Add support for parsing variables

This adds a reader flag and serdi option for extending a syntax with support
for SPARQL-like variables, for storing things like patterns or simple queries.

7 files changed, 90 insertions, 7 deletions

diff --git a/src/env.c b/src/env.c
index 5cbb6a43..3abbc857 100644
--- a/src/env.c
+++ b/src/env.c
@@ -371,6 +371,7 @@ serd_env_expand(const SerdEnv* env, const SerdNode* node)
 		case SERD_CURIE:
 			return expand_curie(env, node);
 		case SERD_BLANK:
+		case SERD_VARIABLE:
 			return serd_node_copy(node);
 		}
 	}
diff --git a/src/n3.c b/src/n3.c
index 4fe0363c..39cd515f 100644
--- a/src/n3.c
+++ b/src/n3.c
@@ -866,10 +866,50 @@ read_literal(SerdReader* reader, SerdNode** dest, bool* ate_dot)
 }
 
 static SerdStatus
+read_VARNAME(SerdReader* reader, SerdNode** dest)
+{
+	// Simplified from SPARQL: VARNAME ::= (PN_CHARS_U | [0-9])+
+	SerdNode*  n  = *dest;
+	SerdStatus st = SERD_SUCCESS;
+	int        c  = 0;
+	peek_byte(reader);
+	while ((c = peek_byte(reader))) {
+		if (is_digit(c) || c == '_') {
+			push_byte(reader, n, eat_byte_safe(reader, c));
+		} else if ((st = read_PN_CHARS(reader, n))) {
+			st = st > SERD_FAILURE ? st : SERD_SUCCESS;
+			break;
+		}
+	}
+
+	return st;
+}
+
+static SerdStatus
+read_Var(SerdReader* reader, SerdNode** dest)
+{
+	if (!(reader->flags & SERD_READ_VARIABLES)) {
+		return r_err(reader, SERD_ERR_BAD_SYNTAX,
+		             "syntax does not support variables\n");
+	} else if (!(*dest = push_node(reader, SERD_VARIABLE, "", 0))) {
+		return SERD_ERR_OVERFLOW;
+	}
+
+	assert(peek_byte(reader) == '$' || peek_byte(reader) == '?');
+	serd_byte_source_advance(&reader->source);
+
+	return read_VARNAME(reader, dest);
+}
+
+static SerdStatus
 read_verb(SerdReader* reader, SerdNode** dest)
 {
 	const size_t orig_stack_size = reader->stack.size;
-	if (peek_byte(reader) == '<') {
+
+	switch (peek_byte(reader)) {
+	case '$': case '?':
+		return read_Var(reader, dest);
+	case '<':
 		return read_IRIREF(reader, dest);
 	}
 
@@ -1030,7 +1070,13 @@ read_object(SerdReader* reader, ReadContext* ctx, bool emit, bool* ate_dot)
 	const int  c      = peek_byte(reader);
 	if (!fancy_syntax(reader)) {
 		switch (c) {
-		case '"': case ':': case '<': case '_': break;
+		case '"': case ':': case '<': case '_':
+			break;
+		case '$': case '?':
+			if (reader->flags & SERD_READ_VARIABLES) {
+				break;
+			}
+			break;
 		default:
 			return r_err(reader, SERD_ERR_BAD_SYNTAX,
 			             "expected: ':', '<', or '_'\n");
@@ -1040,6 +1086,9 @@ read_object(SerdReader* reader, ReadContext* ctx, bool emit, bool* ate_dot)
 	switch (c) {
 	case EOF: case ')':
 		return r_err(reader, SERD_ERR_BAD_SYNTAX, "expected object\n");
+	case '$': case '?':
+		ret = read_Var(reader, &o);
+		break;
 	case '[':
 		simple = false;
 		ret = read_anon(reader, *ctx, false, &o);
@@ -1241,6 +1290,9 @@ read_subject(SerdReader* reader, ReadContext ctx, SerdNode** dest, int* s_type)
 	SerdStatus st      = SERD_SUCCESS;
 	bool       ate_dot = false;
 	switch ((*s_type = peek_byte(reader))) {
+	case '$': case '?':
+		st = read_Var(reader, dest);
+		break;
 	case '[':
 		read_anon(reader, ctx, true, dest);
 		break;
@@ -1558,11 +1610,21 @@ read_nquadsDoc(SerdReader* reader)
 			return SERD_ERR_BAD_SYNTAX;
 		}
 
-		// subject predicate object
 		if ((st = read_subject(reader, ctx, &ctx.subject, &s_type)) ||
-		    !read_ws_star(reader) ||
-		    (st = read_IRIREF(reader, &ctx.predicate)) ||
-		    !read_ws_star(reader) ||
+		    !read_ws_star(reader)) {
+			return st;
+		}
+
+		switch (peek_byte(reader)) {
+		case '$': case '?':
+			st = read_Var(reader, &ctx.predicate);
+			break;
+		case '<':
+			st = read_IRIREF(reader, &ctx.predicate);
+			break;
+		}
+
+		if (st || !read_ws_star(reader) ||
 		    (st = read_object(reader, &ctx, false, &ate_dot))) {
 			return st;
 		}
diff --git a/src/node.c b/src/node.c
index b33a03e9..66964556 100644
--- a/src/node.c
+++ b/src/node.c
@@ -156,7 +156,8 @@ serd_new_simple_node(SerdNodeType type, const char* str, const size_t len)
 {
 	if (!str) {
 		return NULL;
-	} else if (type != SERD_BLANK && type != SERD_CURIE && type != SERD_URI) {
+	} else if (type != SERD_BLANK && type != SERD_CURIE && type != SERD_URI &&
+	           type != SERD_VARIABLE) {
 		return NULL;
 	}
 
diff --git a/src/reader.c b/src/reader.c
index 668cd38d..3dd2f465 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -172,6 +172,7 @@ serd_reader_new(SerdWorld*      world,
 	me->sink          = sink;
 	me->stack         = serd_stack_new(stack_size, serd_node_align);
 	me->syntax        = syntax;
+	me->flags         = flags;
 	me->next_id       = 1;
 	me->strict        = !(flags & SERD_READ_LAX);
 
diff --git a/src/reader.h b/src/reader.h
index 21dc2e04..a5e1f3f0 100644
--- a/src/reader.h
+++ b/src/reader.h
@@ -60,6 +60,7 @@ struct SerdReaderImpl {
 	SerdByteSource    source;
 	SerdStack         stack;
 	SerdSyntax        syntax;
+	SerdReaderFlags   flags;
 	unsigned          next_id;
 	uint8_t*          buf;
 	char*             bprefix;
diff --git a/src/serdi.c b/src/serdi.c
index 7f4880fd..1339e593 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -72,6 +72,7 @@ print_usage(const char* name, bool error)
 	fprintf(os, "  -s INPUT     Parse INPUT as string.\n");
 	fprintf(os, "  -t           Write terser output without newlines.\n");
 	fprintf(os, "  -v           Display version information and exit.\n");
+	fprintf(os, "  -x           Support parsing variable nodes like `?x'.\n");
 	return error ? 1 : 0;
 }
 
@@ -228,6 +229,8 @@ main(int argc, char** argv)
 				return missing_arg(argv[0], 'r');
 			}
 			root_uri = argv[a];
+		} else if (argv[a][1] == 'x') {
+			reader_flags |= SERD_READ_VARIABLES;
 		} else {
 			SERDI_ERRORF("invalid option -- '%s'\n", argv[a] + 1);
 			return print_usage(argv[0], true);
diff --git a/src/writer.c b/src/writer.c
index acb63d7d..d3326647 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -773,6 +773,18 @@ write_blank(SerdWriter* const        writer,
 }
 
 SERD_WARN_UNUSED_RESULT static SerdStatus
+write_variable(SerdWriter* const writer, const SerdNode* node)
+{
+	SerdStatus st = SERD_SUCCESS;
+
+	TRY(st, esink("?", 1, writer));
+	TRY(st, esink(serd_node_string(node), node->n_bytes, writer));
+
+	writer->last_sep = SEP_NONE;
+	return st;
+}
+
+SERD_WARN_UNUSED_RESULT static SerdStatus
 write_node(SerdWriter*        writer,
            const SerdNode*    node,
            const SerdField    field,
@@ -787,6 +799,8 @@ write_node(SerdWriter*        writer,
 		return write_curie(writer, node, field, flags);
 	case SERD_BLANK:
 		return write_blank(writer, node, field, flags);
+	case SERD_VARIABLE:
+		return write_variable(writer, node);
 	}
 
 	return SERD_ERR_INTERNAL;