Download the PHP package dazet/data-map without Composer

Data Mapper

Library for mapping and transforming data structures.

Defining mapper

Mapper configuration is a description of output structure defined as association:

Key defines property name in output structure and Getter is a function that extracts value from input.

Examples

Getter function

Getter generally can be described as interface:

There are 2 forms of defining map:

• Getter can be string which is shorthand for new GetRaw('key').
• Getter can also be a closure or any other callable. It will receive DataMap\Input\Input as first argument and original input as second argument. Getter interface is not required, it's just a hint.

Predefined Getters

new GetRaw($key, $default)

Get value by property path without additional transformation.

new GetString($key, $default)

Gets value and casts to string (if possible) or returns $default.

new GetInteger($key, $default)

Gets value and casts to integer (if possible) or $default.

new GetFloat($key, $default)

Gets value and casts to float (if possible) or $default.

new GetBoolean($key, $default)

Gets value and casts to boolean (true, false, 0, 1, '0', '1') or $default.

new GetDate($key, $default)

Gets value and transform to \DateTimeImmutable (if possible) or $default.

new GetJoinedStrings($glue, $key1, $key2, ...)

Gets string value for given keys an join it using $glue.

new GetMappedCollection($key, $callback)

Gets collection under given $key and maps it with $callback or return [] if entry cannot be mapped.

new GetMappedFlatCollection($key, $callback)

Similar to GetMappedCollection but result is flattened.

new GetTranslated($key, $map, $default)

Gets value and translates it using provided associative array ($map) or $default when translation for value is not available.

GetFiltered::from('key')->...

Gets value and transforms it through filters pipeline.

Using function as filter:

Regular filters will not be called when value becomes null, with exceptions of ifNull, ifEmpty and withNullable.

Custom null handling filter:

GetFiltered has set of built-in filters similar to FilteredInput.

• with(callable $filter): add to pipeline custom filter functions
• withNullable(callable $filter): add to pipeline custom filter functions that will be called even when value has become null
• string(): try cast to string
• int(): try cast to int
• float(): try cast to float
• bool(): try cast to bool
• array(): try cast to array
• explode(string $delimiter)
• implode(string $glue)
• upper()
• lower()
• trim()
• format(string $template): format value with sprintf template
• replace(string $search, string $replace)
• stripTags()
• numberFormat(int $decimals = 0, string $decimalPoint = '.', string $thousandsSeparator = ',')
• round(int $precision = 0)
• floor()
• ceil()
• date(): try cast to DateTimeImmutable
• dateFormat(string $format)
• count()
• ifNull($default)
• ifEmpty($default)

Input abstraction

Input interface defines common abstraction for accessing data from different data structures, so mapping and getters must not depend of underlying data type.

It also allows to create input decorators for additional input processing, like data filtering, transformation, traversing etc.

ArrayInput

Wraps associative arrays and ArrayAccess objects.

ObjectInput

Wraps generic object and fetches data using object public interface: public properties or getters (a public method without parameters that returns some value).

Access method for key example name is resolved in the following order:

• check for public property name
• check for getter name()
• check for getter getName()
• check for getter isName()

RecursiveInput

RecursiveInput allows to traverse trees od data using dot notation ($input->get('root.branch.leaf')). It decorates Input (current leaf) and requires Wrapper to wrap with proper Input next visited leafs (which can be arrays or objects).

FilteredInput

FilteredInput is another Input decorator that allows to transform data after it is extracted from inner structure.

Default input parser supports given filters:

• string: cast value to string if possible or return null |
• int, integer: cast to integer or return null
• float: cast to float or return null
• bool, boolean: resolve value as boolean or return null
• array: cast value to array if possible (from array or iterable) or return null
• explode [delimiter=","]: explode string using delimiter (, by default)
• implode [delimiter=","]: implode array of strings using delimiter (, by default)
• upper: upper case string
• lower: lower case string
• trim, ltrim, rtrim: trim string
• format: format value as string using sprintf
• replace [search] [replace=""]: replace substring in string like str_replace function
• strip_tags: same as strip_tags function
• number_format [decimals=2] [decimal_point="."] [thousands_separator=","]: same as number_format function
• round [precision=0]: same as round function
• floor: floor value, returns float|null
• ceil: floor value, returns float|null
• datetime: try to transform value to DateTimeImmutable or return null
• date_format [format="Y-m-d H:i:s"]: try to transform value to datetime and format as string or return null when value cannot be transformed
• date_modify [modifier]: try to transform value to DateTimeImmutable and then transform it using modifier $datetime->modify($modifier)
• timestamp: try to transform value to datetime and then to timestamp or return null
• json_encode: encode value to json or return null
• json_decode: decode array from json string or return null when failed
• count: return count for array or Countable or null when not countable
• if_null [then]: return default value when mapped value is null
• if_empty [then]: return default value when mapped value is empty

Examples

• default explode by comma: string | explode
• explode by custom string: string | explode "-"
• default implode by comma: array | implode
• implode by custom string: array | implode "-"
• format string like sprintf: string | format "string: %s"
• format money from float: float | format "price: $%01.2f" - transforms 12.3499 to 'price: $12.35'
• cast to string with default value: maybe_string | string | if_null "default"
• cast to date and modify: date_string | date_modify "+1 day"
• calculate md5 of mapped value: key | string | md5
• wrap string after 20 characters: key | string | wordwrap 20
• using native function with custom argument position of mapped value key | string | preg_replace "/\s+/" " " $$

Function as transformation

Default configuration of InputFilterParser allows use any PHP function as transformation. By default mapped value is passed as first argument to that function optionally followed by other arguments defined in filter config. It is also possible to define different argument position of mapped value using $$ as a placeholder.

Output formatting

Mapping output type depends on Formatter used by Mapper.

Built-in formatters:

ArrayFormatter

Returns associative array which is raw result of Mapper transformation.

ObjectConstructor

Tries to create new instance of object using regular constructor. Keys are matched with constructor parameters by variable name.

There is no value type and correctness checking, so you will get TypeError when mapped types does not match. It also fallback to null value when object constructor has parameter that is not in the mapping.

ObjectHydrator

Tries to hydrate instance of object using his public interface, that is:

• by setting public properties values
• by using setters (setSomething or withSomething assuming immutability)

Customizing and extending

Mapper consists of 3 components:

• GetterMap that describes mapping as string => Getter association,
• Wrapper that wraps input mixed structure with proper Input implementation,
• Formatter that formats raw mapping result (associative array) to array, object, XML, JSON and so on.

Implement Input and Wrapper to extract data from specific sources

It is possible to define data extracting for some object type explicitly.

Use only MixedWrapper for better performance

By default Mapper supports nested structure fetching and value filters, which is nice but has some expense in performance (see BENCHMARK.md). But it is possible to create Mapper only with MixedWrapper when these feature are not needed.

Custom filters for FilteredInput

Filter functions list can be extended or overwritten with own implementation.

Custom Formatter

Custom formatter can be used to achieve better object construction performance than generic object formatters. It is also possible to create formatters creating different result types like XML, JSON etc.