src/zserio/JsonTokenizer.h
Line | Count | Source (jump to first uncovered line) |
1 | | #ifndef ZSERIO_JSON_TOKENIZER_H_INC |
2 | | #define ZSERIO_JSON_TOKENIZER_H_INC |
3 | | |
4 | | #include <array> |
5 | | #include <istream> |
6 | | #include <memory> |
7 | | |
8 | | #include "zserio/AnyHolder.h" |
9 | | #include "zserio/CppRuntimeException.h" |
10 | | #include "zserio/JsonDecoder.h" |
11 | | #include "zserio/Types.h" |
12 | | |
13 | | namespace zserio |
14 | | { |
15 | | |
16 | | /** |
17 | | * Tokens used by Json Tokenizer. |
18 | | */ |
19 | | enum class JsonToken : int8_t |
20 | | { |
21 | | UNKNOWN = -1, |
22 | | BEGIN_OF_FILE, |
23 | | END_OF_FILE, |
24 | | BEGIN_OBJECT, |
25 | | END_OBJECT, |
26 | | BEGIN_ARRAY, |
27 | | END_ARRAY, |
28 | | KEY_SEPARATOR, |
29 | | ITEM_SEPARATOR, |
30 | | VALUE |
31 | | }; |
32 | | |
33 | | /** |
34 | | * Exception used to distinguish exceptions from the JsonParser. |
35 | | */ |
36 | | class JsonParserException : public CppRuntimeException |
37 | | { |
38 | | public: |
39 | | using CppRuntimeException::CppRuntimeException; |
40 | | }; |
41 | | |
42 | | /** |
43 | | * Allows to append JsonToken to CppRuntimeException. |
44 | | * |
45 | | * \param exception Exception to modify. |
46 | | * \param token JSON Token to append. |
47 | | * |
48 | | * \return Reference to the exception to allow operator chaining. |
49 | | */ |
50 | | CppRuntimeException& operator<<(CppRuntimeException& exception, JsonToken token); |
51 | | |
52 | | /** |
53 | | * Json Tokenizer used by Json Parser. |
54 | | */ |
55 | | template <typename ALLOC = std::allocator<uint8_t>> |
56 | | class BasicJsonTokenizer |
57 | | { |
58 | | public: |
59 | | /** |
60 | | * Constructor. |
61 | | * |
62 | | * \param in Input stream to tokenize. |
63 | | * \param allocator Allocator to use. |
64 | | */ |
65 | | BasicJsonTokenizer(std::istream& in, const ALLOC& allocator) : |
66 | | m_buffer(), |
67 | | m_in(in), |
68 | | m_decoder(allocator), |
69 | | m_decoderResult(0, allocator), |
70 | | m_content(readContent(allocator)), |
71 | | m_value(allocator) |
72 | 93 | { |
73 | 93 | m_token = m_content.empty() ? JsonToken::END_OF_FILE1 : JsonToken::BEGIN_OF_FILE92 ; |
74 | 93 | } |
75 | | |
76 | | /** |
77 | | * Move to the next token. |
78 | | * |
79 | | * \return Next token. |
80 | | * \throw JsonParserException In case that tokenizing fails - i.e. unknown token is reached. |
81 | | */ |
82 | | JsonToken next(); |
83 | | |
84 | | /** |
85 | | * Gets current token. |
86 | | * |
87 | | * \return Current token. |
88 | | */ |
89 | | JsonToken getToken() const |
90 | 1.87k | { |
91 | 1.87k | return m_token; |
92 | 1.87k | } |
93 | | |
94 | | /** |
95 | | * Gets current value. |
96 | | * |
97 | | * Any holder can be either unset - i.e. beginning or end of the input, |
98 | | * or it can hold one of the types defined in IObserver::visitValue. |
99 | | * |
100 | | * \return Current value as an AnyHolder. |
101 | | */ |
102 | | const AnyHolder<ALLOC>& getValue() const |
103 | 48.4k | { |
104 | 48.4k | return m_value; |
105 | 48.4k | } |
106 | | |
107 | | /** |
108 | | * Gets line number of the current token. |
109 | | * |
110 | | * \return Line number. |
111 | | */ |
112 | | size_t getLine() const |
113 | 48.0k | { |
114 | 48.0k | return m_lineNumber; |
115 | 48.0k | } |
116 | | |
117 | | /** |
118 | | * Gets column number of the current token. |
119 | | * |
120 | | * \return Column number. |
121 | | */ |
122 | | size_t getColumn() const |
123 | 48.0k | { |
124 | 48.0k | return m_tokenColumnNumber; |
125 | 48.0k | } |
126 | | |
127 | | private: |
128 | | string<ALLOC> readContent(const ALLOC& allocator); |
129 | | |
130 | | bool decodeNext(); |
131 | | bool skipWhitespaces(); |
132 | | |
133 | | template <typename T> |
134 | | void setToken(JsonToken token, T&& value); |
135 | | void setToken(JsonToken token, AnyHolder<ALLOC>&& value); |
136 | | void setToken(JsonToken token); |
137 | | void setPosition(size_t newPos, size_t newColumnNumber); |
138 | | void setTokenValue(); |
139 | | |
140 | | static constexpr size_t BUFFER_SIZE = 64 * 1024; |
141 | | std::array<char, BUFFER_SIZE> m_buffer; |
142 | | |
143 | | std::istream& m_in; |
144 | | BasicJsonDecoder<ALLOC> m_decoder; |
145 | | typename BasicJsonDecoder<ALLOC>::DecoderResult m_decoderResult; |
146 | | string<ALLOC> m_content; |
147 | | size_t m_lineNumber = 1; |
148 | | size_t m_columnNumber = 1; |
149 | | size_t m_tokenColumnNumber = 1; |
150 | | size_t m_pos = 0; |
151 | | JsonToken m_token; |
152 | | AnyHolder<ALLOC> m_value; |
153 | | }; |
154 | | |
155 | | template <typename ALLOC> |
156 | | JsonToken BasicJsonTokenizer<ALLOC>::next() |
157 | 49.1k | { |
158 | 49.2k | while (!decodeNext()) |
159 | 60 | { |
160 | 60 | string<ALLOC> newContent = readContent(m_content.get_allocator()); |
161 | 60 | if (newContent.empty()) |
162 | 57 | { |
163 | 57 | if (m_token == JsonToken::END_OF_FILE) |
164 | 52 | { |
165 | 52 | m_tokenColumnNumber = m_columnNumber; |
166 | 52 | } |
167 | 5 | else |
168 | 5 | { |
169 | | // stream is finished but last token is not EOF => value must be at the end |
170 | 5 | setTokenValue(); |
171 | 5 | } |
172 | | |
173 | 57 | return m_token; |
174 | 57 | } |
175 | | |
176 | 3 | m_content = m_content.substr(m_pos) + newContent; |
177 | 3 | m_pos = 0; |
178 | 3 | } |
179 | | |
180 | 49.1k | return m_token; |
181 | 49.1k | } |
182 | | |
183 | | template <typename ALLOC> |
184 | | string<ALLOC> BasicJsonTokenizer<ALLOC>::readContent(const ALLOC& allocator) |
185 | 153 | { |
186 | 153 | const size_t count = static_cast<size_t>(m_in.rdbuf()->sgetn(m_buffer.data(), BUFFER_SIZE)); |
187 | 153 | return string<ALLOC>(m_buffer.data(), count, allocator); |
188 | 153 | } |
189 | | |
190 | | template <typename ALLOC> |
191 | | bool BasicJsonTokenizer<ALLOC>::decodeNext() |
192 | 49.2k | { |
193 | 49.2k | if (!skipWhitespaces()) |
194 | 52 | { |
195 | 52 | return false; |
196 | 52 | } |
197 | | |
198 | 49.1k | m_tokenColumnNumber = m_columnNumber; |
199 | | |
200 | 49.1k | const char nextChar = m_content[m_pos]; |
201 | 49.1k | switch (nextChar) |
202 | 49.1k | { |
203 | 164 | case '{': |
204 | 164 | setToken(JsonToken::BEGIN_OBJECT, nextChar); |
205 | 164 | setPosition(m_pos + 1, m_columnNumber + 1); |
206 | 164 | break; |
207 | 96 | case '}': |
208 | 96 | setToken(JsonToken::END_OBJECT, nextChar); |
209 | 96 | setPosition(m_pos + 1, m_columnNumber + 1); |
210 | 96 | break; |
211 | 44 | case '[': |
212 | 44 | setToken(JsonToken::BEGIN_ARRAY, nextChar); |
213 | 44 | setPosition(m_pos + 1, m_columnNumber + 1); |
214 | 44 | break; |
215 | 38 | case ']': |
216 | 38 | setToken(JsonToken::END_ARRAY, nextChar); |
217 | 38 | setPosition(m_pos + 1, m_columnNumber + 1); |
218 | 38 | break; |
219 | 12.2k | case ':': |
220 | 12.2k | setToken(JsonToken::KEY_SEPARATOR, nextChar); |
221 | 12.2k | setPosition(m_pos + 1, m_columnNumber + 1); |
222 | 12.2k | break; |
223 | 12.1k | case ',': |
224 | 12.1k | setToken(JsonToken::ITEM_SEPARATOR, nextChar); |
225 | 12.1k | setPosition(m_pos + 1, m_columnNumber + 1); |
226 | 12.1k | break; |
227 | 24.4k | default: |
228 | 24.4k | m_decoderResult = m_decoder.decodeValue(StringView(m_content.data()).substr(m_pos)); |
229 | 24.4k | if (m_pos + m_decoderResult.numReadChars >= m_content.size()) |
230 | 8 | { |
231 | 8 | return false; // we are at the end of chunk => read more |
232 | 8 | } |
233 | | |
234 | 24.4k | setTokenValue(); |
235 | 24.4k | break; |
236 | 49.1k | } |
237 | | |
238 | 49.1k | return true; |
239 | 49.1k | } |
240 | | |
241 | | template <typename ALLOC> |
242 | | bool BasicJsonTokenizer<ALLOC>::skipWhitespaces() |
243 | 49.2k | { |
244 | 125k | while (true) |
245 | 125k | { |
246 | 125k | if (m_pos >= m_content.size()) |
247 | 51 | { |
248 | 51 | setToken(JsonToken::END_OF_FILE); |
249 | 51 | return false; |
250 | 51 | } |
251 | | |
252 | 125k | const char nextChar = m_content[m_pos]; |
253 | 125k | switch (nextChar) |
254 | 125k | { |
255 | 63.4k | case ' ': |
256 | 63.4k | case '\t': |
257 | 63.4k | setPosition(m_pos + 1, m_columnNumber + 1); |
258 | 63.4k | break; |
259 | 12.4k | case '\n': |
260 | 12.4k | m_lineNumber++; |
261 | 12.4k | setPosition(m_pos + 1, 1); |
262 | 12.4k | break; |
263 | 3 | case '\r': |
264 | 3 | if (m_pos + 1 >= m_content.size()) |
265 | 1 | { |
266 | 1 | setToken(JsonToken::END_OF_FILE); |
267 | 1 | return false; |
268 | 1 | } |
269 | 2 | m_lineNumber++; |
270 | 2 | setPosition(m_pos + (m_content[m_pos + 1] == '\n' ? 21 : 11 ), 1); |
271 | 2 | break; |
272 | 49.1k | default: |
273 | 49.1k | return true; |
274 | 125k | } |
275 | 125k | } |
276 | 49.2k | } |
277 | | |
278 | | template <typename ALLOC> |
279 | | template <typename T> |
280 | | void BasicJsonTokenizer<ALLOC>::setToken(JsonToken token, T&& value) |
281 | 24.6k | { |
282 | 24.6k | m_token = token; |
283 | 24.6k | m_value.set(std::forward<T>(value)); |
284 | 24.6k | } |
285 | | |
286 | | template <typename ALLOC> |
287 | | void BasicJsonTokenizer<ALLOC>::setToken(JsonToken token, AnyHolder<ALLOC>&& value) |
288 | 24.4k | { |
289 | 24.4k | m_token = token; |
290 | 24.4k | m_value = std::move(value); |
291 | 24.4k | } |
292 | | |
293 | | template <typename ALLOC> |
294 | | void BasicJsonTokenizer<ALLOC>::setToken(JsonToken token) |
295 | 52 | { |
296 | 52 | m_token = token; |
297 | 52 | m_value.reset(); |
298 | 52 | } |
299 | | |
300 | | template <typename ALLOC> |
301 | | void BasicJsonTokenizer<ALLOC>::setPosition(size_t newPos, size_t newColumnNumber) |
302 | 125k | { |
303 | 125k | m_pos = newPos; |
304 | 125k | m_columnNumber = newColumnNumber; |
305 | 125k | } |
306 | | |
307 | | template <typename ALLOC> |
308 | | void BasicJsonTokenizer<ALLOC>::setTokenValue() |
309 | 24.4k | { |
310 | 24.4k | if (!m_decoderResult.value.hasValue()) |
311 | 2 | { |
312 | 2 | throw JsonParserException("JsonTokenizer:") |
313 | 2 | << m_lineNumber << ":" << m_tokenColumnNumber << ": " |
314 | 2 | << (m_decoderResult.integerOverflow ? "Value is outside of the 64-bit integer range!"1 |
315 | 2 | : "Unknown token!"1 ); |
316 | 2 | } |
317 | | |
318 | 24.4k | setToken(JsonToken::VALUE, std::move(m_decoderResult.value)); |
319 | 24.4k | setPosition(m_pos + m_decoderResult.numReadChars, m_columnNumber + m_decoderResult.numReadChars); |
320 | 24.4k | } |
321 | | |
322 | | } // namespace zserio |
323 | | |
324 | | #endif // ZSERIO_JSON_TOKENIZER_H_INC |