diff options
Diffstat (limited to 'lv2/atom/atom.meta.ttl')
| -rw-r--r-- | lv2/atom/atom.meta.ttl | 440 |
1 files changed, 439 insertions, 1 deletions
diff --git a/lv2/atom/atom.meta.ttl b/lv2/atom/atom.meta.ttl index 4f6e69b..7b04180 100644 --- a/lv2/atom/atom.meta.ttl +++ b/lv2/atom/atom.meta.ttl @@ -1,6 +1,8 @@ +@prefix atom: <http://lv2plug.in/ns/ext/atom#> . @prefix dcs: <http://ontologi.es/doap-changeset#> . @prefix doap: <http://usefulinc.com/ns/doap#> . @prefix foaf: <http://xmlns.com/foaf/0.1/> . +@prefix lv2: <http://lv2plug.in/ns/lv2core#> . @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> . <http://lv2plug.in/ns/ext/atom> @@ -100,5 +102,441 @@ rdfs:label "Initial release." ] ] - ] . + ] ; + lv2:documentation """ + +An atom:Atom is a simple generic data container for holding any type of Plain +Old Data (POD). An Atom can contain simple primitive types like integers, +floating point numbers, and strings; as well as structured data like lists and +dictionary-like <q>Objects</q>. Since Atoms are POD, they can be easily copied +(for example, with `memcpy()`) anywhere and are suitable for use in real-time +code. + +Every atom starts with an LV2_Atom header, followed by the contents. This +allows code to process atoms without requiring special code for every type of +data. For example, plugins that mutually understand a type can be used +together in a host that does not understand that type, because the host is only +required to copy atoms, not interpret their contents. Similarly, plugins (such +as routers, delays, or data structures) can meaningfully process atoms of a +type unknown to them. + +Atoms should be used anywhere values of various types must be stored or +transmitted. An atom:AtomPort can be used to transmit atoms via ports. An +atom:AtomPort that contains a atom:Sequence can be used for sample accurate +communication of events, such as MIDI. + +### Serialisation + +Each Atom type defines a binary format for use at runtime, but also a +serialisation that is natural to express in Turtle format. Thus, this +specification defines a powerful real-time appropriate data model, as well as a +portable way to serialise any data in that model. This is particularly useful +for inter-process communication, saving/restoring state, and describing values +in plugin data files. + +### Custom Atom Types + +While it is possible to define new Atom types for any binary format, the +standard types defined here are powerful enough to describe almost anything. +Implementations SHOULD build structures out of the types provided here, rather +than define new binary formats (for example, using atom:Object rather than a +new C `struct` type). Host and tool implementations have support for +serialising all standard types, so new binary formats are an implementation +burden which harms interoperabilty. In particular, plugins SHOULD NOT expect +UI communication or state saving with custom binary types to work. In general, +new Atom types should only be defined where absolutely necessary due to +performance reasons and serialisation is not a concern. + +"""^^lv2:Markdown . + +atom:Atom + lv2:documentation """ + +An LV2_Atom has a 32-bit `size` and `type`, followed by a body of `size` bytes. +Atoms MUST be 64-bit aligned. + +All concrete Atom types (subclasses of this class) MUST define a precise binary +layout for their body. + +The `type` field is the URI of an Atom type mapped to an integer. +Implementations SHOULD gracefully pass through, or ignore, atoms with unknown +types. + +All atoms are POD by definition except references, which as a special case have +`type` 0. An Atom MUST NOT contain a Reference. It is safe to copy any +non-reference Atom with a simple `memcpy`, even if the implementation does not +understand `type`. Though this extension reserves the type 0 for references, +the details of reference handling are currently unspecified. A future revision +of this extension, or a different extension, may define how to use non-POD data +and references. Implementations MUST NOT send references to another +implementation unless the receiver is explicitly known to support references +(e.g. by supporting a feature). + +The special case of a null atom with both `type` and `size` 0 is not considered +a reference. + +"""^^lv2:Markdown . + +atom:Chunk + lv2:documentation """ + +This type is used to indicate a certain amount of space is available. For +example, output ports with a variably sized type are connected to a Chunk so +the plugin knows the size of the buffer available for writing. + +The use of a Chunk should be constrained to a local scope, since +interpreting it is impossible without context. However, if serialised to RDF, +a Chunk may be represented directly as an xsd:base64Binary string, for example: + + :::turtle + [] eg:someChunk "vu/erQ=="^^xsd:base64Binary . + +"""^^lv2:Markdown . + +atom:String + lv2:documentation """ + +The body of an LV2_Atom_String is a C string in UTF-8 encoding, i.e. an array +of bytes (`uint8_t`) terminated with a NULL byte (`'\\0'`). + +This type is for free-form strings, but SHOULD NOT be used for typed data or +text in any language. Use atom:Literal unless translating the string does not +make sense and the string has no meaningful datatype. + +"""^^lv2:Markdown . + +atom:Literal + lv2:documentation """ + +This type is compatible with rdfs:Literal and is capable of expressing a +string in any language or a value of any type. A Literal has a +`datatype` and `lang` followed by string data in UTF-8 +encoding. The length of the string data in bytes is `size - +sizeof(LV2_Atom_Literal)`, including the terminating NULL character. The +`lang` field SHOULD be a URI of the form +<http://lexvo.org/id/iso639-3/LANG> or +<http://lexvo.org/id/iso639-1/LANG> where LANG is a 3-character ISO 693-3 +language code, or a 2-character ISO 693-1 language code, respectively. + +A Literal may have a `datatype` or a `lang`, but never both. + +For example, a Literal can be <q>Hello</q> in English: + + :::c + void set_to_hello_in_english(LV2_Atom_Literal* lit) { + lit->atom.type = map(expand("atom:Literal")); + lit->atom.size = 14; + lit->body.datatype = 0; + lit->body.lang = map("http://lexvo.org/id/iso639-1/en"); + memcpy(LV2_ATOM_CONTENTS(LV2_Atom_Literal, lit), + "Hello", + sizeof("Hello")); // Assumes enough space + } + +or a Turtle string: + + :::c + void set_to_turtle_string(LV2_Atom_Literal* lit, const char* ttl) { + lit->atom.type = map(expand("atom:Literal")); + lit->atom.size = 64; + lit->body.datatype = map("http://www.w3.org/2008/turtle#turtle"); + lit->body.lang = 0; + memcpy(LV2_ATOM_CONTENTS(LV2_Atom_Literal, lit), + ttl, + strlen(ttl) + 1); // Assumes enough space + } + +"""^^lv2:Markdown . + +atom:Path + lv2:documentation """ + +A Path is a URI reference with only a path component: no scheme, authority, +query, or fragment. In particular, paths to files in the same bundle may be +cleanly written in Turtle files as a relative URI. However, implementations +may assume any binary Path (e.g. in an event payload) is a valid file path +which can passed to system functions like fopen() directly, without any +character encoding or escape expansion required. + +Any implemenation that creates a Path atom to transmit to another is +responsible for ensuring it is valid. A Path SHOULD always be absolute, unless +there is some mechanism in place that defines a base path. Since this is not +the case for plugin instances, effectively any Path sent to or received from a +plugin instance MUST be absolute. + +"""^^lv2:Markdown . + +atom:URI + lv2:documentation """ + +This is useful when a URI is needed but mapping is inappropriate, for example +with temporary or relative URIs. Since the ability to distinguish URIs from +plain strings is often necessary, URIs MUST NOT be transmitted as atom:String. + +This is not strictly a URI, since UTF-8 is allowed. Escaping and related +issues are the host's responsibility. + +"""^^lv2:Markdown . + +atom:URID + lv2:documentation """ + +A URID is typically generated with the LV2_URID_Map provided by the host . + +"""^^lv2:Markdown . + +atom:Vector + lv2:documentation """ + +A homogeneous series of atom bodies with equivalent type and size. + +An LV2_Atom_Vector is a 32-bit `child_size` and `child_type` followed by `size +/ child_size` atom bodies. + +For example, an atom:Vector containing 42 elements of type atom:Float: + + :::c + struct VectorOf42Floats { + uint32_t size; // sizeof(LV2_Atom_Vector_Body) + (42 * sizeof(float); + uint32_t type; // map(expand("atom:Vector")) + uint32_t child_size; // sizeof(float) + uint32_t child_type; // map(expand("atom:Float")) + float elems[42]; + }; + +Note that it is possible to construct a valid Atom for each element of the +vector, even by an implementation which does not understand `child_type`. + +If serialised to RDF, a Vector SHOULD have the form: + + :::turtle + eg:someVector + a atom:Vector ; + atom:childType atom:Int ; + rdf:value ( + "1"^^xsd:int + "2"^^xsd:int + "3"^^xsd:int + "4"^^xsd:int + ) . + +"""^^lv2:Markdown . + +atom:Tuple + lv2:documentation """ + +The body of a Tuple is simply a series of complete atoms, each aligned to +64 bits. + +If serialised to RDF, a Tuple SHOULD have the form: + + :::turtle + eg:someVector + a atom:Tuple ; + rdf:value ( + "1"^^xsd:int + "3.5"^^xsd:float + "etc" + ) . + +"""^^lv2:Markdown . + +atom:Property + lv2:documentation """ + +An LV2_Atom_Property has a URID `key` and `context`, and an Atom `value`. This +corresponds to an RDF Property, where the <q>key</q> is the <q>predicate</q> +and the <q>value</q> is the object. + +The `context` field can be used to specify a different context for each +property, where this is useful. Otherwise, it may be 0. + +Properties generally only exist as part of an atom:Object. Accordingly, +they will typically be represented directly as properties in RDF (see +atom:Object). If this is not possible, they may be expressed as partial +reified statements, for example: + + :::turtle + eg:someProperty + rdf:predicate eg:theKey ; + rdf:object eg:theValue . + +"""^^lv2:Markdown . + +atom:Object + lv2:documentation """ + +An <q>Object</q> is an atom with a set of properties. This corresponds to an +RDF Resource, and can be thought of as a dictionary with URID keys. + +An LV2_Atom_Object body has a uint32_t `id` and `type`, followed by a series of +atom:Property bodies (LV2_Atom_Property_Body). The LV2_Atom_Object_Body::otype +field is equivalent to a property with key rdf:type, but is included in the +structure to allow for fast dispatching. + +Code SHOULD check for objects using lv2_atom_forge_is_object() or +lv2_atom_forge_is_blank() if a forge is available, rather than checking the +atom type directly. This will correctly handle the deprecated atom:Resource +and atom:Blank types. + +When serialised to RDF, an Object is represented as a resource, for example: + + :::turtle + eg:someObject + eg:firstPropertyKey "first property value" ; + eg:secondPropertyKey "first loser" ; + eg:andSoOn "and so on" . + +"""^^lv2:Markdown . + +atom:Resource + lv2:documentation """ + +This class is deprecated. Use atom:Object directly instead. + +An atom:Object where the <code>id</code> field is a URID, that is, an Object +with a URI. + +"""^^lv2:Markdown . + +atom:Blank + lv2:documentation """ + +This class is deprecated. Use atom:Object with ID 0 instead. + +An atom:Object where the LV2_Atom_Object::id is a blank node ID (NOT a URI). +The ID of a Blank is valid only within the context the Blank appears in. For +ports this is the context of the associated run() call, i.e. all ports share +the same context so outputs can contain IDs that correspond to IDs of blanks in +the input. + +"""^^lv2:Markdown . + +atom:Sound + lv2:documentation """ + +The format of a atom:Sound is the same as the buffer format for lv2:AudioPort +(except the size may be arbitrary). An atom:Sound inherently depends on the +sample rate, which is assumed to be known from context. Because of this, +directly serialising an atom:Sound is probably a bad idea, use a standard +format like WAV instead. + +"""^^lv2:Markdown . + +atom:Event + lv2:documentation """ + +An Event is typically an element of an atom:Sequence. Note that this is not an Atom type since it begins with a timestamp, not an atom header. + +"""^^lv2:Markdown . + +atom:Sequence + lv2:documentation """ + +A flat sequence of atom:Event, that is, a series of time-stamped Atoms. + +LV2_Atom_Sequence_Body.unit describes the time unit for the contained atoms. +If the unit is known from context (e.g. run() stamps are always audio frames), +this field may be zero. Otherwise, it SHOULD be either units:frame or +units:beat, in which case ev.time.frames or ev.time.beats is valid, +respectively. + +If serialised to RDF, a Sequence has a similar form to atom:Vector, but for +brevity the elements may be assumed to be atom:Event, for example: + + :::turtle + eg:someSequence + a atom:Sequence ; + rdf:value ( + [ + atom:frameTime 1 ; + rdf:value "901A01"^^midi:MidiEvent + ] [ + atom:frameTime 3 ; + rdf:value "902B02"^^midi:MidiEvent + ] + ) . + +"""^^lv2:Markdown . + +atom:AtomPort + lv2:documentation """ + +Ports of this type are connected to an LV2_Atom with a type specified by +atom:bufferType. + +Output ports with a variably sized type MUST be initialised by the host before +every run() to an atom:Chunk with size set to the available space. The plugin +reads this size to know how much space is available for writing. In all cases, +the plugin MUST write a complete atom (including header) to outputs. However, +to be robust, hosts SHOULD initialise output ports to a safe sentinel (e.g. the +null Atom) before calling run(). + +"""^^lv2:Markdown . + +atom:bufferType + lv2:documentation """ + +Indicates that an AtomPort may be connected to a certain Atom type. A port MAY +support several buffer types. The host MUST NOT connect a port to an Atom with +a type not explicitly listed with this property. The value of this property +MUST be a sub-class of atom:Atom. For example, an input port that is connected +directly to an LV2_Atom_Double value is described like so: + + :::turtle + <plugin> + lv2:port [ + a lv2:InputPort , atom:AtomPort ; + atom:bufferType atom:Double ; + ] . + +This property only describes the types a port may be directly connected to. It +says nothing about the expected contents of containers. For that, use +atom:supports. + +"""^^lv2:Markdown . + +atom:supports + lv2:documentation """ + +This property is defined loosely, it may be used to indicate that anything +<q>supports</q> an Atom type, wherever that may be useful. It applies +<q>recursively</q> where collections are involved. + +In particular, this property can be used to describe which event types are +expected by a port. For example, a port that receives MIDI events is described +like so: + + :::turtle + <plugin> + lv2:port [ + a lv2:InputPort , atom:AtomPort ; + atom:bufferType atom:Sequence ; + atom:supports midi:MidiEvent ; + ] . + +"""^^lv2:Markdown . + +atom:eventTransfer + lv2:documentation """ + +Transfer of individual events in a port buffer. Useful as the `format` for a +LV2UI_Write_Function. + +This protocol applies to ports which contain events, usually in an +atom:Sequence. The host must transfer each individual event to the recipient. +The format of the received data is an LV2_Atom, there is no timestamp header. + +"""^^lv2:Markdown . + +atom:atomTransfer + lv2:documentation """ + +Transfer of the complete atom in a port buffer. Useful as the `format` for a +LV2UI_Write_Function. + +This protocol applies to atom ports. The host must transfer the complete atom +contained in the port, including header. + +"""^^lv2:Markdown . |