README

About hydrating

This package aims to help handling raw data (associative arrays of basic types) for editing and creating instances of whatever-the-class-you-want. Therefore, a class offering a solution to this purpose should implement interface Hydrator, ie methods hydrateObject and hydrateNewObject.

• method hydrateNewObject should create an instance of requested class, using parsed data to set its values.
• method hydrateObject should edit a given object, setting its properties using parsed data. This package offers a generic implementation of Hydrator, MetaHydrator, designed to work specifically with Mouf.

How to use?

The main interest of this package is to avoid unpleasant and interminable hours of writing repetitive lines of code to describe how to manage some innput data (ie validate, parse and instantiate from it), and instead make the creation of a data handler a simple configuration task. Configuration, or should I use the terms "dependency injection". This is where the power of Mouf comes particularly handy: this tool, featuring present library, will allow you to create your data parsers/validators with only drag/drop/naming action, occasionally writing little pieces of code (if you wish to implement your own atomic parsers or validators), but without duplication.

How does it work?

Data parsing

Class MetaHydrator uses instances of HydratingHandlerInterface to parse and validate raw input. Such an instance should parse specific key(s) in input array (using a configured instance of ParserInterface), and may throw an exception (Invalid). In SimpleHydratingHandler implementation of HydratingHandlerInterface, one specific key in input array data is handled, ie the value located at that key is parsed and, if parsing was successful, parsed value is validated. Keep in mind that this validation should not consider other parsed values, since a handler does not wait for all values to be parsed before checking for its handled value sanity.

Data validation

An implementation of interface ValidatorInterface must implement method validate; this method will do nothing if input data is correct, and shall throw a InvalidValueException if it is not. Such an exception needs an innerErrorsMap when constructed. A well-formed errors map should be an associative array, keys being strings, and values each being either a descriptor of the field error (such as a DetailedErrorMessage, for instance) or a well-formed errors map. More precisely, its structure should be consistent with input data, where invalid fields would be replaced by those error descriptors.

Applying parsed data

Using an inner simple hydrator (by default, TDBMHydrator seems to be a good choice), the parsed data can finally be used in two different ways: You can whether apply it to an already existing object, (using implementation of method hydrateObject) of your hydrator, or create a new instance of the class you wish to instanciate (method hydrateNewObject).