/src/tinysparql/src/libtinysparql/tracker-serializer-json-ld.c
Line | Count | Source (jump to first uncovered line) |
1 | | /* |
2 | | * Copyright (C) 2020, Red Hat, Inc |
3 | | * |
4 | | * This library is free software; you can redistribute it and/or |
5 | | * modify it under the terms of the GNU Lesser General Public |
6 | | * License as published by the Free Software Foundation; either |
7 | | * version 2.1 of the License, or (at your option) any later version. |
8 | | * |
9 | | * This library is distributed in the hope that it will be useful, |
10 | | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
11 | | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
12 | | * Lesser General Public License for more details. |
13 | | * |
14 | | * You should have received a copy of the GNU Lesser General Public |
15 | | * License along with this library; if not, write to the |
16 | | * Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
17 | | * Boston, MA 02110-1301, USA. |
18 | | * |
19 | | * Author: Carlos Garnacho <carlosg@gnome.org> |
20 | | */ |
21 | | |
22 | | /* Serialization of cursors to the JSON-LD format defined at: |
23 | | * https://www.w3.org/TR/json-ld/ |
24 | | */ |
25 | | |
26 | | #include "config.h" |
27 | | |
28 | | #include "tracker-serializer-json-ld.h" |
29 | | |
30 | | #include <json-glib/json-glib.h> |
31 | | |
32 | | struct _TrackerSerializerJsonLD |
33 | | { |
34 | | TrackerSerializer parent_instance; |
35 | | JsonGenerator *generator; |
36 | | JsonBuilder *builder; |
37 | | GHashTable *resources; |
38 | | GList *recent_resources; |
39 | | GString *data; |
40 | | GPtrArray *vars; |
41 | | gchar *cur_graph; |
42 | | gchar *cur_subject; |
43 | | JsonObject *cur_resource; |
44 | | guint stream_closed : 1; |
45 | | guint cursor_started : 1; |
46 | | guint cursor_finished : 1; |
47 | | guint context_printed : 1; |
48 | | guint needs_separator : 1; |
49 | | }; |
50 | | |
51 | | G_DEFINE_TYPE (TrackerSerializerJsonLD, tracker_serializer_json_ld, |
52 | | TRACKER_TYPE_SERIALIZER) |
53 | | |
54 | 0 | #define MAX_CACHED_RESOURCES 50 |
55 | | |
56 | | static void |
57 | | tracker_serializer_json_ld_finalize (GObject *object) |
58 | 0 | { |
59 | 0 | TrackerSerializerJsonLD *serializer_json_ld = |
60 | 0 | TRACKER_SERIALIZER_JSON_LD (object); |
61 | |
|
62 | 0 | g_input_stream_close (G_INPUT_STREAM (object), NULL, NULL); |
63 | 0 | g_clear_pointer (&serializer_json_ld->cur_graph, g_free); |
64 | 0 | g_clear_pointer (&serializer_json_ld->cur_subject, g_free); |
65 | |
|
66 | 0 | G_OBJECT_CLASS (tracker_serializer_json_ld_parent_class)->finalize (object); |
67 | 0 | } |
68 | | |
69 | | static void |
70 | | generate_namespace_foreach (gpointer key, |
71 | | gpointer value, |
72 | | gpointer user_data) |
73 | 0 | { |
74 | 0 | JsonBuilder *builder = user_data;; |
75 | |
|
76 | 0 | json_builder_set_member_name (builder, key); |
77 | 0 | json_builder_add_string_value (builder, value); |
78 | 0 | } |
79 | | |
80 | | static void |
81 | | finish_objects (TrackerSerializerJsonLD *serializer_json_ld) |
82 | 0 | { |
83 | 0 | GHashTableIter iter; |
84 | 0 | JsonNode *node; |
85 | |
|
86 | 0 | g_hash_table_iter_init (&iter, serializer_json_ld->resources); |
87 | |
|
88 | 0 | while (g_hash_table_iter_next (&iter, NULL, (gpointer*) &node)) { |
89 | 0 | if (g_list_find (serializer_json_ld->recent_resources, node)) { |
90 | 0 | if (serializer_json_ld->needs_separator) |
91 | 0 | g_string_append (serializer_json_ld->data, ",\n"); |
92 | |
|
93 | 0 | json_generator_set_root (serializer_json_ld->generator, node); |
94 | 0 | json_generator_to_gstring (serializer_json_ld->generator, |
95 | 0 | serializer_json_ld->data); |
96 | 0 | serializer_json_ld->needs_separator = TRUE; |
97 | 0 | } |
98 | |
|
99 | 0 | g_hash_table_iter_remove (&iter); |
100 | 0 | } |
101 | |
|
102 | 0 | g_clear_list (&serializer_json_ld->recent_resources, NULL); |
103 | 0 | } |
104 | | |
105 | | static JsonNode * |
106 | | create_node (const gchar *id, |
107 | | TrackerSparqlValueType type, |
108 | | TrackerNamespaceManager *namespaces) |
109 | 0 | { |
110 | 0 | JsonNode *node; |
111 | 0 | JsonObject *object; |
112 | |
|
113 | 0 | node = json_node_new (JSON_NODE_OBJECT); |
114 | 0 | object = json_object_new (); |
115 | |
|
116 | 0 | if (type == TRACKER_SPARQL_VALUE_TYPE_BLANK_NODE) { |
117 | 0 | gchar *bnode_label; |
118 | |
|
119 | 0 | bnode_label = g_strconcat ("_:", id, NULL); |
120 | 0 | g_strdelimit (&bnode_label[2], ":", '_'); |
121 | 0 | json_object_set_string_member (object, "@id", bnode_label); |
122 | 0 | g_free (bnode_label); |
123 | 0 | } else { |
124 | 0 | gchar *compressed; |
125 | |
|
126 | 0 | compressed = tracker_namespace_manager_compress_uri (namespaces, id); |
127 | |
|
128 | 0 | if (compressed) { |
129 | 0 | json_object_set_string_member (object, "@id", compressed); |
130 | 0 | g_free (compressed); |
131 | 0 | } else { |
132 | 0 | json_object_set_string_member (object, "@id", id); |
133 | 0 | } |
134 | 0 | } |
135 | |
|
136 | 0 | json_node_set_object (node, object); |
137 | 0 | json_object_unref (object); |
138 | |
|
139 | 0 | return node; |
140 | 0 | } |
141 | | |
142 | | static JsonNode * |
143 | | create_value_object (const gchar *value, |
144 | | const gchar *langtag, |
145 | | const gchar *datatype) |
146 | 0 | { |
147 | 0 | JsonNode *node; |
148 | 0 | JsonObject *object; |
149 | |
|
150 | 0 | node = json_node_new (JSON_NODE_OBJECT); |
151 | 0 | object = json_object_new (); |
152 | 0 | json_object_set_string_member (object, "@value", value); |
153 | |
|
154 | 0 | if (langtag) |
155 | 0 | json_object_set_string_member (object, "@language", langtag); |
156 | 0 | if (datatype) |
157 | 0 | json_object_set_string_member (object, "@type", datatype); |
158 | |
|
159 | 0 | json_node_set_object (node, object); |
160 | 0 | json_object_unref (object); |
161 | |
|
162 | 0 | return node; |
163 | 0 | } |
164 | | |
165 | | static gboolean |
166 | | serialize_up_to_position (TrackerSerializerJsonLD *serializer_json_ld, |
167 | | gsize pos, |
168 | | GCancellable *cancellable, |
169 | | GError **error) |
170 | 0 | { |
171 | 0 | TrackerSparqlCursor *cursor; |
172 | 0 | TrackerNamespaceManager *namespaces; |
173 | 0 | GError *inner_error = NULL; |
174 | 0 | JsonNode *node; |
175 | 0 | gboolean check_finish = FALSE; |
176 | |
|
177 | 0 | if (!serializer_json_ld->data) |
178 | 0 | serializer_json_ld->data = g_string_new (NULL); |
179 | 0 | if (!serializer_json_ld->generator) |
180 | 0 | serializer_json_ld->generator = json_generator_new (); |
181 | 0 | if (!serializer_json_ld->builder) |
182 | 0 | serializer_json_ld->builder = json_builder_new (); |
183 | 0 | if (!serializer_json_ld->vars) |
184 | 0 | serializer_json_ld->vars = g_ptr_array_new_with_free_func (g_free); |
185 | |
|
186 | 0 | if (!serializer_json_ld->resources) { |
187 | 0 | serializer_json_ld->resources = g_hash_table_new_full (g_str_hash, g_str_equal, |
188 | 0 | g_free, (GDestroyNotify) json_node_unref); |
189 | 0 | } |
190 | |
|
191 | 0 | if (pos < serializer_json_ld->data->len) |
192 | 0 | return TRUE; |
193 | | |
194 | 0 | cursor = tracker_serializer_get_cursor (TRACKER_SERIALIZER (serializer_json_ld)); |
195 | 0 | namespaces = tracker_serializer_get_namespaces (TRACKER_SERIALIZER (serializer_json_ld)); |
196 | |
|
197 | 0 | if (!serializer_json_ld->context_printed) { |
198 | 0 | JsonBuilder *builder; |
199 | |
|
200 | 0 | g_string_append (serializer_json_ld->data, "{"); |
201 | |
|
202 | 0 | builder = json_builder_new (); |
203 | 0 | json_builder_begin_object (builder); |
204 | 0 | tracker_namespace_manager_foreach (namespaces, |
205 | 0 | generate_namespace_foreach, |
206 | 0 | builder); |
207 | 0 | json_builder_end_object (builder); |
208 | |
|
209 | 0 | node = json_builder_get_root (builder); |
210 | 0 | json_generator_set_root (serializer_json_ld->generator, node); |
211 | 0 | g_object_unref (builder); |
212 | 0 | json_node_unref (node); |
213 | |
|
214 | 0 | g_string_append (serializer_json_ld->data, "\"@context\":"); |
215 | 0 | json_generator_to_gstring (serializer_json_ld->generator, |
216 | 0 | serializer_json_ld->data); |
217 | |
|
218 | 0 | g_string_append (serializer_json_ld->data, |
219 | 0 | ",\"@graph\":["); |
220 | |
|
221 | 0 | serializer_json_ld->context_printed = TRUE; |
222 | 0 | } |
223 | |
|
224 | 0 | while (!serializer_json_ld->cursor_finished) { |
225 | 0 | const gchar *graph, *subject, *predicate, *langtag; |
226 | 0 | gboolean graph_changed, subject_changed; |
227 | 0 | TrackerSparqlValueType subject_type, object_type; |
228 | 0 | JsonNode *value = NULL; |
229 | 0 | gchar *prop = NULL; |
230 | |
|
231 | 0 | if (!tracker_sparql_cursor_next (cursor, cancellable, &inner_error)) { |
232 | 0 | if (inner_error) { |
233 | 0 | g_propagate_error (error, inner_error); |
234 | 0 | return FALSE; |
235 | 0 | } else { |
236 | 0 | finish_objects (serializer_json_ld); |
237 | 0 | serializer_json_ld->cursor_finished = TRUE; |
238 | 0 | if (serializer_json_ld->cur_graph) |
239 | 0 | g_string_append (serializer_json_ld->data, "]}"); |
240 | |
|
241 | 0 | g_string_append (serializer_json_ld->data, "]}"); |
242 | |
|
243 | 0 | return TRUE; |
244 | 0 | } |
245 | 0 | } |
246 | | |
247 | 0 | subject = tracker_sparql_cursor_get_string (cursor, 0, NULL); |
248 | 0 | subject_type = tracker_sparql_cursor_get_value_type (cursor, 0); |
249 | 0 | predicate = tracker_sparql_cursor_get_string (cursor, 1, NULL); |
250 | 0 | object_type = tracker_sparql_cursor_get_value_type (cursor, 2); |
251 | 0 | graph = tracker_sparql_cursor_get_string (cursor, 3, NULL); |
252 | |
|
253 | 0 | graph_changed = g_strcmp0 (graph, serializer_json_ld->cur_graph) != 0; |
254 | 0 | subject_changed = g_strcmp0 (subject, serializer_json_ld->cur_subject) != 0; |
255 | |
|
256 | 0 | if (graph_changed) { |
257 | | /* New/different graph */ |
258 | 0 | if (serializer_json_ld->cursor_started && graph_changed) { |
259 | 0 | finish_objects (serializer_json_ld); |
260 | 0 | if (serializer_json_ld->cur_graph) |
261 | 0 | g_string_append (serializer_json_ld->data, "]}"); |
262 | |
|
263 | 0 | g_string_append (serializer_json_ld->data, ",\n"); |
264 | 0 | } |
265 | |
|
266 | 0 | if (graph) { |
267 | 0 | g_string_append_printf (serializer_json_ld->data, |
268 | 0 | "{\"@id\":\"%s\",\"@graph\":[", |
269 | 0 | graph); |
270 | 0 | } |
271 | |
|
272 | 0 | g_clear_pointer (&serializer_json_ld->cur_graph, g_free); |
273 | 0 | serializer_json_ld->cur_graph = g_strdup (graph); |
274 | 0 | serializer_json_ld->needs_separator = FALSE; |
275 | 0 | } |
276 | |
|
277 | 0 | if (subject_changed || graph_changed) { |
278 | 0 | JsonObject *object; |
279 | | |
280 | | /* New/different subject */ |
281 | 0 | if (serializer_json_ld->cursor_started && subject_changed) |
282 | 0 | check_finish = TRUE; |
283 | |
|
284 | 0 | if (g_hash_table_size (serializer_json_ld->resources) > MAX_CACHED_RESOURCES) |
285 | 0 | finish_objects (serializer_json_ld); |
286 | |
|
287 | 0 | node = g_hash_table_lookup (serializer_json_ld->resources, subject); |
288 | |
|
289 | 0 | if (!node) { |
290 | 0 | node = create_node (subject, subject_type, namespaces); |
291 | |
|
292 | 0 | g_hash_table_insert (serializer_json_ld->resources, g_strdup (subject), node); |
293 | 0 | serializer_json_ld->recent_resources = |
294 | 0 | g_list_prepend (serializer_json_ld->recent_resources, node); |
295 | 0 | } |
296 | |
|
297 | 0 | object = json_node_get_object (node); |
298 | 0 | serializer_json_ld->cur_resource = object; |
299 | |
|
300 | 0 | g_clear_pointer (&serializer_json_ld->cur_subject, g_free); |
301 | 0 | serializer_json_ld->cur_subject = g_strdup (subject); |
302 | 0 | } |
303 | |
|
304 | 0 | if (g_strcmp0 (predicate, TRACKER_PREFIX_RDF "type") == 0) { |
305 | 0 | const gchar *type; |
306 | 0 | gchar *compressed; |
307 | |
|
308 | 0 | type = tracker_sparql_cursor_get_string (cursor, 2, NULL); |
309 | 0 | compressed = tracker_namespace_manager_compress_uri (namespaces, type); |
310 | |
|
311 | 0 | prop = g_strdup ("@type"); |
312 | |
|
313 | 0 | value = json_node_alloc (); |
314 | 0 | json_node_init_string (value, compressed); |
315 | 0 | g_free (compressed); |
316 | 0 | } else { |
317 | 0 | TrackerNamespaceManager *namespaces; |
318 | 0 | const gchar *res; |
319 | |
|
320 | 0 | namespaces = tracker_serializer_get_namespaces (TRACKER_SERIALIZER (serializer_json_ld)); |
321 | 0 | prop = tracker_namespace_manager_compress_uri (namespaces, predicate); |
322 | |
|
323 | 0 | switch (object_type) { |
324 | 0 | case TRACKER_SPARQL_VALUE_TYPE_URI: |
325 | 0 | case TRACKER_SPARQL_VALUE_TYPE_BLANK_NODE: |
326 | 0 | res = tracker_sparql_cursor_get_string (cursor, 2, NULL); |
327 | |
|
328 | 0 | node = g_hash_table_lookup (serializer_json_ld->resources, |
329 | 0 | res); |
330 | |
|
331 | 0 | if (node && |
332 | 0 | serializer_json_ld->recent_resources->next != NULL && |
333 | 0 | serializer_json_ld->cur_resource != json_node_get_object (node) && |
334 | 0 | g_list_find (serializer_json_ld->recent_resources, node)) { |
335 | | /* This is still a "root" node, make it part of this tree */ |
336 | 0 | serializer_json_ld->recent_resources = |
337 | 0 | g_list_remove (serializer_json_ld->recent_resources, node); |
338 | 0 | value = json_node_ref (node); |
339 | 0 | } else { |
340 | | /* Unknown object, or one already referenced elsewhere */ |
341 | 0 | value = create_node (res, object_type, namespaces); |
342 | 0 | } |
343 | 0 | break; |
344 | 0 | case TRACKER_SPARQL_VALUE_TYPE_DATETIME: |
345 | 0 | res = tracker_sparql_cursor_get_string (cursor, 2, NULL); |
346 | 0 | value = create_value_object (res, NULL, TRACKER_PREFIX_XSD "dateTime"); |
347 | 0 | break; |
348 | 0 | case TRACKER_SPARQL_VALUE_TYPE_STRING: |
349 | 0 | res = tracker_sparql_cursor_get_langstring (cursor, 2, &langtag, NULL); |
350 | 0 | if (langtag) { |
351 | 0 | value = create_value_object (res, langtag, TRACKER_PREFIX_RDF "langString"); |
352 | 0 | } else { |
353 | 0 | value = json_node_alloc (); |
354 | 0 | json_node_init_string (value, res); |
355 | 0 | } |
356 | 0 | break; |
357 | 0 | case TRACKER_SPARQL_VALUE_TYPE_INTEGER: |
358 | 0 | value = json_node_alloc (); |
359 | 0 | json_node_init_int (value, tracker_sparql_cursor_get_integer (cursor, 2)); |
360 | 0 | break; |
361 | 0 | case TRACKER_SPARQL_VALUE_TYPE_DOUBLE: |
362 | 0 | value = json_node_alloc (); |
363 | 0 | json_node_init_double (value, tracker_sparql_cursor_get_double (cursor, 2)); |
364 | 0 | break; |
365 | 0 | case TRACKER_SPARQL_VALUE_TYPE_BOOLEAN: |
366 | 0 | value = json_node_alloc (); |
367 | 0 | json_node_init_boolean (value, tracker_sparql_cursor_get_boolean (cursor, 2)); |
368 | 0 | break; |
369 | 0 | case TRACKER_SPARQL_VALUE_TYPE_UNBOUND: |
370 | 0 | break; |
371 | 0 | } |
372 | 0 | } |
373 | | |
374 | 0 | if (prop && value) { |
375 | 0 | JsonNode *prev; |
376 | 0 | JsonArray *array; |
377 | |
|
378 | 0 | prev = json_object_get_member (serializer_json_ld->cur_resource, |
379 | 0 | prop); |
380 | |
|
381 | 0 | if (!prev) { |
382 | 0 | json_object_set_member (serializer_json_ld->cur_resource, |
383 | 0 | prop, value); |
384 | 0 | } else if (JSON_NODE_HOLDS_ARRAY (prev)) { |
385 | 0 | array = json_node_get_array (prev); |
386 | 0 | json_array_add_element (array, value); |
387 | 0 | } else if (!json_node_equal (prev, value)) { |
388 | 0 | array = json_array_new (); |
389 | 0 | json_array_add_element (array, json_node_ref (prev)); |
390 | 0 | json_array_add_element (array, value); |
391 | |
|
392 | 0 | json_object_set_array_member (serializer_json_ld->cur_resource, |
393 | 0 | prop, array); |
394 | 0 | } |
395 | 0 | } |
396 | |
|
397 | 0 | g_free (prop); |
398 | 0 | serializer_json_ld->cursor_started = TRUE; |
399 | |
|
400 | 0 | if (check_finish && serializer_json_ld->data->len > pos) |
401 | 0 | break; |
402 | 0 | } |
403 | | |
404 | 0 | return TRUE; |
405 | 0 | } |
406 | | |
407 | | static gssize |
408 | | tracker_serializer_json_ld_read (GInputStream *istream, |
409 | | gpointer buffer, |
410 | | gsize count, |
411 | | GCancellable *cancellable, |
412 | | GError **error) |
413 | 0 | { |
414 | 0 | TrackerSerializerJsonLD *serializer_json_ld = TRACKER_SERIALIZER_JSON_LD (istream); |
415 | 0 | gsize bytes_copied; |
416 | |
|
417 | 0 | if (serializer_json_ld->stream_closed || |
418 | 0 | (serializer_json_ld->cursor_finished && |
419 | 0 | serializer_json_ld->data->len == 0)) |
420 | 0 | return 0; |
421 | | |
422 | 0 | if (!serialize_up_to_position (serializer_json_ld, |
423 | 0 | count, |
424 | 0 | cancellable, |
425 | 0 | error)) |
426 | 0 | return -1; |
427 | | |
428 | 0 | bytes_copied = MIN (count, serializer_json_ld->data->len); |
429 | |
|
430 | 0 | memcpy (buffer, |
431 | 0 | serializer_json_ld->data->str, |
432 | 0 | bytes_copied); |
433 | 0 | g_string_erase (serializer_json_ld->data, 0, bytes_copied); |
434 | |
|
435 | 0 | return bytes_copied; |
436 | 0 | } |
437 | | |
438 | | static gboolean |
439 | | tracker_serializer_json_ld_close (GInputStream *istream, |
440 | | GCancellable *cancellable, |
441 | | GError **error) |
442 | 0 | { |
443 | 0 | TrackerSerializerJsonLD *serializer_json_ld = TRACKER_SERIALIZER_JSON_LD (istream); |
444 | |
|
445 | 0 | if (serializer_json_ld->data) { |
446 | 0 | g_string_free (serializer_json_ld->data, TRUE); |
447 | 0 | serializer_json_ld->data = NULL; |
448 | 0 | } |
449 | |
|
450 | 0 | g_clear_object (&serializer_json_ld->generator); |
451 | 0 | g_clear_object (&serializer_json_ld->builder); |
452 | 0 | serializer_json_ld->stream_closed = TRUE; |
453 | 0 | g_clear_pointer (&serializer_json_ld->vars, g_ptr_array_unref); |
454 | |
|
455 | 0 | return TRUE; |
456 | 0 | } |
457 | | |
458 | | static void |
459 | | tracker_serializer_json_ld_class_init (TrackerSerializerJsonLDClass *klass) |
460 | 0 | { |
461 | 0 | GObjectClass *object_class = G_OBJECT_CLASS (klass); |
462 | 0 | GInputStreamClass *istream_class = G_INPUT_STREAM_CLASS (klass); |
463 | |
|
464 | 0 | object_class->finalize = tracker_serializer_json_ld_finalize; |
465 | |
|
466 | 0 | istream_class->read_fn = tracker_serializer_json_ld_read; |
467 | 0 | istream_class->close_fn = tracker_serializer_json_ld_close; |
468 | 0 | } |
469 | | |
470 | | static void |
471 | | tracker_serializer_json_ld_init (TrackerSerializerJsonLD *serializer) |
472 | 0 | { |
473 | 0 | } |