls to make sure that we can find the files we want. This can be passed in on the command line and / is a good way to think of this value. The default is actually for plated to provide a relative url such as ../../ that would get you to the root of the site from the current page. You should be using _root as the prefix for all urls when including css or js or images etc in a html file.


Similar to root but intended to explicitly link to the root of the named site. This is intended for use in links that can never be relative to the current url so must be full and explicit urls, eg in RSS feeds.


This will be set to the filename of the input source file that we are currently rendering.


This will be set to url of the file that we are currently rendering.


This will be set to url of the directory containing the file that we are currently rendering.


This is an object containing flag data for the defined chunks, most chunks will not have any flags defined but if they are they can be found in for example _flags.chunkname.flagname which would hold the value of flagname for the given chunkname. This is really just for internal processing and should not need to be referenced by the user.

As well as the provided chunks there are some special names exist to trigger plugin behaviour and need to be defined with the correct configuration data, for instance


Is the name of a chunk containing configuration data in json format that enables the docs plugin to create pages such as the ones that you are currently reading. See html.plated.plugins for documentation on how to use them and what data must be provided. All you need to know is that any chunk name that begins with an underscore belongs to the plated system itself and must only be created according to the documentation.

As well as chunknames there are two other things we can do with special characters at the start of a line.

#^-this is a comment line

The comment line will simply be ignored, in case you wish to include some comments about what a chunk will be used for.


Allows the redefinition of the magic string, for the rest of the file, in this case, it would be ## instead of #^ but any string could be used. This is intended as an escape clause in case the magic string is undesirable in a certain file. It can also be changed on the command line when plated is invoked if you wish to change it globally.


A special string in the filename triggers chunk file processing, by default this is ^. but it can be changed to another character or string when plated is invoked. ^. seems to be a reasonably safe yet generally unused sequence of characters in file names.

Chunk files are text files containing chunks of text assigned to symbolic chunknames, these chunks can then be referenced by tags contained in other chunks and expanded to build final output files.

This is intended to be simple macro expansion only rather than a complex programming language. Any programming would take place in a plugin, for instance we include a blog plugin that helps build blog style websites on top of this base chunk system.

There are two basic types of chunk files, directory scope and file scope. Directory scope chunk files contain chunks that can be referenced by any other chunk file in their directory as well as any sub directories. Directory scope chunks declared inside a sub directory will replace any chunks defined higher up the path, this allows us to adjust and group data using hierarchical directories.


Are all examples of directory scoped chunk files, they do not create any output files themselves but are parsed to provide directory scope chunks which can be used in other chunk files. It does not matter what extension is used, primarily it is intended to be used as a clue to editors to provide appropriate language highlighting. By convention css chunks would be placed inside ^.css and html chunks inside ^.html but this is in no way enforced or necessary for plated to work.


Are all examples of file scoped chunk files. Each one will create a file in the output directory with the same path but the name altered slightly to remove ^. from it. For example index.^.html becomes index.html

The extension used selects the chunk name that will be rendered into the output file. So index.^.html will render the html chunk into the output file and layout.^.css will render the css chunk.

Usually pages in a directory will share the same layout, so a html chunk will be declared at the directory level with the file just adding a content chunk to be rendered inside this page structure. The cascading chunk system means you are free to alter this in sub directories but often pages have a lot in common so keeping a bunch of generic layout chunks in the root with sub directories just picking from them works perfectly. The idea is to provide data organisational tools but not to dictate the use.


Once we have some chunks defined we need to provide a way of refering to them inside other chunks as a macro expansion.

#^chunkname trim=ends
I am expanded
Expand the {chunkname} to its contents.

If the above mainchunk was rendered then {chunkname} would be replace with the contents of the chunk that was defined as #^chunkname this macro expansion is recursive so you can include chunks within chunks within chunks. Combined with the cascading chunks this provides a huge amount of flexibility without any additional programming logic. Similar to {{}moustache} templates but with slightly less logic and a bit more recursion.

If {chunkname} does not exist then the text will be left untouched as {chunkname} also there must be no white space in this macro expansion. so { chunkname } will never expand to anything.

This may sound dangerous but we are able to get away with {} even inside C like languages that contain {} all over the place. If this scares you then you are free to redefine {} to {{}} when invoking plated but I assure you it is not necessary.


When using a json chunk a . can be used to pull out a value from the object


If it is an array then the first item could be picked out with a number and then its member.


Negative numbers are allowed in which case it counts backwards from the end, in this case the last name of the last object in the array would be used.

Finally a json chunk may have another chunk applied as a layout.


In this case plate is a chunk name that renders with {_it} being synonymous with {jsonchunk} This is similar to calling a macro with a number of values.


If a template is applied to an array then it is applied repeatedly to each item in that array. This allows for simple formatting of json data held within an array. The loop happens auto-magically with {_it} expanding to a different value each time.

All of these templating expansions are intended for use by plugins which provide arrays or objects of data for you to display.

If a plate is applied to empty data then the empty string is returned. Eg no expansion happens, this can help with layout logic removing some chunks and showing others depending on their existence.

Finally because we also need to be able to talk about these macros here without them accidently expanding then we have a simple way to escape them.


No matter how valid the chunkname is it will not expand because it is contained within the comment tags {[}{]} these tags will be removed after the text is rendered.


If multiple [ are used instead of one then it allows one level macro replacement per extra [ so we can control expansion this way. In the above example {chunkname} will expand once but any macros within that chunk will stay untouched. A tad complex but escape syntax is always a pain.

Due to the way this expands you must be careful to balance any use of {[}{]} within this chunk so as not to accidentally close the tag prematurely. This documentation for instance is designed to be used inside such a chunk so all square brackets inside {} have been carefully balanced to stop anything from going wrong.


As well as chunknames we can also combine some logic operators to control how macros expand. This is mostly of use with json chunks where you wish to make a choice between a number of possibilities. For instance


The above is an example of a value?one:two style logic test. First obj.count is compared to the value 1 this will work if it is a string or a number due to the loose typing used. If true then obj will be rendered with the template chunk showit if false then instead obj will be rendered with the template chunk hideit. You can guess what these two templates are intended to do.

The operators are evaluated left to right with no precedence and are C like, hence == rather than just a single = sign. Here are all the possible operators that can be used.

    returns true if a is less than b
    returns true if a is less than or equal to b
    returns true if a is more than b
    returns true if a is more than or equal to b
    returns true if a is equal to b
    a and b returns b if a is true else returns a
    a or b returns b if a is false else returns a
    returns a if it exists else return an empty string

The last one is useful for making macros invisible if they refer to empty data. Normally macros remain in the output if they are invalid. So {a} on its own would either expand to something else or remain as {a} in the output. {a||} is just using the || operator to make it go away if empty.


        title="Feed Title",

A chunk of this name must be created in a directory scope file for this plugin to parse it. posts_per_page is the number of posts per page, we will create as many pages as we need.

posts_per_feed specifys the number of posts to publish in the feed.json file. Which will be published using any data suplied in feed which can contain any valid jsonfeed values. The base url must also be suplied since feeds are expected to be copied to other domains. This url is intentionally seperate from _root as it must be explictly set and we can not get away with relative paths here.

Every directory within this blog directory will now be treated as a blogpost.

See source/blog for an example of this all fits together. Inside each of these directories we look for.

    "title":"my title",
    "author":"my name",

Which contains metadata about the blog post, the feed object can contain any valid jsonfeed settings, by we try and set useful defaults from the rest of the metadata. All of these values can also be used in your templates to render the blog posts.

#^_blog_post_body form=markdown
This is my blog post body.

Our blog body is to be found in this chunk name, probably best to use markdown as it makes writing blog posts easier.

When it comes to generating the pages then the following chunks should be setup in base directory.

This is a page of blog posts, eg the front page.

Within this chunk we provide _blog_page_older as the url of an older page and _blog_page_newer as the url of a newer page. If there is no newer or older page then this will be empty. _list will contain an array of blog posts that we intend to display in this page. It will be at least one post and no more than the posts_per_page setting.

This is a single blog post, when viewed on its own page.

Within this chunk we provide _blog_post_older as the url of an older page and _blog_post_newer as the url of a newer page. If there is no newer or older page then this will be empty. _blog_post_body will contain the _blog_post_body as defined in the blog post directory.



A chunk of this name must be created in a directory scope file for this plugin to parse it. include is a list of prefixes to include and exclude is a list of prefixes within the include to exclude.

The above configuration is assumed to be within a file fra/^.index and spa/^.index so it would include all chunkfiles from root but exclude files in fra or spa, ie itself. This way we can have a default english site and a spanish translation under spa/ or french under fra/

The files are copied into the current directory without the prefix used in include.

An example can be found in test-source/006-copy



A chunk of this name must be created in a directory scope file for this plugin to parse it. We should ignore any paths containing the key strings in the ignore object and will include (recursively) the keys in the dirs object but only if the filename ends in the given string.

So in the above case we will scan ../js for all .js files but ignore everything in node_modules. One should always ignore everything in node_modules.

These files are searched for special auto doc documentation syntax where any line that begins with --[#name.of.the.chunk will begin a special documentation chunk and ]] will end it. In both cases the string must be at the start of a line.

Each of these chunks will then be rendered into its own page as well as its parent pages, we use dot notation to describe this relationship. In the case of name.of.the.chunk it will exist in name.of.the name.of name and the always present /

Take a look at the source code that generates this site documentation in source/docs for an example of how this can be themed and presented.


#^importedchunk import=dir/dir

The content if this chunk is unimportant as it will be replaced by the chunk referenced from another file via the import=dir flag.



A chunk of this name must be created in a directory scope file for this plugin to parse it. files is a map of redirections we require.

Perform a redirect of files, using simple html javascript redirection, to another url. We mostly make use of the automatic use of an index.html inside a given directory so in the case above ./from will get redirected to ./to note that these can contain one level of macro expansion so {_root}dir is a reasonable redirection.

Multiple redirections can be performed in a single json configuration but be aware that we end up actually creating files to perform these redirections so be careful not to clash with other files.

An example can be found in test-source/005-redirect