NAME

Web::DataService::Configuration::Node - how to configure data service nodes

SYNOPSIS

This page describes the role that data service nodes play in the Web::DataService framework, and how to configure them. It includes a list of the attributes that you can use to define them.

NODE DEFINITIONS

Data service nodes are the fundamental organizing elements of a data service definition under this framework. They correspond to the various resources provided by the data service. You define them by calling the method define_node on a data service instance and specifying one or more sets of attributes. Any node for which the attributes role and method are both defined will correspond to a data service operation (we call these "operation nodes") while other nodes may correspond to documentation pages, images, stylesheets, or other files.

The only required attribute for a node is path, which provides a unique key. Most attributes of data service nodes are inherited path-wise. That is, a node with path "foo/bar" will inherit from the node with path "foo" any attributes whose values are not explicitly specified in its definition. The node with path "/" functions as the root, and its attribute values provide defaults for all of the other nodes. Any node attribute can be explicitly disabled for a particular node by specifying its value as the empty string. Those attributes that are not inherited will be noted below.

Most of the attributes listed here may also be specified in the application configuration file, in the same manner as data service attributes. In that case, they will provide default values for the nodes. Attributes for which you wish to give the same value to all (or most) nodes can be conveniently specified in this way, or you can specify them when defining the root node "/". You can always override them at a lower level of the node hierarchy, if you wish.

The node attributes do not have accessor methods. Rather, you can retrieve the value of any node attribute by calling the node_attr method on a data service instance or a request instance:

    # for example, you could use either of the following:
    
    $node_title = $ds->node_attr($node_path, 'title');
    $node_title = $request->node_attr('title');

For example, each operation subroutine is called as a method of a request object. This request object itself has many attributes (with accessors) that are derived from the data service node that matches the request. However, you can if necessary query for arbitrary node attributes:

    sub my_operation {
        my ($request) = @_;
        my $default_limit = $request->node_attr('default_limit');
        ...

Node Lists

In the process of fully documenting a data service, it is important to document the relationships between nodes. For this reason, Web::DataService provides "node lists" which can be used in documentation templates to display lists of related nodes (see NODELIST).

A default node list is created for each node, whose primary purpose is to list all or some of its child nodes. The name of this list is the same as the node path. You can place any node in its parent's list by giving it a place attribute with a value greater than zero. You can also add other nodes to these default lists, or create arbitrarily named lists, by calling the list_node method. See "LIST ELEMENT ATTRIBUTES" in LIST ELEMENT ATTRIBUTES below for a list of the attributes that you can use in defining list elements.

NODE ATTRIBUTES

With the exception of path, each of these attributes is optional.

path

This attribute is required for every node, and must be unique among the nodes defined for this data service. For each incoming request, the URL path and parameters are processed in various ways (depending upon which data service features and special parameters have been enabled) to extract a path which is compared to the set of defined nodes. If one matches, then the attributes of that node will be used to generate the appropriate response. Otherwise, a 404 error ("not found") will be returned to the client.

disabled

If this attribute is given a true value, then it will never match any request. Any request corresponding to this node's path (whether it is an operation request or a documentation request) will return a 404 error. The attribute is inherited, so any children of this node will be likewise disabled. You can use this to define placeholder nodes for functionality to be added later, or to remove existing functions from your application and leave the code in place to be re-activated later.

undocumented

If this attribute is given a true value, then any request asking for documentation about this node will return a 404 error. The node will still be active for requests that ask for its operation. This allows you to provide undocumented data service operations, while explicitly noting that fact in your code.

title

The value of this attribute must be a string, which will be used as the title for any documentation generated about the node. If not specified, the node path will be used instead. This attribute is not inherited. Best practice is to define a specific and informative title for each node.

doc_string

You can set this attribute either directly or by including one or more documentation strings after the node attribute hash in the call to define_node. The default documentation templates use this value as the main description on each documentation page. If you wish to use a longer description than can be easily conveyed in a call to define_node, then create a specific documentation template for this node.

place

If this attribute is given a numeric value greater than zero, then this node will be included in its parent node's default node list. Node lists are ordered by place value and secondarily by order of definition, so you can use the values of this attribute to order the nodes as you choose. This attribute is not inherited.

If you wish to include a node in a list other than the default node list for its parent node, use a separate call to list_node.

list

This attribute is valid with list_node routine, where it specifies the name of a node list in which the specified node will be listed. If the attribute place is also given, it determines the location in the list at which the node will be placed.

method

A node that has both this attribute and the attribute "role" is considered to be an "operation node". The attribute value must be the name of a subroutine (not a code reference) in the package specified by role. An "operation request" that matches this node will result in the creation of a "request object" whose class contains the appropriate role, followed by a method call to the specified subroutine.

role

The value of this attribute must be the name of a Moo Role defined by an already-loaded package. Any operation methods defined for this node and/or its children must occur in that package. All operation nodes must either have this attribute specified explicitly or inherit its value from a parent node.

arg

This attribute is only relevant for operation nodes. If specified, its value will be provided as an argument when the specified method is called to carry out the operation. By means of this attribute, you can arrange for more than one node to call the same method, and have that method behave differently depending upon which argument it receives.

ruleset

This attribute is only relevant for operation nodes. Its value must be the name of a ruleset defined for this data service instance. This ruleset will be automatically used to validate the URL parameters for any request that matches this node, and will also be used in the process of generating documentation about the node.

If the attribute is not specified, a ruleset name will be automatically generated by taking the node path, changing any slashes into colons, and adding the ruleset prefix (if any has been defined for this data service). If the resulting name corresponds to a defined ruleset, that ruleset will be used. You will probably find it convenient to use these auto-generated ruleset names in most cases, and will rarely need to specify this attribute.

output

This attribute is only relevant for operation nodes. Its value must be the name of an output block defined for this data service, or the names of more then one output block separated by commas and optional whitespace. For example:

    output => 'block1, block2'

This block or blocks will make up the fixed output of this node's operation.

output_label

The value attribute will be used to label the fixed output blocks in the generated documentation for this node. Its value must be a string. If not specified, it defaults to basic.

optional_output

If specified, the value of this attribute must be the name of a single output map (in other words, a set) defined for this data service. This will be used, in conjunction with the value of the special parameter show, to select additional output blocks to be included in a response. This attribute is useless unless the special parameter show is enabled, and will only be used when responding to requests that include a value for that parameter.

usage

The value of this attribute must be either a single string or an array of strings. These strings will be used to generate usage examples as part of the documentation for this node. Each string will be parsed to determine the following list of elements:

  • A node path

  • A format

  • A parameter string (optional)

For example, the following string:

    single.json?state=wi

will be parsed as: path="single", format="json", params="state=wi". These elements are then used to construct a data service operation URL based on the set of features and special parameters that are enabled for this data service. Note that you must use this syntax no matter how your data service is configured (i.e. even if you have specified a format parameter instead of format suffix). This attribute is not inherited.

file_path

If this attribute is specified, then this node will be a "file node". The value must be a filename relative to the "public file" directory established by the foundation framework (for Dancer, this is the directory "public" under the application root). A request that exactly matches this node will return the contents of this file, a 404 error if the file does not exist, or a 500 error if it exists but is not readable. It is an error to specify both file_path and file_dir for a single node, or to specify either of them along with method.

file_dir

If this attribute is specified, then this node will be a "file node". The value must be a directory path relative to the "public file" directory (see "file_dir"). A request whose path exactly matches this node will result in a 404 error, but one whose path has this node's path as a prefix will look up the remainder of the path in this directory. If the indicated file exists and is readable, its contents will be returned. If it exists but is not readable, a 500 error will be returned. Otherewise, a 404 error results.

public_access

If this attribute is given a true value, then all response messages generated in association with this node will have the CORS header "Access-control-allow-origin" set to "*". Until we provide better means of controlling the CORS header in a later version of this framework, we suggest that you always set this to true for the root node.

default_format

The value of this attribute must be the name of one of the formats defined for this data service. If no response format can be determined from the request URL and/or parameters, then the specified format will be used for any operation request matching this node. If the data service will only be returning data in a single format, then you should set the value of this attribute in the root node to the name of that format.

default_limit

The value of this attribute will put a limit on the size of the result set for all operation requests matching this node, unless overridden by the special parameter "limit". The purpose of this attribute is to prevent badly-composed requests from accidentally generating an enormous result set. A client can always include limit=all in the request parameters to retrieve the full result set. However, provided that clients leave that parameter off unless needed, this attribute provides a backstop. The value of this attribute must be a positive integer. Unless you want a hard limit that clients cannot override, you should make sure that the special parameter limit is enabled if you use this attribute (it is included in the standard set).

default_header

By default, text format output includes a header unless the client explicitly turns it off by including header=no in the request parameters. If this parameter is set to a false value, then no header will be provided for text format responses matching this node unless explicitly requested by the client using header=yes. Unless you want to disable headers entirely, you should make sure that the special parameter header is enabled if you use this attribute (it is included in the standard set).

default_datainfo

By default, information about the dataset is included in a response only if the client requests it by including datainfo=yes in the request parameters. If this attribute is set to a true value, then this information will be included by default for all operation requests matching this node unless the client specifies datainfo=no (assuming that the special parameter datainfo is active). If you want clients to have control over whether or not this information is provided, you should make sure that the special parameter datainfo is enabled (it is included in the standard set).

default_count

By default, a count of the number of records found and returned is included in a response only if the client requests it by including count=yes in the request parameters. If this attribute is set to a true value, then this information will be included by default for all operation requests matching this node unless the client specifies count=no. If you want clients to have control over whether or not this information is provided, you should make sure that the special parameter count is enabled (it is included in the standard set).

default_linebreak

The value of this attribute must be either 'crlf', 'cr', or 'lf'. If not specified, it defaults to 'crlf'. The specified character sequence will be used to separate the lines of any text format output from requests that match this node, unless overridden by the special parameter linebreak.

default_save_filename

The value of this attribute will used for the 'content-disposition' header of the response message for requests matching this node, if the special parameter save is given with a basic 'true' value and not a filename. The name of the requested response format will automatically be appended as a suffix, so no suffix should be included in the attribute value. For requests given through a web browser, most browsers will offer to save the file under this name.

stream_threshold

The value of this attribute must be a positive integer. It is only relevant if the feature stream_data is enabled for this data service. Any response whose length exceeds the value of this attribute will be streamed to the client instead of sent as a single message. This feature is a good idea to enable for any service that can produce responses of more than a few hundred kilobytes. If the feature is enabled but this attribute is not specified, it defaults to 100Kb.

allow_method

This is a set-valued attribute. The individual values must be HTTP method types (i.e. GET, POST), specifying which HTTP methods are valid for requests matching this node. If GET is allowed, then HEAD is allowed automatically as well. If not specified, then the methods GET and HEAD are allowed.

allow_format

This is a set-valued attribute. The individual values must be the names of response formats defined for this data service, specifying which ones are valid for requests matching this node. If not specified, then all defined formats are allowed.

allow_vocab

This is a set-valued attribute. The individual values must be the names of vocabularies defined for this data service, specifying which ones are valid for requests matching this node. If not specified, then all defined vocabularies are allowed.

doc_template

The value of this attribute must be a file pathname relative to the documentation template directory. The specified template file will be used to respond to any documentation requests matching this node. If not specified, then an automatic path will be constructed by starting with the node path and adding "_doc" followed by the filename suffix specified by the templating plugin. If no file is found under that name, then the node path followed by "/index" and the same suffix is tried. You will probably find it easiest to name your documentation files according to one of these two patterns, so that you will rarely if ever need to specify a value for this attribute.

doc_default_op_template

When a request for documentation matches this node, if the template specified by the "doc_template" attribute is not found, and if the automatic paths are not found either, then the value of this attribute is tried next if this is an operation node. If specified, the attribute value must be a file pathname relative to the documentation template directory. The contents of the template should be a generic "operation documentation" template that can be filled in from the node attributes such as "doc_string". In most cases, you will want to specify this attribute at the root node so that its value will be inherited by all of the other nodes.

doc_default_template

When a request for documentation matches this node, if none of the other template paths correspond to an actual template file on disk, then the value of this attribute will be tried as a final default. The contents of this template might say something like "no documentation can be found". If not specified, the default value is 'doc_not_found' followed by the appropriate suffix for the selected templating engine (i.e. '.tt' for Template Toolkit).

doc_defs

If specified, the value of this attribute must be a file path relative to the documentation template directory. This file will be evaluated before each documentation template is rendered. Its purpose is to define standard elements for use by the documentation template, the header, and the footer. You may set this to the empty string if you do not wish a definition file to be used.

If not specified, the default value is 'doc_defs' followed by the appropriate suffix for the selected templating engine (i.e. '.tt' for Template Toolkit). In most cases, you will want to either use the default or specify this attribute at the root node.

doc_header

If specified, the value of this attribute must be a file path relative to the documentation template directory. This file will be evaluated before each documentation template is rendered, but after the file specified by "doc_defs". Its purpose is to generate a header for the documentation pages. You may set this to the empty string if you do not wish a header to be applied to the documentation pages.

If not specified, the default value is 'doc_header' followed by the appropriate suffix for the selected templating engine (i.e. '.tt' for Template Toolkit). In most cases, you will want to either use the default or specify this attribute at the root node.

If specified, the value of this attribute must be a file path relative to the documentation template directory. This file will be evaluated after each documentation template is rendered. Its purpose is to generate a footer for the documentation pages. You may set this to the empty string if you do not wish a footer to be applied to the documentation pages.

If not specified, the default value is 'doc_footer' followed by the appropriate suffix for the selected templating engine (i.e. '.tt' for Template Toolkit). In most cases, you will want to either use the default or specify this attribute at the root node.

doc_stylesheet

If specified, the value of this attribute must be an absolute or relative URL (not a file path) which should refer to a stylesheet file to go with the documentation pages. If you wish the data service to provide this file, you will need to define a data service node with one of the attributes file_dir or file_path (and enable the send_files feature). Typically, this node should have the path "css" or "css/dsdoc.css", and its attribute should point to a similarly-named subdirectory of the public file directory set up by the foundation framework ("public" in the case of Dancer).

If not specified, the default value is a URL generated using the appropriate pattern for this data service for the node path "/css/dsdoc.css". The default installation of this framework includes an appropriate CSS file under that name, which you can edit however you choose.

LIST ELEMENT ATTRIBUTES

The entries in a node list can have some or all of the following attributes. The only required attributes are path, list and place. Whenever you define a node with define_node and give it a non-zero value for place, a list element is automatically created with these values. This element is placed into the node list corresponding to the newly defined node's path parent.

If you wish to add other (i.e. non-hierarchical) entries to these node lists, or to define your own lists, you can use the list_node method and provide some or all of the following attributes:

path [for list element]

This attribute specifies the path of a "target" node to be displayed in the list. It must correspond to a node of this data service.

list

This attribute specifies the name of the list in which this entry will appear. This name may be a node path, or any other string you choose. If the named list does not already exist, it will be created.

place [for list element]

This attribute specifies where in the list the new record should appear. Its value must be numeric. Node lists are sorted by the value of place, and secondarily by the order in which the nodes were defined. If you give this attribute a value of zero, then this list entry will not be displayed. You may want to do this to create placeholder entries that are inactive for now.

title [for list element]

This optional attribute provides the title for the list entry. If not given, the title of the target node will be used by default.

usage [for list element]

This optional attribute can be used to provide usage examples for the list entry. If not given, the usage examples (if any) of the target node will be used by default.

doc_string [for list element]

This optional attribute can be used to provide a documentation string for the list entry. If not given, the documentation string (if any) of the target node will be used by default. The usual way to set this attribute is by including one or more documentation strings after the list element attribute hash in the call to list_node.

AUTHOR

mmcclenn "at" cpan.org

BUGS

Please report any bugs or feature requests to bug-web-dataservice at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Web-DataService. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

COPYRIGHT & LICENSE

Copyright 2014 Michael McClennen, all rights reserved.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.