1 /*
2 * This file is part of the MicroPython project, http://micropython.org/
3 *
4 * The MIT License (MIT)
5 *
6 * Copyright (c) 2014-2019 Damien P. George
7 *
8 * Permission is hereby granted, free of charge, to any person obtaining a copy
9 * of this software and associated documentation files (the "Software"), to deal
10 * in the Software without restriction, including without limitation the rights
11 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
12 * copies of the Software, and to permit persons to whom the Software is
13 * furnished to do so, subject to the following conditions:
14 *
15 * The above copyright notice and this permission notice shall be included in
16 * all copies or substantial portions of the Software.
17 *
18 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
19 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
20 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
21 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
22 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
23 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
24 * THE SOFTWARE.
25 */
26
27 #include <stdio.h>
28
29 #include "py/objlist.h"
30 #include "py/objstringio.h"
31 #include "py/parsenum.h"
32 #include "py/runtime.h"
33 #include "py/stream.h"
34
35 #if MICROPY_PY_UJSON
36
37 #if MICROPY_PY_UJSON_SEPARATORS
38
39 enum {
40 DUMP_MODE_TO_STRING = 1,
41 DUMP_MODE_TO_STREAM = 2,
42 };
43
mod_ujson_dump_helper(size_t n_args,const mp_obj_t * pos_args,mp_map_t * kw_args,unsigned int mode)44 STATIC mp_obj_t mod_ujson_dump_helper(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args, unsigned int mode) {
45 enum { ARG_separators };
46 static const mp_arg_t allowed_args[] = {
47 { MP_QSTR_separators, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_NONE} },
48 };
49
50 mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
51 mp_arg_parse_all(n_args - mode, pos_args + mode, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
52
53 mp_print_ext_t print_ext;
54
55 if (args[ARG_separators].u_obj == mp_const_none) {
56 print_ext.item_separator = ", ";
57 print_ext.key_separator = ": ";
58 } else {
59 mp_obj_t *items;
60 mp_obj_get_array_fixed_n(args[ARG_separators].u_obj, 2, &items);
61 print_ext.item_separator = mp_obj_str_get_str(items[0]);
62 print_ext.key_separator = mp_obj_str_get_str(items[1]);
63 }
64
65 if (mode == DUMP_MODE_TO_STRING) {
66 // dumps(obj)
67 vstr_t vstr;
68 vstr_init_print(&vstr, 8, &print_ext.base);
69 mp_obj_print_helper(&print_ext.base, pos_args[0], PRINT_JSON);
70 return mp_obj_new_str_from_vstr(&mp_type_str, &vstr);
71 } else {
72 // dump(obj, stream)
73 print_ext.base.data = MP_OBJ_TO_PTR(pos_args[1]);
74 print_ext.base.print_strn = mp_stream_write_adaptor;
75 mp_get_stream_raise(pos_args[1], MP_STREAM_OP_WRITE);
76 mp_obj_print_helper(&print_ext.base, pos_args[0], PRINT_JSON);
77 return mp_const_none;
78 }
79 }
80
mod_ujson_dump(size_t n_args,const mp_obj_t * pos_args,mp_map_t * kw_args)81 STATIC mp_obj_t mod_ujson_dump(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
82 return mod_ujson_dump_helper(n_args, pos_args, kw_args, DUMP_MODE_TO_STREAM);
83 }
84 STATIC MP_DEFINE_CONST_FUN_OBJ_KW(mod_ujson_dump_obj, 2, mod_ujson_dump);
85
mod_ujson_dumps(size_t n_args,const mp_obj_t * pos_args,mp_map_t * kw_args)86 STATIC mp_obj_t mod_ujson_dumps(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
87 return mod_ujson_dump_helper(n_args, pos_args, kw_args, DUMP_MODE_TO_STRING);
88 }
89 STATIC MP_DEFINE_CONST_FUN_OBJ_KW(mod_ujson_dumps_obj, 1, mod_ujson_dumps);
90
91 #else
92
mod_ujson_dump(mp_obj_t obj,mp_obj_t stream)93 STATIC mp_obj_t mod_ujson_dump(mp_obj_t obj, mp_obj_t stream) {
94 mp_get_stream_raise(stream, MP_STREAM_OP_WRITE);
95 mp_print_t print = {MP_OBJ_TO_PTR(stream), mp_stream_write_adaptor};
96 mp_obj_print_helper(&print, obj, PRINT_JSON);
97 return mp_const_none;
98 }
99 STATIC MP_DEFINE_CONST_FUN_OBJ_2(mod_ujson_dump_obj, mod_ujson_dump);
100
mod_ujson_dumps(mp_obj_t obj)101 STATIC mp_obj_t mod_ujson_dumps(mp_obj_t obj) {
102 vstr_t vstr;
103 mp_print_t print;
104 vstr_init_print(&vstr, 8, &print);
105 mp_obj_print_helper(&print, obj, PRINT_JSON);
106 return mp_obj_new_str_from_vstr(&mp_type_str, &vstr);
107 }
108 STATIC MP_DEFINE_CONST_FUN_OBJ_1(mod_ujson_dumps_obj, mod_ujson_dumps);
109
110 #endif
111
112 // The function below implements a simple non-recursive JSON parser.
113 //
114 // The JSON specification is at http://www.ietf.org/rfc/rfc4627.txt
115 // The parser here will parse any valid JSON and return the correct
116 // corresponding Python object. It allows through a superset of JSON, since
117 // it treats commas and colons as "whitespace", and doesn't care if
118 // brackets/braces are correctly paired. It will raise a ValueError if the
119 // input is outside it's specs.
120 //
121 // Most of the work is parsing the primitives (null, false, true, numbers,
122 // strings). It does 1 pass over the input stream. It tries to be fast and
123 // small in code size, while not using more RAM than necessary.
124
125 typedef struct _ujson_stream_t {
126 mp_obj_t stream_obj;
127 mp_uint_t (*read)(mp_obj_t obj, void *buf, mp_uint_t size, int *errcode);
128 int errcode;
129 byte cur;
130 } ujson_stream_t;
131
132 #define S_EOF (0) // null is not allowed in json stream so is ok as EOF marker
133 #define S_END(s) ((s).cur == S_EOF)
134 #define S_CUR(s) ((s).cur)
135 #define S_NEXT(s) (ujson_stream_next(&(s)))
136
ujson_stream_next(ujson_stream_t * s)137 STATIC byte ujson_stream_next(ujson_stream_t *s) {
138 mp_uint_t ret = s->read(s->stream_obj, &s->cur, 1, &s->errcode);
139 if (s->errcode != 0) {
140 mp_raise_OSError(s->errcode);
141 }
142 if (ret == 0) {
143 s->cur = S_EOF;
144 }
145 return s->cur;
146 }
147
mod_ujson_load(mp_obj_t stream_obj)148 STATIC mp_obj_t mod_ujson_load(mp_obj_t stream_obj) {
149 const mp_stream_p_t *stream_p = mp_get_stream_raise(stream_obj, MP_STREAM_OP_READ);
150 ujson_stream_t s = {stream_obj, stream_p->read, 0, 0};
151 vstr_t vstr;
152 vstr_init(&vstr, 8);
153 mp_obj_list_t stack; // we use a list as a simple stack for nested JSON
154 stack.len = 0;
155 stack.items = NULL;
156 mp_obj_t stack_top = MP_OBJ_NULL;
157 const mp_obj_type_t *stack_top_type = NULL;
158 mp_obj_t stack_key = MP_OBJ_NULL;
159 S_NEXT(s);
160 for (;;) {
161 cont:
162 if (S_END(s)) {
163 break;
164 }
165 mp_obj_t next = MP_OBJ_NULL;
166 bool enter = false;
167 byte cur = S_CUR(s);
168 S_NEXT(s);
169 switch (cur) {
170 case ',':
171 case ':':
172 case ' ':
173 case '\t':
174 case '\n':
175 case '\r':
176 goto cont;
177 case 'n':
178 if (S_CUR(s) == 'u' && S_NEXT(s) == 'l' && S_NEXT(s) == 'l') {
179 S_NEXT(s);
180 next = mp_const_none;
181 } else {
182 goto fail;
183 }
184 break;
185 case 'f':
186 if (S_CUR(s) == 'a' && S_NEXT(s) == 'l' && S_NEXT(s) == 's' && S_NEXT(s) == 'e') {
187 S_NEXT(s);
188 next = mp_const_false;
189 } else {
190 goto fail;
191 }
192 break;
193 case 't':
194 if (S_CUR(s) == 'r' && S_NEXT(s) == 'u' && S_NEXT(s) == 'e') {
195 S_NEXT(s);
196 next = mp_const_true;
197 } else {
198 goto fail;
199 }
200 break;
201 case '"':
202 vstr_reset(&vstr);
203 for (; !S_END(s) && S_CUR(s) != '"';) {
204 byte c = S_CUR(s);
205 if (c == '\\') {
206 c = S_NEXT(s);
207 switch (c) {
208 case 'b':
209 c = 0x08;
210 break;
211 case 'f':
212 c = 0x0c;
213 break;
214 case 'n':
215 c = 0x0a;
216 break;
217 case 'r':
218 c = 0x0d;
219 break;
220 case 't':
221 c = 0x09;
222 break;
223 case 'u': {
224 mp_uint_t num = 0;
225 for (int i = 0; i < 4; i++) {
226 c = (S_NEXT(s) | 0x20) - '0';
227 if (c > 9) {
228 c -= ('a' - ('9' + 1));
229 }
230 num = (num << 4) | c;
231 }
232 vstr_add_char(&vstr, num);
233 goto str_cont;
234 }
235 }
236 }
237 vstr_add_byte(&vstr, c);
238 str_cont:
239 S_NEXT(s);
240 }
241 if (S_END(s)) {
242 goto fail;
243 }
244 S_NEXT(s);
245 next = mp_obj_new_str(vstr.buf, vstr.len);
246 break;
247 case '-':
248 case '0':
249 case '1':
250 case '2':
251 case '3':
252 case '4':
253 case '5':
254 case '6':
255 case '7':
256 case '8':
257 case '9': {
258 bool flt = false;
259 vstr_reset(&vstr);
260 for (;;) {
261 vstr_add_byte(&vstr, cur);
262 cur = S_CUR(s);
263 if (cur == '.' || cur == 'E' || cur == 'e') {
264 flt = true;
265 } else if (cur == '+' || cur == '-' || unichar_isdigit(cur)) {
266 // pass
267 } else {
268 break;
269 }
270 S_NEXT(s);
271 }
272 if (flt) {
273 next = mp_parse_num_decimal(vstr.buf, vstr.len, false, false, NULL);
274 } else {
275 next = mp_parse_num_integer(vstr.buf, vstr.len, 10, NULL);
276 }
277 break;
278 }
279 case '[':
280 next = mp_obj_new_list(0, NULL);
281 enter = true;
282 break;
283 case '{':
284 next = mp_obj_new_dict(0);
285 enter = true;
286 break;
287 case '}':
288 case ']': {
289 if (stack_top == MP_OBJ_NULL) {
290 // no object at all
291 goto fail;
292 }
293 if (stack.len == 0) {
294 // finished; compound object
295 goto success;
296 }
297 stack.len -= 1;
298 stack_top = stack.items[stack.len];
299 stack_top_type = mp_obj_get_type(stack_top);
300 goto cont;
301 }
302 default:
303 goto fail;
304 }
305 if (stack_top == MP_OBJ_NULL) {
306 stack_top = next;
307 stack_top_type = mp_obj_get_type(stack_top);
308 if (!enter) {
309 // finished; single primitive only
310 goto success;
311 }
312 } else {
313 // append to list or dict
314 if (stack_top_type == &mp_type_list) {
315 mp_obj_list_append(stack_top, next);
316 } else {
317 if (stack_key == MP_OBJ_NULL) {
318 stack_key = next;
319 if (enter) {
320 goto fail;
321 }
322 } else {
323 mp_obj_dict_store(stack_top, stack_key, next);
324 stack_key = MP_OBJ_NULL;
325 }
326 }
327 if (enter) {
328 if (stack.items == NULL) {
329 mp_obj_list_init(&stack, 1);
330 stack.items[0] = stack_top;
331 } else {
332 mp_obj_list_append(MP_OBJ_FROM_PTR(&stack), stack_top);
333 }
334 stack_top = next;
335 stack_top_type = mp_obj_get_type(stack_top);
336 }
337 }
338 }
339 success:
340 // eat trailing whitespace
341 while (unichar_isspace(S_CUR(s))) {
342 S_NEXT(s);
343 }
344 if (!S_END(s)) {
345 // unexpected chars
346 goto fail;
347 }
348 if (stack_top == MP_OBJ_NULL || stack.len != 0) {
349 // not exactly 1 object
350 goto fail;
351 }
352 vstr_clear(&vstr);
353 return stack_top;
354
355 fail:
356 mp_raise_ValueError(MP_ERROR_TEXT("syntax error in JSON"));
357 }
358 STATIC MP_DEFINE_CONST_FUN_OBJ_1(mod_ujson_load_obj, mod_ujson_load);
359
mod_ujson_loads(mp_obj_t obj)360 STATIC mp_obj_t mod_ujson_loads(mp_obj_t obj) {
361 mp_buffer_info_t bufinfo;
362 mp_get_buffer_raise(obj, &bufinfo, MP_BUFFER_READ);
363 vstr_t vstr = {bufinfo.len, bufinfo.len, (char *)bufinfo.buf, true};
364 mp_obj_stringio_t sio = {{&mp_type_stringio}, &vstr, 0, MP_OBJ_NULL};
365 return mod_ujson_load(MP_OBJ_FROM_PTR(&sio));
366 }
367 STATIC MP_DEFINE_CONST_FUN_OBJ_1(mod_ujson_loads_obj, mod_ujson_loads);
368
369 STATIC const mp_rom_map_elem_t mp_module_ujson_globals_table[] = {
370 { MP_ROM_QSTR(MP_QSTR___name__), MP_ROM_QSTR(MP_QSTR_ujson) },
371 { MP_ROM_QSTR(MP_QSTR_dump), MP_ROM_PTR(&mod_ujson_dump_obj) },
372 { MP_ROM_QSTR(MP_QSTR_dumps), MP_ROM_PTR(&mod_ujson_dumps_obj) },
373 { MP_ROM_QSTR(MP_QSTR_load), MP_ROM_PTR(&mod_ujson_load_obj) },
374 { MP_ROM_QSTR(MP_QSTR_loads), MP_ROM_PTR(&mod_ujson_loads_obj) },
375 };
376
377 STATIC MP_DEFINE_CONST_DICT(mp_module_ujson_globals, mp_module_ujson_globals_table);
378
379 const mp_obj_module_t mp_module_ujson = {
380 .base = { &mp_type_module },
381 .globals = (mp_obj_dict_t *)&mp_module_ujson_globals,
382 };
383
384 #endif // MICROPY_PY_UJSON
385