embulk-parser-csv_guessable 0.1.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +12 -0
- data/LICENSE.txt +21 -0
- data/README.md +53 -0
- data/build.gradle +97 -0
- data/config/checkstyle/checkstyle.xml +128 -0
- data/config/checkstyle/default.xml +108 -0
- data/gradle/wrapper/gradle-wrapper.jar +0 -0
- data/gradle/wrapper/gradle-wrapper.properties +6 -0
- data/gradlew +169 -0
- data/gradlew.bat +84 -0
- data/lib/embulk/guess/csv_guessable.rb +61 -0
- data/lib/embulk/parser/csv_guessable.rb +3 -0
- data/libs/embulk-standards-0.8.22.jar +0 -0
- data/src/main/java/org/embulk/parser/csv_guessable/CsvGuessableParserPlugin.java +371 -0
- data/src/main/java/org/embulk/parser/csv_guessable/CsvTokenizer.java +512 -0
- data/src/test/java/org/embulk/parser/csv_guessable/TestCsvGuessableParserPlugin.java +81 -0
- data/src/test/resources/data/test.csv +3 -0
- data/src/test/resources/data/test_alias.yml +3 -0
- data/src/test/resources/yml/guess_from_header.yml +9 -0
- data/src/test/resources/yml/original-csv.yml +12 -0
- data/src/test/resources/yml/replace_column_name.yml +13 -0
- metadata +100 -0
@@ -0,0 +1,512 @@
|
|
1
|
+
package org.embulk.parser.csv_guessable;
|
2
|
+
|
3
|
+
import com.google.common.base.Preconditions;
|
4
|
+
import java.util.List;
|
5
|
+
import java.util.ArrayList;
|
6
|
+
import java.util.Deque;
|
7
|
+
import java.util.ArrayDeque;
|
8
|
+
import org.embulk.spi.DataException;
|
9
|
+
import org.embulk.spi.util.LineDecoder;
|
10
|
+
import org.embulk.config.ConfigException;
|
11
|
+
|
12
|
+
public class CsvTokenizer
|
13
|
+
{
|
14
|
+
static enum RecordState
|
15
|
+
{
|
16
|
+
NOT_END, END,
|
17
|
+
}
|
18
|
+
|
19
|
+
static enum ColumnState
|
20
|
+
{
|
21
|
+
BEGIN, VALUE, QUOTED_VALUE, AFTER_QUOTED_VALUE, FIRST_TRIM, LAST_TRIM_OR_VALUE,
|
22
|
+
}
|
23
|
+
|
24
|
+
private static final char END_OF_LINE = '\0';
|
25
|
+
static final char NO_QUOTE = '\0';
|
26
|
+
static final char NO_ESCAPE = '\0';
|
27
|
+
|
28
|
+
private final char delimiterChar;
|
29
|
+
private final String delimiterFollowingString;
|
30
|
+
private final char quote;
|
31
|
+
private final char escape;
|
32
|
+
private final String newline;
|
33
|
+
private final boolean trimIfNotQuoted;
|
34
|
+
private final long maxQuotedSizeLimit;
|
35
|
+
private final String commentLineMarker;
|
36
|
+
private final LineDecoder input;
|
37
|
+
private final String nullStringOrNull;
|
38
|
+
|
39
|
+
private RecordState recordState = RecordState.END; // initial state is end of a record. nextRecord() must be called first
|
40
|
+
private long lineNumber = 0;
|
41
|
+
|
42
|
+
private String line = null;
|
43
|
+
private int linePos = 0;
|
44
|
+
private boolean wasQuotedColumn = false;
|
45
|
+
private List<String> quotedValueLines = new ArrayList<>();
|
46
|
+
private Deque<String> unreadLines = new ArrayDeque<>();
|
47
|
+
|
48
|
+
public CsvTokenizer(LineDecoder input, CsvGuessableParserPlugin.PluginTask task)
|
49
|
+
{
|
50
|
+
String delimiter = task.getDelimiter();
|
51
|
+
if (delimiter.length() == 0) {
|
52
|
+
throw new ConfigException("Empty delimiter is not allowed");
|
53
|
+
} else {
|
54
|
+
this.delimiterChar = delimiter.charAt(0);
|
55
|
+
if (delimiter.length() > 1) {
|
56
|
+
delimiterFollowingString = delimiter.substring(1);
|
57
|
+
} else {
|
58
|
+
delimiterFollowingString = null;
|
59
|
+
}
|
60
|
+
}
|
61
|
+
quote = task.getQuoteChar().or(CsvGuessableParserPlugin.QuoteCharacter.noQuote()).getCharacter();
|
62
|
+
escape = task.getEscapeChar().or(CsvGuessableParserPlugin.EscapeCharacter.noEscape()).getCharacter();
|
63
|
+
newline = task.getNewline().getString();
|
64
|
+
trimIfNotQuoted = task.getTrimIfNotQuoted();
|
65
|
+
maxQuotedSizeLimit = task.getMaxQuotedSizeLimit();
|
66
|
+
commentLineMarker = task.getCommentLineMarker().orNull();
|
67
|
+
nullStringOrNull = task.getNullString().orNull();
|
68
|
+
this.input = input;
|
69
|
+
}
|
70
|
+
|
71
|
+
public long getCurrentLineNumber()
|
72
|
+
{
|
73
|
+
return lineNumber;
|
74
|
+
}
|
75
|
+
|
76
|
+
public boolean skipHeaderLine()
|
77
|
+
{
|
78
|
+
boolean skipped = input.poll() != null;
|
79
|
+
if (skipped) {
|
80
|
+
lineNumber++;
|
81
|
+
}
|
82
|
+
return skipped;
|
83
|
+
}
|
84
|
+
|
85
|
+
// returns skipped line
|
86
|
+
public String skipCurrentLine()
|
87
|
+
{
|
88
|
+
String skippedLine;
|
89
|
+
if (quotedValueLines.isEmpty()) {
|
90
|
+
skippedLine = line;
|
91
|
+
} else {
|
92
|
+
// recover lines of quoted value
|
93
|
+
skippedLine = quotedValueLines.remove(0); // TODO optimize performance
|
94
|
+
unreadLines.addAll(quotedValueLines);
|
95
|
+
lineNumber -= quotedValueLines.size();
|
96
|
+
if (line != null) {
|
97
|
+
unreadLines.add(line);
|
98
|
+
lineNumber -= 1;
|
99
|
+
}
|
100
|
+
quotedValueLines.clear();
|
101
|
+
}
|
102
|
+
recordState = RecordState.END;
|
103
|
+
return skippedLine;
|
104
|
+
}
|
105
|
+
|
106
|
+
public boolean nextFile()
|
107
|
+
{
|
108
|
+
boolean next = input.nextFile();
|
109
|
+
if (next) {
|
110
|
+
lineNumber = 0;
|
111
|
+
}
|
112
|
+
return next;
|
113
|
+
}
|
114
|
+
|
115
|
+
// used by guess-csv
|
116
|
+
public boolean nextRecord()
|
117
|
+
{
|
118
|
+
return nextRecord(true);
|
119
|
+
}
|
120
|
+
|
121
|
+
public boolean nextRecord(boolean skipEmptyLine)
|
122
|
+
{
|
123
|
+
// If at the end of record, read the next line and initialize the state
|
124
|
+
if (recordState != RecordState.END) {
|
125
|
+
throw new TooManyColumnsException("Too many columns");
|
126
|
+
}
|
127
|
+
|
128
|
+
boolean hasNext = nextLine(skipEmptyLine);
|
129
|
+
if (hasNext) {
|
130
|
+
recordState = RecordState.NOT_END;
|
131
|
+
return true;
|
132
|
+
} else {
|
133
|
+
return false;
|
134
|
+
}
|
135
|
+
}
|
136
|
+
|
137
|
+
private boolean nextLine(boolean skipEmptyLine)
|
138
|
+
{
|
139
|
+
while (true) {
|
140
|
+
if (!unreadLines.isEmpty()) {
|
141
|
+
line = unreadLines.removeFirst();
|
142
|
+
} else {
|
143
|
+
line = input.poll();
|
144
|
+
if (line == null) {
|
145
|
+
return false;
|
146
|
+
}
|
147
|
+
}
|
148
|
+
linePos = 0;
|
149
|
+
lineNumber++;
|
150
|
+
|
151
|
+
boolean skip = skipEmptyLine && (
|
152
|
+
line.isEmpty() ||
|
153
|
+
(commentLineMarker != null && line.startsWith(commentLineMarker)));
|
154
|
+
if (!skip) {
|
155
|
+
return true;
|
156
|
+
}
|
157
|
+
}
|
158
|
+
}
|
159
|
+
|
160
|
+
public boolean hasNextColumn()
|
161
|
+
{
|
162
|
+
return recordState == RecordState.NOT_END;
|
163
|
+
}
|
164
|
+
|
165
|
+
public String nextColumn()
|
166
|
+
{
|
167
|
+
if (!hasNextColumn()) {
|
168
|
+
throw new TooFewColumnsException("Too few columns");
|
169
|
+
}
|
170
|
+
|
171
|
+
// reset last state
|
172
|
+
wasQuotedColumn = false;
|
173
|
+
quotedValueLines.clear();
|
174
|
+
|
175
|
+
// local state
|
176
|
+
int valueStartPos = linePos;
|
177
|
+
int valueEndPos = 0; // initialized by VALUE state and used by LAST_TRIM_OR_VALUE and
|
178
|
+
StringBuilder quotedValue = null; // initial by VALUE or FIRST_TRIM state and used by QUOTED_VALUE state
|
179
|
+
ColumnState columnState = ColumnState.BEGIN;
|
180
|
+
|
181
|
+
while (true) {
|
182
|
+
final char c = nextChar();
|
183
|
+
|
184
|
+
switch (columnState) {
|
185
|
+
case BEGIN:
|
186
|
+
// TODO optimization: state is BEGIN only at the first character of a column.
|
187
|
+
// this block can be out of the looop.
|
188
|
+
if (isDelimiter(c)) {
|
189
|
+
// empty value
|
190
|
+
if (delimiterFollowingString == null) {
|
191
|
+
return "";
|
192
|
+
} else if (isDelimiterFollowingFrom(linePos)) {
|
193
|
+
linePos += delimiterFollowingString.length();
|
194
|
+
return "";
|
195
|
+
}
|
196
|
+
// not a delimiter
|
197
|
+
}
|
198
|
+
if (isEndOfLine(c)) {
|
199
|
+
// empty value
|
200
|
+
recordState = RecordState.END;
|
201
|
+
return "";
|
202
|
+
|
203
|
+
} else if (isSpace(c) && trimIfNotQuoted) {
|
204
|
+
columnState = ColumnState.FIRST_TRIM;
|
205
|
+
|
206
|
+
} else if (isQuote(c)) {
|
207
|
+
valueStartPos = linePos; // == 1
|
208
|
+
wasQuotedColumn = true;
|
209
|
+
quotedValue = new StringBuilder();
|
210
|
+
columnState = ColumnState.QUOTED_VALUE;
|
211
|
+
|
212
|
+
} else {
|
213
|
+
columnState = ColumnState.VALUE;
|
214
|
+
}
|
215
|
+
break;
|
216
|
+
|
217
|
+
case FIRST_TRIM:
|
218
|
+
if (isDelimiter(c)) {
|
219
|
+
// empty value
|
220
|
+
if (delimiterFollowingString == null) {
|
221
|
+
return "";
|
222
|
+
} else if (isDelimiterFollowingFrom(linePos)) {
|
223
|
+
linePos += delimiterFollowingString.length();
|
224
|
+
return "";
|
225
|
+
}
|
226
|
+
// not a delimiter
|
227
|
+
}
|
228
|
+
if (isEndOfLine(c)) {
|
229
|
+
// empty value
|
230
|
+
recordState = RecordState.END;
|
231
|
+
return "";
|
232
|
+
|
233
|
+
} else if (isQuote(c)) {
|
234
|
+
// column has heading spaces and quoted. TODO should this be rejected?
|
235
|
+
valueStartPos = linePos;
|
236
|
+
wasQuotedColumn = true;
|
237
|
+
quotedValue = new StringBuilder();
|
238
|
+
columnState = ColumnState.QUOTED_VALUE;
|
239
|
+
|
240
|
+
} else if (isSpace(c)) {
|
241
|
+
// skip this character
|
242
|
+
|
243
|
+
} else {
|
244
|
+
valueStartPos = linePos - 1;
|
245
|
+
columnState = ColumnState.VALUE;
|
246
|
+
}
|
247
|
+
break;
|
248
|
+
|
249
|
+
case VALUE:
|
250
|
+
if (isDelimiter(c)) {
|
251
|
+
if (delimiterFollowingString == null) {
|
252
|
+
return line.substring(valueStartPos, linePos - 1);
|
253
|
+
} else if (isDelimiterFollowingFrom(linePos)) {
|
254
|
+
String value = line.substring(valueStartPos, linePos - 1);
|
255
|
+
linePos += delimiterFollowingString.length();
|
256
|
+
return value;
|
257
|
+
}
|
258
|
+
// not a delimiter
|
259
|
+
}
|
260
|
+
if (isEndOfLine(c)) {
|
261
|
+
recordState = RecordState.END;
|
262
|
+
return line.substring(valueStartPos, linePos);
|
263
|
+
|
264
|
+
} else if (isSpace(c) && trimIfNotQuoted) {
|
265
|
+
valueEndPos = linePos - 1; // this is possibly end of value
|
266
|
+
columnState = ColumnState.LAST_TRIM_OR_VALUE;
|
267
|
+
|
268
|
+
// TODO not implemented yet foo""bar""baz -> [foo, bar, baz].append
|
269
|
+
//} else if (isQuote(c)) {
|
270
|
+
// // In RFC4180, If fields are not enclosed with double quotes, then
|
271
|
+
// // double quotes may not appear inside the fields. But they are often
|
272
|
+
// // included in the fields. We should care about them later.
|
273
|
+
|
274
|
+
} else {
|
275
|
+
// keep VALUE state
|
276
|
+
}
|
277
|
+
break;
|
278
|
+
|
279
|
+
case LAST_TRIM_OR_VALUE:
|
280
|
+
if (isDelimiter(c)) {
|
281
|
+
if (delimiterFollowingString == null) {
|
282
|
+
return line.substring(valueStartPos, valueEndPos);
|
283
|
+
} else if (isDelimiterFollowingFrom(linePos)) {
|
284
|
+
linePos += delimiterFollowingString.length();
|
285
|
+
return line.substring(valueStartPos, valueEndPos);
|
286
|
+
} else {
|
287
|
+
// not a delimiter
|
288
|
+
}
|
289
|
+
}
|
290
|
+
if (isEndOfLine(c)) {
|
291
|
+
recordState = RecordState.END;
|
292
|
+
return line.substring(valueStartPos, valueEndPos);
|
293
|
+
|
294
|
+
} else if (isSpace(c)) {
|
295
|
+
// keep LAST_TRIM_OR_VALUE state
|
296
|
+
|
297
|
+
} else {
|
298
|
+
// this spaces are not trailing spaces. go back to VALUE state
|
299
|
+
columnState = ColumnState.VALUE;
|
300
|
+
}
|
301
|
+
break;
|
302
|
+
|
303
|
+
case QUOTED_VALUE:
|
304
|
+
if (isEndOfLine(c)) {
|
305
|
+
// multi-line quoted value
|
306
|
+
quotedValue.append(line.substring(valueStartPos, linePos));
|
307
|
+
quotedValue.append(newline);
|
308
|
+
quotedValueLines.add(line);
|
309
|
+
if (!nextLine(false)) {
|
310
|
+
throw new InvalidValueException("Unexpected end of line during parsing a quoted value");
|
311
|
+
}
|
312
|
+
valueStartPos = 0;
|
313
|
+
|
314
|
+
} else if (isQuote(c)) {
|
315
|
+
char next = peekNextChar();
|
316
|
+
if (isQuote(next)) { // escaped quote
|
317
|
+
quotedValue.append(line.substring(valueStartPos, linePos));
|
318
|
+
valueStartPos = ++linePos;
|
319
|
+
} else {
|
320
|
+
quotedValue.append(line.substring(valueStartPos, linePos - 1));
|
321
|
+
columnState = ColumnState.AFTER_QUOTED_VALUE;
|
322
|
+
}
|
323
|
+
|
324
|
+
} else if (isEscape(c)) { // isQuote must be checked first in case of quote == escape
|
325
|
+
// In RFC 4180, CSV's escape char is '\"'. But '\\' is often used.
|
326
|
+
char next = peekNextChar();
|
327
|
+
if (isEndOfLine(c)) {
|
328
|
+
// escape end of line. TODO assuming multi-line quoted value without newline?
|
329
|
+
quotedValue.append(line.substring(valueStartPos, linePos));
|
330
|
+
quotedValueLines.add(line);
|
331
|
+
if (!nextLine(false)) {
|
332
|
+
throw new InvalidValueException("Unexpected end of line during parsing a quoted value");
|
333
|
+
}
|
334
|
+
valueStartPos = 0;
|
335
|
+
} else if (isQuote(next) || isEscape(next)) { // escaped quote
|
336
|
+
quotedValue.append(line.substring(valueStartPos, linePos - 1));
|
337
|
+
quotedValue.append(next);
|
338
|
+
valueStartPos = ++linePos;
|
339
|
+
}
|
340
|
+
|
341
|
+
} else {
|
342
|
+
if ((linePos - valueStartPos) + quotedValue.length() > maxQuotedSizeLimit) {
|
343
|
+
throw new QuotedSizeLimitExceededException("The size of the quoted value exceeds the limit size ("+maxQuotedSizeLimit+")");
|
344
|
+
}
|
345
|
+
// keep QUOTED_VALUE state
|
346
|
+
}
|
347
|
+
break;
|
348
|
+
|
349
|
+
case AFTER_QUOTED_VALUE:
|
350
|
+
if (isDelimiter(c)) {
|
351
|
+
if (delimiterFollowingString == null) {
|
352
|
+
return quotedValue.toString();
|
353
|
+
} else if (isDelimiterFollowingFrom(linePos)) {
|
354
|
+
linePos += delimiterFollowingString.length();
|
355
|
+
return quotedValue.toString();
|
356
|
+
}
|
357
|
+
// not a delimiter
|
358
|
+
}
|
359
|
+
if (isEndOfLine(c)) {
|
360
|
+
recordState = RecordState.END;
|
361
|
+
return quotedValue.toString();
|
362
|
+
|
363
|
+
} else if (isSpace(c)) {
|
364
|
+
// column has trailing spaces and quoted. TODO should this be rejected?
|
365
|
+
|
366
|
+
} else {
|
367
|
+
throw new InvalidValueException(String.format("Unexpected extra character '%c' after a value quoted by '%c'", c, quote));
|
368
|
+
}
|
369
|
+
break;
|
370
|
+
|
371
|
+
default:
|
372
|
+
assert false;
|
373
|
+
}
|
374
|
+
}
|
375
|
+
}
|
376
|
+
|
377
|
+
public String nextColumnOrNull()
|
378
|
+
{
|
379
|
+
String v = nextColumn();
|
380
|
+
if (nullStringOrNull == null) {
|
381
|
+
if (v.isEmpty()) {
|
382
|
+
if (wasQuotedColumn) {
|
383
|
+
return "";
|
384
|
+
}
|
385
|
+
else {
|
386
|
+
return null;
|
387
|
+
}
|
388
|
+
}
|
389
|
+
else {
|
390
|
+
return v;
|
391
|
+
}
|
392
|
+
}
|
393
|
+
else {
|
394
|
+
if (v.equals(nullStringOrNull)) {
|
395
|
+
return null;
|
396
|
+
}
|
397
|
+
else {
|
398
|
+
return v;
|
399
|
+
}
|
400
|
+
}
|
401
|
+
}
|
402
|
+
|
403
|
+
public boolean wasQuotedColumn()
|
404
|
+
{
|
405
|
+
return wasQuotedColumn;
|
406
|
+
}
|
407
|
+
|
408
|
+
private char nextChar()
|
409
|
+
{
|
410
|
+
Preconditions.checkState(line != null, "nextColumn is called after end of file");
|
411
|
+
|
412
|
+
if (linePos >= line.length()) {
|
413
|
+
return END_OF_LINE;
|
414
|
+
} else {
|
415
|
+
return line.charAt(linePos++);
|
416
|
+
}
|
417
|
+
}
|
418
|
+
|
419
|
+
private char peekNextChar()
|
420
|
+
{
|
421
|
+
Preconditions.checkState(line != null, "peekNextChar is called after end of file");
|
422
|
+
|
423
|
+
if (linePos >= line.length()) {
|
424
|
+
return END_OF_LINE;
|
425
|
+
} else {
|
426
|
+
return line.charAt(linePos);
|
427
|
+
}
|
428
|
+
}
|
429
|
+
|
430
|
+
private boolean isSpace(char c)
|
431
|
+
{
|
432
|
+
return c == ' ';
|
433
|
+
}
|
434
|
+
|
435
|
+
private boolean isDelimiterFollowingFrom(int pos)
|
436
|
+
{
|
437
|
+
if (line.length() < pos + delimiterFollowingString.length()) {
|
438
|
+
return false;
|
439
|
+
}
|
440
|
+
for (int i = 0; i < delimiterFollowingString.length(); i++) {
|
441
|
+
if (delimiterFollowingString.charAt(i) != line.charAt(pos + i)) {
|
442
|
+
return false;
|
443
|
+
}
|
444
|
+
}
|
445
|
+
return true;
|
446
|
+
}
|
447
|
+
|
448
|
+
private boolean isDelimiter(char c)
|
449
|
+
{
|
450
|
+
return c == delimiterChar;
|
451
|
+
}
|
452
|
+
|
453
|
+
private boolean isEndOfLine(char c)
|
454
|
+
{
|
455
|
+
return c == END_OF_LINE;
|
456
|
+
}
|
457
|
+
|
458
|
+
private boolean isQuote(char c)
|
459
|
+
{
|
460
|
+
return quote != NO_QUOTE && c == quote;
|
461
|
+
}
|
462
|
+
|
463
|
+
private boolean isEscape(char c)
|
464
|
+
{
|
465
|
+
return escape != NO_ESCAPE && c == escape;
|
466
|
+
}
|
467
|
+
|
468
|
+
public static class InvalidFormatException
|
469
|
+
extends DataException
|
470
|
+
{
|
471
|
+
public InvalidFormatException(String message)
|
472
|
+
{
|
473
|
+
super(message);
|
474
|
+
}
|
475
|
+
}
|
476
|
+
|
477
|
+
public static class InvalidValueException
|
478
|
+
extends DataException
|
479
|
+
{
|
480
|
+
public InvalidValueException(String message)
|
481
|
+
{
|
482
|
+
super(message);
|
483
|
+
}
|
484
|
+
}
|
485
|
+
|
486
|
+
public static class QuotedSizeLimitExceededException
|
487
|
+
extends InvalidValueException
|
488
|
+
{
|
489
|
+
public QuotedSizeLimitExceededException(String message)
|
490
|
+
{
|
491
|
+
super(message);
|
492
|
+
}
|
493
|
+
}
|
494
|
+
|
495
|
+
public class TooManyColumnsException
|
496
|
+
extends InvalidFormatException
|
497
|
+
{
|
498
|
+
public TooManyColumnsException(String message)
|
499
|
+
{
|
500
|
+
super(message);
|
501
|
+
}
|
502
|
+
}
|
503
|
+
|
504
|
+
public class TooFewColumnsException
|
505
|
+
extends InvalidFormatException
|
506
|
+
{
|
507
|
+
public TooFewColumnsException(String message)
|
508
|
+
{
|
509
|
+
super(message);
|
510
|
+
}
|
511
|
+
}
|
512
|
+
}
|