Alexandria 2.25.0
SDC-CH common library for the Euclid project
AsciiReaderHelper.cpp
Go to the documentation of this file.
1/*
2 * Copyright (C) 2012-2022 Euclid Science Ground Segment
3 *
4 * This library is free software; you can redistribute it and/or modify it under
5 * the terms of the GNU Lesser General Public License as published by the Free
6 * Software Foundation; either version 3.0 of the License, or (at your option)
7 * any later version.
8 *
9 * This library is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
11 * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
12 * details.
13 *
14 * You should have received a copy of the GNU Lesser General Public License
15 * along with this library; if not, write to the Free Software Foundation, Inc.,
16 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17 */
18
25#include "AsciiReaderHelper.h"
28#include "NdArray/NdArray.h"
29#include <boost/algorithm/string.hpp>
30#include <boost/lexical_cast.hpp>
31#include <boost/spirit/include/qi.hpp>
32#include <boost/tokenizer.hpp>
33#include <set>
34#include <sstream>
35
36namespace Euclid {
37namespace Table {
38
40
42
43size_t countColumns(std::istream& in, const std::string& comment) {
44 StreamRewinder rewinder{in};
45 size_t count = 0;
46
47 while (in) {
48 std::string line;
49 getline(in, line);
50 // Remove any comments
51 size_t comment_pos = line.find(comment);
52 if (comment_pos != std::string::npos) {
53 line = line.substr(0, comment_pos);
54 }
55 boost::trim(line);
56 if (!line.empty()) {
57 std::string token;
58 std::stringstream line_stream(line);
59 line_stream >> boost::io::quoted(token);
60 while (line_stream) {
61 line_stream >> boost::io::quoted(token);
62 ++count;
63 }
64 break;
65 }
66 }
67 if (count == 0) {
68 throw Elements::Exception() << "No data lines found";
69 }
70 return count;
71}
72
76 // Boolean
77 {"bool", typeid(bool)},
78 {"boolean", typeid(bool)},
79 // Integers
80 {"int", typeid(int32_t)},
81 {"long", typeid(int64_t)},
82 {"int32", typeid(int32_t)},
83 {"int64", typeid(int64_t)},
84 // Floating point
85 {"float", typeid(float)},
86 {"double", typeid(double)},
87 // Strings
88 {"string", typeid(std::string)},
89 // Arrays
90 {"[bool]", typeid(std::vector<bool>)},
91 {"[boolean]", typeid(std::vector<bool>)},
92 {"[int]", typeid(std::vector<int32_t>)},
93 {"[long]", typeid(std::vector<int64_t>)},
94 {"[int32]", typeid(std::vector<int32_t>)},
95 {"[int64]", typeid(std::vector<int64_t>)},
96 {"[float]", typeid(std::vector<float>)},
97 {"[double]", typeid(std::vector<double>)},
98 // NdArrays
99 {"[int+]", typeid(NdArray<int32_t>)},
100 {"[long+]", typeid(NdArray<int64_t>)},
101 {"[int32+]", typeid(NdArray<int32_t>)},
102 {"[int64+]", typeid(NdArray<int64_t>)},
103 {"[float+]", typeid(NdArray<float>)},
104 {"[double+]", typeid(NdArray<double>)},
105};
106
108 for (auto p = KeywordTypeMap.begin(); p != KeywordTypeMap.end(); ++p) {
109 if (p->first == keyword) {
110 return p->second;
111 }
112 }
113 throw Elements::Exception() << "Unknown column type keyword " << keyword;
114}
115
117 StreamRewinder rewinder{in};
119 while (in) {
120 std::string line;
121 getline(in, line);
122 boost::trim(line);
123 if (line.empty()) {
124 continue; // We skip empty lines
125 }
126 if (boost::starts_with(line, comment)) {
127 // If we have a comment we remove all comment characters and check if we have
128 // a column description
129 boost::replace_all(line, comment, "");
130 boost::trim(line);
131 if (boost::starts_with(line, "Column:")) {
132 line.erase(0, 7);
133 boost::trim(line);
134 if (!line.empty()) {
135 std::string token;
136 std::stringstream line_stream(line);
137 std::string name;
138 line_stream >> boost::io::quoted(name);
139 if (descriptions.count(name) != 0) {
140 throw Elements::Exception() << "Duplicate column name " << name;
141 }
142 line_stream >> boost::io::quoted(token);
143 std::type_index type = typeid(std::string);
144 if (line_stream) {
145 if (!boost::starts_with(token, "(") && token != "-") {
146 type = keywordToType(token);
147 line_stream >> boost::io::quoted(token);
148 }
149 }
150 std::string unit = "";
151 if (line_stream) {
152 if (boost::starts_with(token, "(")) {
153 unit = token;
154 unit.erase(unit.begin());
155 unit.erase(unit.end() - 1);
156 line_stream >> boost::io::quoted(token);
157 }
158 }
159 if (line_stream && token == "-") {
160 line_stream >> boost::io::quoted(token);
161 }
163 while (line_stream) {
164 desc << token << ' ';
165 line_stream >> boost::io::quoted(token);
166 }
167 std::string desc_str = desc.str();
168 boost::trim(desc_str);
169 descriptions.emplace(std::piecewise_construct, std::forward_as_tuple(name),
170 std::forward_as_tuple(name, type, unit, desc_str));
171 }
172 }
173 } else {
174 break; // here we reached the first data line
175 }
176 }
177 return descriptions;
178}
179
180std::vector<std::string> autoDetectColumnNames(std::istream& in, const std::string& comment, size_t columns_number) {
181 StreamRewinder rewinder{in};
183
184 // Find the last comment line and at the same time read the names of the
185 // column info description comments
186 std::string last_comment{};
187 std::vector<std::string> desc_names{};
188 while (in) {
189 std::string line;
190 getline(in, line);
191 boost::trim(line);
192 if (line.empty()) {
193 continue; // We skip empty lines
194 }
195 if (boost::starts_with(line, comment)) {
196 // If we have a comment we remove all comment characters and check if we have
197 // the correct number of tokens
198 boost::replace_all(line, comment, "");
199 boost::trim(line);
200 if (!line.empty()) {
201 last_comment = line;
202 }
203 if (boost::starts_with(line, "Column:")) {
204 std::string temp = line;
205 temp.erase(0, 7);
206 boost::trim(temp);
207 auto space_i = temp.find(' ');
208 if (space_i > 0) {
209 temp = temp.substr(0, space_i);
210 }
211 desc_names.emplace_back(std::move(temp));
212 }
213 } else {
214 break; // here we reached the first data line
215 }
216 }
217
218 // Check if the last comment line contains the names of the columns
219 if (!last_comment.empty()) {
220 std::stringstream line_stream(last_comment);
221 std::string token;
222 line_stream >> boost::io::quoted(token);
223 while (line_stream) {
224 names.push_back(token);
225 line_stream >> boost::io::quoted(token);
226 }
227 if (names.size() != columns_number) {
228 names.clear();
229 }
230 }
231
232 // If the names are empty we fill them with the column descriprion ones
233 if (names.empty()) {
234 if (desc_names.size() != 0 && desc_names.size() != columns_number) {
235 logger.warn() << "Number of column descriptions does not matches the number"
236 << " of the columns";
237 }
238 names = desc_names;
239 }
240
241 if (names.size() < columns_number) {
242 for (size_t i = names.size() + 1; i <= columns_number; ++i) {
243 names.push_back("col" + std::to_string(i));
244 }
245 }
246 // Check for duplicate names
248 for (auto name : names) {
249 if (!set.insert(name).second) {
250 throw Elements::Exception() << "Duplicate column name " << name;
251 }
252 }
253 return names;
254}
255
256namespace {
257
258template <typename T>
259std::vector<T> convertStringToVector(const std::string& str) {
260 std::vector<T> result{};
261 boost::char_separator<char> sep{","};
262 boost::tokenizer<boost::char_separator<char>> tok{str, sep};
263 for (auto& s : tok) {
264 result.push_back(boost::get<T>(convertToCellType(s, typeid(T))));
265 }
266 return result;
267}
268
269template <typename T>
270NdArray<T> convertStringToNdArray(const std::string& str) {
271 if (str.empty()) {
272 throw Elements::Exception() << "Cannot convert an empty string to a NdArray";
273 } else if (str[0] != '<') {
274 throw Elements::Exception() << "Unexpected initial character for a NdArray: " << str[0];
275 }
276
277 auto closing_char = str.find('>');
278 if (closing_char == std::string::npos) {
279 throw Elements::Exception() << "Could not find '>'";
280 }
281
282 auto shape_str = str.substr(1, closing_char - 1);
283 auto shape_i = convertStringToVector<int32_t>(shape_str);
284 auto data = convertStringToVector<T>(str.substr(closing_char + 1));
285
286 std::vector<size_t> shape_u;
287 std::copy(shape_i.begin(), shape_i.end(), std::back_inserter(shape_u));
288 return NdArray<T>(shape_u, data);
289}
290
291} // namespace
292
294 // Boolean
295 {typeid(bool),
296 [](const std::string& value) {
297 if (value == "true" || value == "t" || value == "yes" || value == "y" || value == "1") {
298 return true;
299 } else if (value == "false" || value == "f" || value == "no" || value == "n" || value == "0") {
300 return false;
301 }
302 throw Elements::Exception() << "Invalid boolean value " << value;
303 }},
304 // Integers
305 {typeid(int32_t), boost::lexical_cast<int32_t, const std::string&>},
306 {typeid(int64_t), boost::lexical_cast<int64_t, const std::string&>},
307 // Floating point
308 {typeid(float), boost::lexical_cast<float, const std::string&>},
309 {typeid(double), boost::lexical_cast<double, const std::string&>},
310 // String
311 {typeid(std::string), boost::lexical_cast<std::string, const std::string&>},
312 // Arrays
313 {typeid(std::vector<bool>), convertStringToVector<bool>},
314 {typeid(std::vector<int32_t>), convertStringToVector<int32_t>},
315 {typeid(std::vector<int64_t>), convertStringToVector<int64_t>},
316 {typeid(std::vector<float>), convertStringToVector<float>},
317 {typeid(std::vector<double>), convertStringToVector<double>},
318 // NdArray
319 {typeid(NdArray<int32_t>), convertStringToNdArray<int32_t>},
320 {typeid(NdArray<int64_t>), convertStringToNdArray<int64_t>},
321 {typeid(NdArray<float>), convertStringToNdArray<float>},
322 {typeid(NdArray<double>), convertStringToNdArray<double>},
323};
324
326 try {
327 auto i = sCellConverter.find(type);
328 if (i == sCellConverter.end()) {
329 throw Elements::Exception() << "Unknown type name " << type.name();
330 }
331 return i->second(value);
332 } catch (boost::bad_lexical_cast const&) {
333 throw Elements::Exception() << "Cannot convert " << value << " to " << type.name();
334 }
335}
336
337bool hasNextRow(std::istream& in, const std::string& comment) {
338 StreamRewinder rewinder{in};
339 while (in) {
340 std::string line;
341 getline(in, line);
342 size_t comment_pos = line.find(comment);
343 if (comment_pos != std::string::npos) {
344 line = line.substr(0, comment_pos);
345 }
346 boost::trim(line);
347 if (!line.empty()) {
348 return true;
349 }
350 }
351 return false;
352}
353
355 StreamRewinder rewinder{in};
356 std::size_t count = 0;
357 while (in) {
358 std::string line;
359 getline(in, line);
360 size_t comment_pos = line.find(comment);
361 if (comment_pos != std::string::npos) {
362 line = line.substr(0, comment_pos);
363 }
364 boost::trim(line);
365 if (!line.empty()) {
366 ++count;
367 }
368 }
369 return count;
370}
371
374 size_t comment_pos = line.find(comment);
375
376 if (comment_pos != std::string::npos) {
377 line = line.substr(0, comment_pos);
378 }
379 boost::trim(line);
380 if (!line.empty()) {
381 std::stringstream line_stream(line);
382 size_t count = 0;
383 std::string token;
384 line_stream >> boost::io::quoted(token);
385 while (line_stream) {
386 cells.emplace_back(token);
387 line_stream >> boost::io::quoted(token);
388 ++count;
389 }
390 }
391 return cells;
392}
393
395 StreamRewinder rewinder{in};
396 std::string line(comment);
397 while (in && boost::starts_with(line, comment)) {
398 getline(in, line);
399 }
400 return splitLine(line, comment);
401}
402
404 namespace qi = boost::spirit::qi;
405 double d;
406 long l;
407
408 auto it1 = token.begin(), it2 = it1;
409 if (qi::parse(it1, token.end(), qi::long_, l) && it1 == token.end()) {
410 return {typeid(int64_t), 0};
411 }
412 if (qi::parse(it2, token.end(), qi::double_, d) && it2 == token.end()) {
413 return {typeid(double), 0};
414 }
415 return {typeid(std::string), std::size_t(0)};
416}
417
418} // namespace Table
419} // end of namespace Euclid
T back_inserter(T... args)
T begin(T... args)
static Logging getLogger(const std::string &name="")
void warn(const std::string &logMessage)
NdArray(std::vector< size_t > shape_)
boost::variant< bool, int32_t, int64_t, float, double, std::string, std::vector< bool >, std::vector< int32_t >, std::vector< int64_t >, std::vector< float >, std::vector< double >, NdArray::NdArray< int32_t >, NdArray::NdArray< int64_t >, NdArray::NdArray< float >, NdArray::NdArray< double > > cell_type
The possible cell types.
Definition: Row.h:71
This class gets a stream as argument during construction and when it is deleted it sets the position ...
T copy(T... args)
T count(T... args)
T emplace_back(T... args)
T emplace(T... args)
T empty(T... args)
T end(T... args)
T erase(T... args)
T find(T... args)
T forward_as_tuple(T... args)
T move(T... args)
T name(T... args)
std::vector< std::string > splitLine(std::string line, const std::string &comment)
std::type_index keywordToType(const std::string &keyword)
bool hasNextRow(std::istream &in, const std::string &comment)
std::map< std::string, ColumnDescription > autoDetectColumnDescriptions(std::istream &in, const std::string &comment)
Reads the column descriptions of the given stream.
size_t countColumns(std::istream &in, const std::string &comment)
Returns the number of whitespace separated tokens of the first non commented line.
std::vector< std::string > autoDetectColumnNames(std::istream &in, const std::string &comment, size_t columns_number)
Reads the column names of the given stream.
std::pair< std::type_index, std::size_t > guessColumnType(const std::string &token)
Row::cell_type convertToCellType(const std::string &value, std::type_index type)
Converts the given value to a Row::cell_type of the given type.
std::vector< std::string > firstDataLine(std::istream &in, const std::string &comment)
const std::vector< std::pair< std::string, std::type_index > > KeywordTypeMap
std::size_t countRemainingRows(std::istream &in, const std::string &comment)
const std::map< std::type_index, std::function< Row::cell_type(const std::string &)> > sCellConverter
static Elements::Logging logger
std::string quoted(const std::string &str)
T push_back(T... args)
T str(T... args)
T substr(T... args)
T to_string(T... args)