chris@16: /*
chris@16:   Copyright 2011-2012 David Robillard <http://drobilla.net>
chris@16: 
chris@16:   Permission to use, copy, modify, and/or distribute this software for any
chris@16:   purpose with or without fee is hereby granted, provided that the above
chris@16:   copyright notice and this permission notice appear in all copies.
chris@16: 
chris@16:   THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
chris@16:   WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
chris@16:   MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
chris@16:   ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
chris@16:   WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
chris@16:   ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
chris@16:   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
chris@16: */
chris@16: 
chris@16: /**
chris@16:    @file serd.h API for Serd, a lightweight RDF syntax library.
chris@16: */
chris@16: 
chris@16: #ifndef SERD_SERD_H
chris@16: #define SERD_SERD_H
chris@16: 
chris@16: #include <stdarg.h>
chris@16: #include <stddef.h>
chris@16: #include <stdint.h>
chris@16: #include <stdio.h>
chris@16: 
chris@16: #ifdef SERD_SHARED
chris@16: #    ifdef _WIN32
chris@16: #        define SERD_LIB_IMPORT __declspec(dllimport)
chris@16: #        define SERD_LIB_EXPORT __declspec(dllexport)
chris@16: #    else
chris@16: #        define SERD_LIB_IMPORT __attribute__((visibility("default")))
chris@16: #        define SERD_LIB_EXPORT __attribute__((visibility("default")))
chris@16: #    endif
chris@16: #    ifdef SERD_INTERNAL
chris@16: #        define SERD_API SERD_LIB_EXPORT
chris@16: #    else
chris@16: #        define SERD_API SERD_LIB_IMPORT
chris@16: #    endif
chris@16: #else
chris@16: #    define SERD_API
chris@16: #endif
chris@16: 
chris@16: #ifdef __cplusplus
chris@16: extern "C" {
chris@16: #else
chris@16: #    include <stdbool.h>
chris@16: #endif
chris@16: 
chris@16: /**
chris@16:    @defgroup serd Serd
chris@16:    A lightweight RDF syntax library.
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Environment.
chris@16: 
chris@16:    Represents the state required to resolve a CURIE or relative URI, e.g. the
chris@16:    base URI and set of namespace prefixes at a particular point.
chris@16: */
chris@16: typedef struct SerdEnvImpl SerdEnv;
chris@16: 
chris@16: /**
chris@16:    RDF reader.
chris@16: 
chris@16:    Parses RDF by calling user-provided sink functions as input is consumed
chris@16:    (much like an XML SAX parser).
chris@16: */
chris@16: typedef struct SerdReaderImpl SerdReader;
chris@16: 
chris@16: /**
chris@16:    RDF writer.
chris@16: 
chris@16:    Provides a number of functions to allow writing RDF syntax out to some
chris@16:    stream.  These functions are deliberately compatible with the sink functions
chris@16:    used by SerdReader, so a reader can be directly connected to a writer to
chris@16:    re-serialise a document with minimal overhead.
chris@16: */
chris@16: typedef struct SerdWriterImpl SerdWriter;
chris@16: 
chris@16: /**
chris@16:    Return status code.
chris@16: */
chris@16: typedef enum {
chris@16: 	SERD_SUCCESS,         /**< No error */
chris@16: 	SERD_FAILURE,         /**< Non-fatal failure */
chris@16: 	SERD_ERR_UNKNOWN,     /**< Unknown error */
chris@16: 	SERD_ERR_BAD_SYNTAX,  /**< Invalid syntax */
chris@16: 	SERD_ERR_BAD_ARG,     /**< Invalid argument */
chris@16: 	SERD_ERR_NOT_FOUND,   /**< Not found */
chris@16: 	SERD_ERR_ID_CLASH,    /**< Encountered clashing blank node IDs */
chris@16: 	SERD_ERR_BAD_CURIE,   /**< Invalid CURIE (e.g. prefix does not exist) */
chris@16: 	SERD_ERR_INTERNAL     /**< Unexpected internal error (should not happen) */
chris@16: } SerdStatus;
chris@16: 
chris@16: /**
chris@16:    RDF syntax type.
chris@16: */
chris@16: typedef enum {
chris@16: 	/**
chris@16: 	   Turtle - Terse RDF Triple Language (UTF-8).
chris@16: 	   @see <a href="http://www.w3.org/TeamSubmission/turtle/">Turtle</a>
chris@16: 	*/
chris@16: 	SERD_TURTLE = 1,
chris@16: 
chris@16: 	/**
chris@16: 	   NTriples - Line-based RDF triples (ASCII).
chris@16: 	   @see <a href="http://www.w3.org/TR/rdf-testcases#ntriples">NTriples</a>
chris@16: 	*/
chris@16: 	SERD_NTRIPLES = 2
chris@16: } SerdSyntax;
chris@16: 
chris@16: /**
chris@16:    Flags indication inline abbreviation information for a statement.
chris@16: */
chris@16: typedef enum {
chris@16: 	SERD_EMPTY_S      = 1 << 1,  /**< Empty blank node subject */
chris@16: 	SERD_EMPTY_O      = 1 << 2,  /**< Empty blank node object */
chris@16: 	SERD_ANON_S_BEGIN = 1 << 3,  /**< Start of anonymous subject */
chris@16: 	SERD_ANON_O_BEGIN = 1 << 4,  /**< Start of anonymous object */
chris@16: 	SERD_ANON_CONT    = 1 << 5,  /**< Continuation of anonymous node */
chris@16: 	SERD_LIST_S_BEGIN = 1 << 6,  /**< Start of list subject */
chris@16: 	SERD_LIST_O_BEGIN = 1 << 7,  /**< Start of list object */
chris@16: 	SERD_LIST_CONT    = 1 << 8   /**< Continuation of list */
chris@16: } SerdStatementFlag;
chris@16: 
chris@16: /**
chris@16:    Bitwise OR of SerdNodeFlag values.
chris@16: */
chris@16: typedef uint32_t SerdStatementFlags;
chris@16: 
chris@16: /**
chris@16:    Type of a syntactic RDF node.
chris@16: 
chris@16:    This is more precise than the type of an abstract RDF node.  An abstract
chris@16:    node is either a resource, literal, or blank.  In syntax there are two ways
chris@16:    to refer to a resource (by URI or CURIE) and two ways to refer to a blank
chris@16:    (by ID or anonymously).  Anonymous (inline) blank nodes are expressed using
chris@16:    SerdStatementFlags rather than this type.
chris@16: */
chris@16: typedef enum {
chris@16: 	/**
chris@16: 	   The type of a nonexistent node.
chris@16: 
chris@16: 	   This type is useful as a sentinel, but is never emitted by the reader.
chris@16: 	*/
chris@16: 	SERD_NOTHING = 0,
chris@16: 
chris@16: 	/**
chris@16: 	   Literal value.
chris@16: 
chris@16: 	   A literal optionally has either a language, or a datatype (not both).
chris@16: 	*/
chris@16: 	SERD_LITERAL = 1,
chris@16: 
chris@16: 	/**
chris@16: 	   URI (absolute or relative).
chris@16: 
chris@16: 	   Value is an unquoted URI string, which is either a relative reference
chris@16: 	   with respect to the current base URI (e.g. "foo/bar"), or an absolute
chris@16: 	   URI (e.g. "http://example.org/foo").
chris@16: 	   @see <a href="http://tools.ietf.org/html/rfc3986">RFC3986</a>.
chris@16: 	*/
chris@16: 	SERD_URI = 2,
chris@16: 
chris@16: 	/**
chris@16: 	   CURIE, a shortened URI.
chris@16: 
chris@16: 	   Value is an unquoted CURIE string relative to the current environment,
chris@16: 	   e.g. "rdf:type".
chris@16: 	   @see <a href="http://www.w3.org/TR/curie">CURIE Syntax 1.0</a>
chris@16: 	*/
chris@16: 	SERD_CURIE = 3,
chris@16: 
chris@16: 	/**
chris@16: 	   A blank node.
chris@16: 
chris@16: 	   Value is a blank node ID, e.g. "id3", which is meaningful only within
chris@16: 	   this serialisation.
chris@16: 	   @see <a href="http://www.w3.org/TeamSubmission/turtle#nodeID">Turtle
chris@16: 	   <tt>nodeID</tt></a>
chris@16: 	*/
chris@16: 	SERD_BLANK = 4,
chris@16: 
chris@16: } SerdType;
chris@16: 
chris@16: /**
chris@16:    Flags indicating certain string properties relevant to serialisation.
chris@16: */
chris@16: typedef enum {
chris@16: 	SERD_HAS_NEWLINE = 1,      /**< Contains line breaks ('\\n' or '\\r') */
chris@16: 	SERD_HAS_QUOTE   = 1 << 1  /**< Contains quotes ('"') */
chris@16: } SerdNodeFlag;
chris@16: 
chris@16: /**
chris@16:    Bitwise OR of SerdNodeFlag values.
chris@16: */
chris@16: typedef uint32_t SerdNodeFlags;
chris@16: 
chris@16: /**
chris@16:    A syntactic RDF node.
chris@16: */
chris@16: typedef struct {
chris@16: 	const uint8_t* buf;      /**< Value string */
chris@16: 	size_t         n_bytes;  /**< Size in bytes (not including null) */
chris@16: 	size_t         n_chars;  /**< Length in characters (not including null)*/
chris@16: 	SerdNodeFlags  flags;    /**< Node flags (e.g. string properties) */
chris@16: 	SerdType       type;     /**< Node type */
chris@16: } SerdNode;
chris@16: 
chris@16: /**
chris@16:    An unterminated string fragment.
chris@16: */
chris@16: typedef struct {
chris@16: 	const uint8_t* buf;  /**< Start of chunk */
chris@16: 	size_t         len;  /**< Length of chunk in bytes */
chris@16: } SerdChunk;
chris@16: 
chris@16: /**
chris@16:    An error description.
chris@16: */
chris@16: typedef struct {
chris@16: 	SerdStatus     status;    /**< Error code */
chris@16: 	const uint8_t* filename;  /**< File where error was encountered, or NULL */
chris@16: 	unsigned       line;      /**< Line where error was encountered, or 0 */
chris@16: 	unsigned       col;       /**< Column where error was encountered */
chris@16: 	const char*    fmt;       /**< Message format string (printf style) */
chris@16: 	va_list*       args;      /**< Arguments for fmt */
chris@16: } SerdError;
chris@16: 
chris@16: /**
chris@16:    A parsed URI.
chris@16: 
chris@16:    This struct directly refers to chunks in other strings, it does not own any
chris@16:    memory itself.  Thus, URIs can be parsed and/or resolved against a base URI
chris@16:    in-place without allocating memory.
chris@16: */
chris@16: typedef struct {
chris@16: 	SerdChunk scheme;     /**< Scheme */
chris@16: 	SerdChunk authority;  /**< Authority */
chris@16: 	SerdChunk path_base;  /**< Path prefix if relative */
chris@16: 	SerdChunk path;       /**< Path suffix */
chris@16: 	SerdChunk query;      /**< Query */
chris@16: 	SerdChunk fragment;   /**< Fragment */
chris@16: } SerdURI;
chris@16: 
chris@16: /**
chris@16:    Syntax style options.
chris@16: 
chris@16:    The style of the writer output can be controlled by ORing together
chris@16:    values from this enumeration.  Note that some options are only supported
chris@16:    for some syntaxes (e.g. NTriples does not support abbreviation and is
chris@16:    always ASCII).
chris@16: */
chris@16: typedef enum {
chris@16: 	SERD_STYLE_ABBREVIATED = 1,       /**< Abbreviate triples when possible. */
chris@16: 	SERD_STYLE_ASCII       = 1 << 1,  /**< Escape all non-ASCII characters. */
chris@16: 	SERD_STYLE_RESOLVED    = 1 << 2,  /**< Resolve URIs against base URI. */
chris@16: 	SERD_STYLE_CURIED      = 1 << 3,  /**< Shorten URIs into CURIEs. */
chris@16: 	SERD_STYLE_BULK        = 1 << 4   /**< Write output in pages. */
chris@16: } SerdStyle;
chris@16: 
chris@16: /**
chris@16:    @name String Utilities
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Return a string describing a status code.
chris@16: */
chris@16: SERD_API
chris@16: const uint8_t*
chris@16: serd_strerror(SerdStatus status);
chris@16: 
chris@16: /**
chris@16:    Measure a UTF-8 string.
chris@16:    @return Length of @c str in characters (except NULL).
chris@16:    @param str A null-terminated UTF-8 string.
chris@16:    @param n_bytes (Output) Set to the size of @c str in bytes (except NULL).
chris@16:    @param flags (Output) Set to the applicable flags.
chris@16: */
chris@16: SERD_API
chris@16: size_t
chris@16: serd_strlen(const uint8_t* str, size_t* n_bytes, SerdNodeFlags* flags);
chris@16: 
chris@16: /**
chris@16:    Parse a string to a double.
chris@16: 
chris@16:    The API of this function is identical to the standard C strtod function,
chris@16:    except this function is locale-independent and always matches the lexical
chris@16:    format used in the Turtle grammar (the decimal point is always ".").
chris@16: */
chris@16: SERD_API
chris@16: double
chris@16: serd_strtod(const char* str, char** endptr);
chris@16: 
chris@16: /**
chris@16:    Decode a base64 string.
chris@16:    This function can be used to deserialise a blob node created with
chris@16:    serd_node_new_blob().
chris@16: 
chris@16:    @param str Base64 string to decode.
chris@16:    @param len The length of @c str.
chris@16:    @param size Set to the size of the returned blob in bytes.
chris@16:    @return A newly allocated blob which must be freed with free().
chris@16: */
chris@16: SERD_API
chris@16: void*
chris@16: serd_base64_decode(const uint8_t* str, size_t len, size_t* size);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name URI
chris@16:    @{
chris@16: */
chris@16: 
chris@16: static const SerdURI SERD_URI_NULL = {{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}};
chris@16: 
chris@16: /**
chris@16:    Return the local path for @c uri, or NULL if @c uri is not a file URI.
chris@16:    Note this (inappropriately named) function only removes the file scheme if
chris@16:    necessary, and returns @c uri unmodified if it is an absolute path.  Percent
chris@16:    encoding and other issues are not handled, to properly convert a file URI to
chris@16:    a path, use serd_file_uri_parse().
chris@16: */
chris@16: SERD_API
chris@16: const uint8_t*
chris@16: serd_uri_to_path(const uint8_t* uri);
chris@16: 
chris@16: /**
chris@16:    Get the unescaped path and hostname from a file URI.
chris@16:    @param uri A file URI.
chris@16:    @param hostname If non-NULL, set to the hostname, if present.
chris@16:    @return The path component of the URI.
chris@16: 
chris@16:    Both the returned path and @c hostname (if applicable) are owned by the
chris@16:    caller and must be freed with free().
chris@16: */
chris@16: SERD_API
chris@16: uint8_t*
chris@16: serd_file_uri_parse(const uint8_t* uri, uint8_t** hostname);
chris@16: 
chris@16: /**
chris@16:    Return true iff @c utf8 starts with a valid URI scheme.
chris@16: */
chris@16: SERD_API
chris@16: bool
chris@16: serd_uri_string_has_scheme(const uint8_t* utf8);
chris@16: 
chris@16: /**
chris@16:    Parse @c utf8, writing result to @c out.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_uri_parse(const uint8_t* utf8, SerdURI* out);
chris@16: 
chris@16: /**
chris@16:    Set @c out to @c uri resolved against @c base.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_uri_resolve(const SerdURI* uri, const SerdURI* base, SerdURI* out);
chris@16: 
chris@16: /**
chris@16:    Sink function for raw string output.
chris@16: */
chris@16: typedef size_t (*SerdSink)(const void* buf, size_t len, void* stream);
chris@16: 
chris@16: /**
chris@16:    Serialise @c uri with a series of calls to @c sink.
chris@16: */
chris@16: SERD_API
chris@16: size_t
chris@16: serd_uri_serialise(const SerdURI* uri, SerdSink sink, void* stream);
chris@16: 
chris@16: /**
chris@16:    Serialise @c uri relative to @c base with a series of calls to @c sink.
chris@16: 
chris@16:    The @c uri is written as a relative URI iff if it a child of @c base and @c
chris@16:    root.  The optional @c root parameter must be a prefix of @c base and can be
chris@16:    used keep up-references ("../") within a certain namespace.
chris@16: */
chris@16: SERD_API
chris@16: size_t
chris@16: serd_uri_serialise_relative(const SerdURI* uri,
chris@16:                             const SerdURI* base,
chris@16:                             const SerdURI* root,
chris@16:                             SerdSink       sink,
chris@16:                             void*          stream);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name Node
chris@16:    @{
chris@16: */
chris@16: 
chris@16: static const SerdNode SERD_NODE_NULL = { 0, 0, 0, 0, SERD_NOTHING };
chris@16: 
chris@16: /**
chris@16:    Make a (shallow) node from @c str.
chris@16: 
chris@16:    This measures, but does not copy, @c str.  No memory is allocated.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_from_string(SerdType type, const uint8_t* str);
chris@16: 
chris@16: /**
chris@16:    Make a deep copy of @c node.
chris@16: 
chris@16:    @return a node that the caller must free with @ref serd_node_free.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_copy(const SerdNode* node);
chris@16: 
chris@16: /**
chris@16:    Return true iff @c a is equal to @c b.
chris@16: */
chris@16: SERD_API
chris@16: bool
chris@16: serd_node_equals(const SerdNode* a, const SerdNode* b);
chris@16: 
chris@16: /**
chris@16:    Simple wrapper for serd_node_new_uri to resolve a URI node.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_uri_from_node(const SerdNode* uri_node,
chris@16:                             const SerdURI*  base,
chris@16:                             SerdURI*        out);
chris@16: 
chris@16: /**
chris@16:    Simple wrapper for serd_node_new_uri to resolve a URI string.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_uri_from_string(const uint8_t* str,
chris@16:                               const SerdURI* base,
chris@16:                               SerdURI*       out);
chris@16: 
chris@16: /**
chris@16:    Create a new file URI node from a file system path and optional hostname.
chris@16: 
chris@16:    Backslashes in Windows paths will be converted and '%' will always be
chris@16:    percent encoded.  If @c escape is true, all other invalid characters will be
chris@16:    percent encoded as well.
chris@16: 
chris@16:    If @c path is relative, @c hostname is ignored.
chris@16:    If @c out is not NULL, it will be set to the parsed URI.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_file_uri(const uint8_t* path,
chris@16:                        const uint8_t* hostname,
chris@16:                        SerdURI*       out,
chris@16:                        bool           escape);
chris@16: 
chris@16: /**
chris@16:    Create a new node by serialising @c uri into a new string.
chris@16: 
chris@16:    @param uri The URI to parse and serialise.
chris@16: 
chris@16:    @param base Base URI to resolve @c uri against (or NULL for no resolution).
chris@16: 
chris@16:    @param out Set to the parsing of the new URI (i.e. points only to
chris@16:    memory owned by the new returned node).
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_uri(const SerdURI* uri, const SerdURI* base, SerdURI* out);
chris@16: 
chris@16: /**
chris@16:    Create a new node by serialising @c d into an xsd:decimal string.
chris@16: 
chris@16:    The resulting node will always contain a `.', start with a digit, and end
chris@16:    with a digit (i.e. will have a leading and/or trailing `0' if necessary).
chris@16:    It will never be in scientific notation.  A maximum of @c frac_digits digits
chris@16:    will be written after the decimal point, but trailing zeros will
chris@16:    automatically be omitted (except one if @c d is a round integer).
chris@16: 
chris@16:    Note that about 16 and 8 fractional digits are required to precisely
chris@16:    represent a double and float, respectively.
chris@16: 
chris@16:    @param d The value for the new node.
chris@16:    @param frac_digits The maximum number of digits after the decimal place.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_decimal(double d, unsigned frac_digits);
chris@16: 
chris@16: /**
chris@16:    Create a new node by serialising @c i into an xsd:integer string.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_integer(int64_t i);
chris@16: 
chris@16: /**
chris@16:    Create a node by serialising @c buf into an xsd:base64Binary string.
chris@16:    This function can be used to make a serialisable node out of arbitrary
chris@16:    binary data, which can be decoded using serd_base64_decode().
chris@16: 
chris@16:    @param buf Raw binary input data.
chris@16:    @param size Size of @c buf.
chris@16:    @param wrap_lines Wrap lines at 76 characters to conform to RFC 2045.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_node_new_blob(const void* buf, size_t size, bool wrap_lines);
chris@16: 
chris@16: /**
chris@16:    Free any data owned by @c node.
chris@16: 
chris@16:    Note that if @c node is itself dynamically allocated (which is not the case
chris@16:    for nodes created internally by serd), it will not be freed.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_node_free(SerdNode* node);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name Event Handlers
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Sink (callback) for errors.
chris@16: 
chris@16:    @param handle Handle for user data.
chris@16:    @param error Error description.
chris@16: */
chris@16: typedef SerdStatus (*SerdErrorSink)(void*            handle,
chris@16:                                     const SerdError* error);
chris@16: 
chris@16: /**
chris@16:    Sink (callback) for base URI changes.
chris@16: 
chris@16:    Called whenever the base URI of the serialisation changes.
chris@16: */
chris@16: typedef SerdStatus (*SerdBaseSink)(void*           handle,
chris@16:                                    const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Sink (callback) for namespace definitions.
chris@16: 
chris@16:    Called whenever a prefix is defined in the serialisation.
chris@16: */
chris@16: typedef SerdStatus (*SerdPrefixSink)(void*           handle,
chris@16:                                      const SerdNode* name,
chris@16:                                      const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Sink (callback) for statements.
chris@16: 
chris@16:    Called for every RDF statement in the serialisation.
chris@16: */
chris@16: typedef SerdStatus (*SerdStatementSink)(void*              handle,
chris@16:                                         SerdStatementFlags flags,
chris@16:                                         const SerdNode*    graph,
chris@16:                                         const SerdNode*    subject,
chris@16:                                         const SerdNode*    predicate,
chris@16:                                         const SerdNode*    object,
chris@16:                                         const SerdNode*    object_datatype,
chris@16:                                         const SerdNode*    object_lang);
chris@16: 
chris@16: /**
chris@16:    Sink (callback) for anonymous node end markers.
chris@16: 
chris@16:    This is called to indicate that the anonymous node with the given
chris@16:    @c value will no longer be referred to by any future statements
chris@16:    (i.e. the anonymous serialisation of the node is finished).
chris@16: */
chris@16: typedef SerdStatus (*SerdEndSink)(void*           handle,
chris@16:                                   const SerdNode* node);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name Environment
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Create a new environment.
chris@16: */
chris@16: SERD_API
chris@16: SerdEnv*
chris@16: serd_env_new(const SerdNode* base_uri);
chris@16: 
chris@16: /**
chris@16:    Free @c ns.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_env_free(SerdEnv* env);
chris@16: 
chris@16: /**
chris@16:    Get the current base URI.
chris@16: */
chris@16: SERD_API
chris@16: const SerdNode*
chris@16: serd_env_get_base_uri(const SerdEnv* env,
chris@16:                       SerdURI*       out);
chris@16: 
chris@16: /**
chris@16:    Set the current base URI.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_env_set_base_uri(SerdEnv*        env,
chris@16:                       const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Set a namespace prefix.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_env_set_prefix(SerdEnv*        env,
chris@16:                     const SerdNode* name,
chris@16:                     const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Set a namespace prefix.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_env_set_prefix_from_strings(SerdEnv*       env,
chris@16:                                  const uint8_t* name,
chris@16:                                  const uint8_t* uri);
chris@16: 
chris@16: /**
chris@16:    Qualify @c uri into a CURIE if possible.
chris@16: */
chris@16: SERD_API
chris@16: bool
chris@16: serd_env_qualify(const SerdEnv*  env,
chris@16:                  const SerdNode* uri,
chris@16:                  SerdNode*       prefix,
chris@16:                  SerdChunk*      suffix);
chris@16: 
chris@16: /**
chris@16:    Expand @c curie.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_env_expand(const SerdEnv*  env,
chris@16:                 const SerdNode* curie,
chris@16:                 SerdChunk*      uri_prefix,
chris@16:                 SerdChunk*      uri_suffix);
chris@16: 
chris@16: /**
chris@16:    Expand @c node, which must be a CURIE or URI, to a full URI.
chris@16: */
chris@16: SERD_API
chris@16: SerdNode
chris@16: serd_env_expand_node(const SerdEnv*  env,
chris@16:                      const SerdNode* node);
chris@16: 
chris@16: /**
chris@16:    Call @c func for each prefix defined in @c env.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_env_foreach(const SerdEnv* env,
chris@16:                  SerdPrefixSink func,
chris@16:                  void*          handle);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name Reader
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Create a new RDF reader.
chris@16: */
chris@16: SERD_API
chris@16: SerdReader*
chris@16: serd_reader_new(SerdSyntax        syntax,
chris@16:                 void*             handle,
chris@16:                 void              (*free_handle)(void*),
chris@16:                 SerdBaseSink      base_sink,
chris@16:                 SerdPrefixSink    prefix_sink,
chris@16:                 SerdStatementSink statement_sink,
chris@16:                 SerdEndSink       end_sink);
chris@16: 
chris@16: /**
chris@16:    Set a function to be called when errors occur during reading.
chris@16: 
chris@16:    The @p error_sink will be called with @p handle as its first argument.  If
chris@16:    no error function is set, errors are printed to stderr in GCC style.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_reader_set_error_sink(SerdReader*   reader,
chris@16:                            SerdErrorSink error_sink,
chris@16:                            void*         handle);
chris@16: 
chris@16: /**
chris@16:    Return the @c handle passed to @ref serd_reader_new.
chris@16: */
chris@16: SERD_API
chris@16: void*
chris@16: serd_reader_get_handle(const SerdReader* reader);
chris@16: 
chris@16: /**
chris@16:    Set a prefix to be added to all blank node identifiers.
chris@16: 
chris@16:    This is useful when multiple files are to be parsed into the same output
chris@16:    (e.g. a store, or other files).  Since Serd preserves blank node IDs, this
chris@16:    could cause conflicts where two non-equivalent blank nodes are merged,
chris@16:    resulting in corrupt data.  By setting a unique blank node prefix for each
chris@16:    parsed file, this can be avoided, while preserving blank node names.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_reader_add_blank_prefix(SerdReader*    reader,
chris@16:                              const uint8_t* prefix);
chris@16: 
chris@16: /**
chris@16:    Set the URI of the default graph.
chris@16: 
chris@16:    If this is set, the reader will emit quads with the graph set to the given
chris@16:    node for any statements that are not in a named graph (which is currently
chris@16:    all of them since Serd currently does not support any graph syntaxes).
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_reader_set_default_graph(SerdReader*     reader,
chris@16:                               const SerdNode* graph);
chris@16: 
chris@16: /**
chris@16:    Read a file at a given @c uri.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_read_file(SerdReader*    reader,
chris@16:                       const uint8_t* uri);
chris@16: 
chris@16: /**
chris@16:    Start an incremental read from a file handle.
chris@16: 
chris@16:    Iff @p bulk is true, @p file will be read a page at a time.  This is more
chris@16:    efficient, but uses a page of memory and means that an entire page of input
chris@16:    must be ready before any callbacks will fire.  To react as soon as input
chris@16:    arrives, set @p bulk to false.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_start_stream(SerdReader*    me,
chris@16:                          FILE*          file,
chris@16:                          const uint8_t* name,
chris@16:                          bool           bulk);
chris@16: 
chris@16: /**
chris@16:    Read a single "chunk" of data during an incremental read.
chris@16: 
chris@16:    This function will read a single top level description, and return.  This
chris@16:    may be a directive, statement, or several statements; essentially it reads
chris@16:    until a '.' is encountered.  This is particularly useful for reading
chris@16:    directly from a pipe or socket.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_read_chunk(SerdReader* me);
chris@16: 
chris@16: /**
chris@16:    Finish an incremental read from a file handle.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_end_stream(SerdReader* me);
chris@16: 
chris@16: /**
chris@16:    Read @c file.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_read_file_handle(SerdReader*    reader,
chris@16:                              FILE*          file,
chris@16:                              const uint8_t* name);
chris@16: 
chris@16: /**
chris@16:    Read @c utf8.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_reader_read_string(SerdReader* me, const uint8_t* utf8);
chris@16: 
chris@16: /**
chris@16:    Free @c reader.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_reader_free(SerdReader* reader);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @name Writer
chris@16:    @{
chris@16: */
chris@16: 
chris@16: /**
chris@16:    Create a new RDF writer.
chris@16: */
chris@16: SERD_API
chris@16: SerdWriter*
chris@16: serd_writer_new(SerdSyntax     syntax,
chris@16:                 SerdStyle      style,
chris@16:                 SerdEnv*       env,
chris@16:                 const SerdURI* base_uri,
chris@16:                 SerdSink       sink,
chris@16:                 void*          stream);
chris@16: 
chris@16: /**
chris@16:    Free @c writer.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_writer_free(SerdWriter* writer);
chris@16: 
chris@16: /**
chris@16:    Return the env used by @c writer.
chris@16: */
chris@16: SERD_API
chris@16: SerdEnv*
chris@16: serd_writer_get_env(SerdWriter* writer);
chris@16: 
chris@16: /**
chris@16:    A convenience sink function for writing to a FILE*.
chris@16: 
chris@16:    This function can be used as a SerdSink when writing to a FILE*.  The
chris@16:    @c stream parameter must be a FILE* opened for writing.
chris@16: */
chris@16: SERD_API
chris@16: size_t
chris@16: serd_file_sink(const void* buf, size_t len, void* stream);
chris@16: 
chris@16: /**
chris@16:    A convenience sink function for writing to a string.
chris@16: 
chris@16:    This function can be used as a SerdSink to write to a SerdChunk which is
chris@16:    resized as necessary with realloc().  The @c stream parameter must point to
chris@16:    an initialized SerdChunk.  When the write is finished, the string should be
chris@16:    retrieved with serd_chunk_sink_finish().
chris@16: */
chris@16: SERD_API
chris@16: size_t
chris@16: serd_chunk_sink(const void* buf, size_t len, void* stream);
chris@16: 
chris@16: /**
chris@16:    Finish a serialisation to a chunk with serd_chunk_sink().
chris@16: 
chris@16:    The returned string is the result of the serialisation, which is NULL
chris@16:    terminated (by this function) and owned by the caller.
chris@16: */
chris@16: SERD_API
chris@16: uint8_t*
chris@16: serd_chunk_sink_finish(SerdChunk* stream);
chris@16: 
chris@16: /**
chris@16:    Set a function to be called when errors occur during writing.
chris@16: 
chris@16:    The @p error_sink will be called with @p handle as its first argument.  If
chris@16:    no error function is set, errors are printed to stderr.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_writer_set_error_sink(SerdWriter*   writer,
chris@16:                            SerdErrorSink error_sink,
chris@16:                            void*         handle);
chris@16: 
chris@16: /**
chris@16:    Set a prefix to be removed from matching blank node identifiers.
chris@16: */
chris@16: SERD_API
chris@16: void
chris@16: serd_writer_chop_blank_prefix(SerdWriter*    writer,
chris@16:                               const uint8_t* prefix);
chris@16: 
chris@16: /**
chris@16:    Set the current output base URI (and emit directive if applicable).
chris@16: 
chris@16:    Note this function can be safely casted to SerdBaseSink.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_set_base_uri(SerdWriter*     writer,
chris@16:                          const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Set the current root URI.
chris@16: 
chris@16:    The root URI should be a prefix of the base URI.  The path of the root URI
chris@16:    is the highest path any relative up-reference can refer to.  For example,
chris@16:    with root <file:///foo/root> and base <file:///foo/root/base>,
chris@16:    <file:///foo/root> will be written as <../>, but <file:///foo> will be
chris@16:    written non-relatively as <file:///foo>.  If the root is not explicitly set,
chris@16:    it defaults to the base URI, so no up-references will be created at all.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_set_root_uri(SerdWriter*     writer,
chris@16:                          const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Set a namespace prefix (and emit directive if applicable).
chris@16: 
chris@16:    Note this function can be safely casted to SerdPrefixSink.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_set_prefix(SerdWriter*     writer,
chris@16:                        const SerdNode* name,
chris@16:                        const SerdNode* uri);
chris@16: 
chris@16: /**
chris@16:    Write a statement.
chris@16: 
chris@16:    Note this function can be safely casted to SerdStatementSink.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_write_statement(SerdWriter*        writer,
chris@16:                             SerdStatementFlags flags,
chris@16:                             const SerdNode*    graph,
chris@16:                             const SerdNode*    subject,
chris@16:                             const SerdNode*    predicate,
chris@16:                             const SerdNode*    object,
chris@16:                             const SerdNode*    object_datatype,
chris@16:                             const SerdNode*    object_lang);
chris@16: 
chris@16: /**
chris@16:    Mark the end of an anonymous node's description.
chris@16: 
chris@16:    Note this function can be safely casted to SerdEndSink.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_end_anon(SerdWriter*     writer,
chris@16:                      const SerdNode* node);
chris@16: 
chris@16: /**
chris@16:    Finish a write.
chris@16: */
chris@16: SERD_API
chris@16: SerdStatus
chris@16: serd_writer_finish(SerdWriter* writer);
chris@16: 
chris@16: /**
chris@16:    @}
chris@16:    @}
chris@16: */
chris@16: 
chris@16: #ifdef __cplusplus
chris@16: }  /* extern "C" */
chris@16: #endif
chris@16: 
chris@16: #endif  /* SERD_SERD_H */