Contains readers for maps where there is no DBKeyCodec for key type.
An Input representing an object field.
Type class for types that can be serialized to Output
(format-agnostic "output stream") and deserialized
from Input
(format-agnostic "input stream").
Type class for types that can be serialized to Output
(format-agnostic "output stream") and deserialized
from Input
(format-agnostic "input stream"). GenCodec
is supposed to capture generic structure of serialized
objects, without being bound to particular format like JSON. The actual format is determined by implementation
of Input
and Output
.
There are convenient macros for automatic derivation of GenCodec
instances (materialize
and materializeRecursively
).
However, GenCodec
instances still need to be explicitly declared and won't be derived "automagically".
If you want fully automatic derivation, use GenCodec.Auto
.
Typeclass which expresses ability to convert between MongoDB JSON object keys and values of some type.
Typeclass which expresses ability to convert between MongoDB JSON object keys and values of some type. Every type which has a natural, unambiguous string representation should have a DBKeyCodec.
Convenience abstract class for companion objects of types that have a GenCodec.
Convenience abstract class for companion objects of types that have a GenCodec. Unfortunately, due to compiler and language limitations this only works for non-generic case classes without default constructor arguments.
Represents an abstract source from which a value may be deserialized (read).
Represents an abstract source from which a value may be deserialized (read).
Each of the read
methods tries to read a value of specified type and may throw an exception
(usually com.avsystem.commons.serialization.GenCodec.ReadFailure) when reading is not successful.
An Input value should be assumed to be stateful. If any of the readX
methods have already been called,
the Input instance can no longer be used and MUST be discarded.
In order to ignore the value kept in this Input, skip()
MUST be called.
In summary: every Input MUST be fully exhausted by either calling one of the read
methods which returns
successful value or by calling skip()
. Also, ListInput and ObjectInput instances returned from this
Input must also be fully exhausted on their own.
Represents the type of value inside and Input that can be read from it.
Represents the type of value inside and Input that can be read from it.
It is possible to distinguish only between four types (null, simple value, object and list) even though any of these types may have different representations. For example, InputType.Simple is returned for more than one actual value types (numbers, booleans, strings, timestamps, binary, etc.).
It's not possible to distinguish between them based only on InputType because not every Input implementation is able to do that. For example, JSON must represent 64-bit integers as strings and therefore it can't distinguish between strings and numbers in general.
Represents an abstract source of sequence of values that can be deserialized.
Represents an abstract source of sequence of values that can be deserialized.
ListInput instance is stateful and MUST be read strictly sequentially. This means, you MUST fully exhaust
an Input instance returned by nextElement()
before calling nextElement()
again. For this reason,
ListInput is not an Iterator
despite having similar interface
(Iterator
would easily allow e.g. conversion to List[Input]
which would be illegal).
ListInput MUST always be fully exhausted. In order to ignore any remaining elements, skipRemaining() may be used.
Represents an abstract sink for serialization of sequences of values.
Represents an abstract sink for serialization of sequences of values. Any ListOutput instance
must be assumed to be stateful and used in strictly sequential manner. After all elements have been written,
finish()
must be called to explicitly mark that the list is complete.
Represents an abstract source of key-value mappings that can be deserialized.
Represents an abstract source of key-value mappings that can be deserialized.
ObjectInput instance is stateful and MUST be read strictly sequentially. This means, you MUST fully exhaust
any Input instance returned by nextField()
before calling nextField()
again. For this reason,
ObjectInput is not an Iterator
despite having similar interface
(Iterator
would easily allow e.g. conversion to List[(String, Input)]
which would be illegal).
ObjectInput MUST always be fully exhausted. In order to ignore any remaining key-value mappings,
skipRemaining()
may be used.
NOTE: The order of keys returned by subsequent invocations of nextField()
may be arbitrary and in particular
may not match the order in which keys were written to corresponding ObjectOutput.
Represents an abstract sink for serialization of string-to-value mappings.
Represents an abstract sink for serialization of string-to-value mappings. Any ObjectOutput instance
must be assumed to be stateful and used in strictly sequential manner. After all key-value pairs have been
written, finish()
must be called to explicitly mark that the object is complete.
Represents an abstract sink to which a value may be serialized (written).
Represents an abstract sink to which a value may be serialized (written).
An Output instance should be assumed to be stateful. After calling any of the write
methods, it MUST NOT be
reused. This means that Output instance can be used only to write a single value. However, if the value
to write is complex, one can use writeList
/writeSet
or writeObject
/writeMap
.
Base trait for outputs which allow writing of multiple values in sequence, i.e.
Base trait for outputs which allow writing of multiple values in sequence, i.e. ListOutput and ObjectOutput.
Author: ghik Created: 19/11/15.
Contains readers for maps where there is no DBKeyCodec for key type. In such case, we assume reading from a list of key-value pairs instead of JSON object.