svn-gvsig-desktop / trunk / org.gvsig.desktop / org.gvsig.desktop.compat.cdc / org.gvsig.fmap.dal / org.gvsig.fmap.dal.file / org.gvsig.fmap.dal.file.csv / src / main / java / org / gvsig / fmap / dal / store / csv / virtualrows / RandomAccessFileReader.java @ 47506
History | View | Annotate | Download (36.9 KB)
1 |
package org.gvsig.fmap.dal.store.csv.virtualrows; |
---|---|
2 |
|
3 |
import java.io.BufferedReader; |
4 |
import java.io.File; |
5 |
import java.io.IOException; |
6 |
import java.io.RandomAccessFile; |
7 |
import java.io.Reader; |
8 |
import java.io.UncheckedIOException; |
9 |
import java.nio.ByteBuffer; |
10 |
import java.nio.CharBuffer; |
11 |
import java.nio.channels.Channels; |
12 |
import java.nio.charset.Charset; |
13 |
import java.util.Iterator; |
14 |
import java.util.NoSuchElementException; |
15 |
import java.util.Spliterator; |
16 |
import java.util.Spliterators; |
17 |
import java.util.function.Function; |
18 |
import java.util.function.Predicate; |
19 |
import java.util.stream.Stream; |
20 |
import java.util.stream.StreamSupport; |
21 |
import org.apache.commons.io.FilenameUtils; |
22 |
import org.apache.commons.io.IOUtils; |
23 |
import org.apache.commons.lang3.StringUtils; |
24 |
import org.gvsig.tools.ToolsLocator; |
25 |
import org.gvsig.tools.i18n.I18nManager; |
26 |
import org.gvsig.tools.library.impl.DefaultLibrariesInitializer; |
27 |
import org.gvsig.tools.observer.Observable; |
28 |
import org.gvsig.tools.task.SimpleTaskStatus; |
29 |
import org.gvsig.tools.task.TaskStatus; |
30 |
import org.gvsig.tools.task.TaskStatusManager; |
31 |
|
32 |
/**
|
33 |
*
|
34 |
* @author gvSIG Team
|
35 |
*/
|
36 |
public class RandomAccessFileReader extends Reader { |
37 |
|
38 |
public static final Predicate<String> FILTER_NONE = (String t) -> false; |
39 |
|
40 |
protected static final int INDEX_HEADER_FILESIZE = 0; |
41 |
protected static final int INDEX_HEADER_INDEXCREATIONCOST = 1; |
42 |
|
43 |
protected static final int MAX_BUFFER_FOR_LINE = 50*1024; //50K |
44 |
|
45 |
protected RandomAccessFile raf; |
46 |
protected Reader reader; |
47 |
protected long currentPosition; |
48 |
protected final Charset charset; |
49 |
protected long lastModified; |
50 |
|
51 |
public RandomAccessFileReader(File f, String charsetName) throws IOException { |
52 |
this(new RandomAccessFile(f, "r"), Charset.forName(charsetName)); |
53 |
this.lastModified = f.lastModified();
|
54 |
} |
55 |
|
56 |
public RandomAccessFileReader(File f, Charset charset) throws IOException { |
57 |
this(new RandomAccessFile(f, "r"), charset); |
58 |
this.lastModified = f.lastModified();
|
59 |
} |
60 |
|
61 |
public RandomAccessFileReader(RandomAccessFile raf, String charsetName) throws IOException { |
62 |
this(raf, Charset.forName(charsetName)); |
63 |
this.lastModified = -1; |
64 |
} |
65 |
|
66 |
public RandomAccessFileReader(RandomAccessFile raf, Charset charset) throws IOException { |
67 |
this.charset = charset;
|
68 |
this.raf = raf;
|
69 |
this.reader = null; |
70 |
this.currentPosition = 0; |
71 |
this.lastModified = -1; |
72 |
} |
73 |
|
74 |
public Charset getCharset() { |
75 |
return this.charset; |
76 |
} |
77 |
|
78 |
@Override
|
79 |
public int read(char[] cbuf, int off, int len) throws IOException { |
80 |
if (this.reader == null) { |
81 |
this.createReader();
|
82 |
} |
83 |
int n = this.reader.read(cbuf, off, len); |
84 |
if (n > 0) { |
85 |
// Update current position (bytes) adding the read characters.
|
86 |
CharBuffer charBuffer = CharBuffer.wrap(cbuf, off, len); |
87 |
ByteBuffer byteBuffer = this.charset.encode(charBuffer); |
88 |
this.currentPosition += byteBuffer.limit();
|
89 |
} |
90 |
return n;
|
91 |
} |
92 |
|
93 |
protected void createReader() { |
94 |
this.reader = Channels.newReader(this.raf.getChannel(), this.charset.name()); |
95 |
} |
96 |
|
97 |
// protected InputStream is;
|
98 |
// protected void createReader() {
|
99 |
// if( this.is==null ) {
|
100 |
// this.is = new InputStream() {
|
101 |
// @Override
|
102 |
// public int read() throws IOException {
|
103 |
// return raf.read();
|
104 |
// }
|
105 |
// };
|
106 |
// }
|
107 |
// this.reader = new InputStreamReader(this.is, charset);
|
108 |
// }
|
109 |
@Override
|
110 |
public void close() throws IOException { |
111 |
// IOUtils.closeQuietly(this.is);
|
112 |
IOUtils.closeQuietly(this.reader);
|
113 |
IOUtils.closeQuietly(this.raf);
|
114 |
} |
115 |
|
116 |
public long getFilePointer() throws IOException { |
117 |
return this.raf.getFilePointer(); |
118 |
} |
119 |
|
120 |
public long getCurrentPosition() { |
121 |
return this.currentPosition; |
122 |
} |
123 |
|
124 |
public void rewind() throws IOException { |
125 |
this.raf.seek(0); |
126 |
this.reader = null; |
127 |
this.currentPosition = 0; |
128 |
} |
129 |
|
130 |
public void seek(long position) throws IOException { |
131 |
this.raf.seek(position);
|
132 |
this.reader = null; |
133 |
this.currentPosition = position;
|
134 |
} |
135 |
|
136 |
public String readLine() throws IOException { |
137 |
StringBuilder buffer = new StringBuilder(); |
138 |
int c = -1; |
139 |
boolean eol = false; |
140 |
|
141 |
while (!eol) {
|
142 |
switch (c = this.read()) { |
143 |
case -1: |
144 |
case '\n': |
145 |
eol = true;
|
146 |
break;
|
147 |
case '\r': |
148 |
eol = true;
|
149 |
long cur = raf.getFilePointer();
|
150 |
if ((raf.read()) != '\n') { |
151 |
raf.seek(cur); |
152 |
} |
153 |
break;
|
154 |
default:
|
155 |
buffer.append((char) c);
|
156 |
break;
|
157 |
} |
158 |
} |
159 |
if ((c == -1) && (buffer.length() == 0)) { |
160 |
return null; |
161 |
} |
162 |
return buffer.toString();
|
163 |
} |
164 |
|
165 |
public long countLines(Predicate<String> filter, SimpleTaskStatus status) throws IOException { |
166 |
if (raf.length() == 0) { |
167 |
return 0; |
168 |
} |
169 |
long savedpos = this.getCurrentPosition(); |
170 |
long count = -1; |
171 |
if (status != null) { |
172 |
I18nManager i18n = ToolsLocator.getI18nManager(); |
173 |
status.message(i18n.getTranslation("_Calculating_number_of_lines"));
|
174 |
status.setIndeterminate(); |
175 |
} |
176 |
BufferedReader breader = new BufferedReader(this, MAX_BUFFER_FOR_LINE); |
177 |
try {
|
178 |
String line;
|
179 |
count = 0;
|
180 |
while ((line = breader.readLine()) != null) { |
181 |
if (status != null) { |
182 |
if (status.isCancellationRequested()) {
|
183 |
return -1; |
184 |
} |
185 |
// status.incrementCurrentValue();
|
186 |
if((count % 1000) == 0){ |
187 |
status.setCurValue(count); |
188 |
} |
189 |
} |
190 |
if (filter.test(line)) {
|
191 |
continue;
|
192 |
} |
193 |
count++; |
194 |
} |
195 |
if (status != null) { |
196 |
status.setCurValue(count); |
197 |
status.message("");
|
198 |
status.setIndeterminate(); |
199 |
} |
200 |
} finally {
|
201 |
this.seek(savedpos);
|
202 |
} |
203 |
return count;
|
204 |
} |
205 |
|
206 |
public boolean isRecomemendedTheRecreationOfTheLinesIndex(File index) { |
207 |
RandomAccessFileIndex line_idx = null;
|
208 |
try {
|
209 |
if (this.lastModified > 0 && this.lastModified > index.lastModified()) { |
210 |
return true; |
211 |
} |
212 |
line_idx = new RandomAccessFileIndex();
|
213 |
line_idx.open(index); |
214 |
if (this.raf.length() != line_idx.getHeader(INDEX_HEADER_FILESIZE)) { |
215 |
return true; |
216 |
} |
217 |
long creationCost = line_idx.getHeader(INDEX_HEADER_FILESIZE);
|
218 |
if (creationCost < 2000) { // < 2 sec. |
219 |
return true; |
220 |
} |
221 |
if (line_idx.get(-1) == 0) { |
222 |
// if last index == 0, index is corrupt
|
223 |
return true; |
224 |
} |
225 |
// FIXME: isValidIndexOfLines not full implemented
|
226 |
// Podria comprobarse que una muestra de 4 o 5 bloques de datos
|
227 |
// repartidas por el fichero tengan el checksum correcto
|
228 |
return false; |
229 |
} catch (IOException ex) { |
230 |
return true; |
231 |
} finally {
|
232 |
IOUtils.closeQuietly(line_idx); |
233 |
} |
234 |
} |
235 |
|
236 |
public RandomAccessFileIndex createOrOpenIndexOfLines(File index, Predicate<String> filter, SimpleTaskStatus status) throws IOException { |
237 |
return this.createOrOpenIndexOfLines(index, false, filter, status); |
238 |
} |
239 |
|
240 |
public RandomAccessFileIndex createOrOpenIndexOfLines(File index, boolean safe, Predicate<String> filter, SimpleTaskStatus status) throws IOException { |
241 |
return createOrOpenIndexOfLines(index, safe, filter, status, null); |
242 |
} |
243 |
|
244 |
public RandomAccessFileIndex createOrOpenIndexOfLines(File index, boolean safe, Predicate<String> filter, SimpleTaskStatus status, Function<BufferedReader,Integer> numberOfLines) throws IOException { |
245 |
if (this.isRecomemendedTheRecreationOfTheLinesIndex(index)) { |
246 |
return this.createIndexOfLines(index, safe, filter, status, numberOfLines); |
247 |
} |
248 |
return new RandomAccessFileIndex(index); |
249 |
} |
250 |
|
251 |
public RandomAccessFileIndex createIndexOfLines(File index, Predicate<String> filter, SimpleTaskStatus status) throws IOException { |
252 |
return this.createIndexOfLines(index, false, filter, status); |
253 |
} |
254 |
|
255 |
public RandomAccessFileIndex createIndexOfLines(File index, boolean safe, Predicate<String> filter, SimpleTaskStatus status) throws IOException { |
256 |
return createIndexOfLines(index, safe, filter, status, null); |
257 |
} |
258 |
|
259 |
public RandomAccessFileIndex createIndexOfLines(File index, boolean safe, Predicate<String> filter, SimpleTaskStatus status, Function<BufferedReader,Integer> numberOfLines) throws IOException { |
260 |
long countLines = this.countLines(filter, status); |
261 |
if (countLines < 1) { |
262 |
return null; |
263 |
} |
264 |
RandomAccessFileIndex line_idx = new RandomAccessFileIndex();
|
265 |
line_idx.create(index, countLines); |
266 |
|
267 |
long savedpos = this.getCurrentPosition(); |
268 |
try {
|
269 |
if (status != null) { |
270 |
I18nManager i18n = ToolsLocator.getI18nManager(); |
271 |
status.push(); |
272 |
status.message(i18n.getTranslation("_Creating_the_index_of_the_lines"));
|
273 |
status.setRangeOfValues(0, line_idx.size64());
|
274 |
status.setCurValue(0);
|
275 |
} |
276 |
long t1 = System.currentTimeMillis(); |
277 |
String line = null; |
278 |
int lineno = 0; |
279 |
long position = 0; |
280 |
// line_idx.set(lineno++, position);
|
281 |
if (safe) {
|
282 |
// Don't use buffered reader, slow and safe calculate position
|
283 |
int x = (int) (countLines / 100); |
284 |
while (lineno < countLines) { //true ) { |
285 |
line = this.readLine();
|
286 |
if (line == null) { |
287 |
break;
|
288 |
} |
289 |
if (filter.test(line)) {
|
290 |
continue;
|
291 |
} |
292 |
line_idx.set(lineno++, position); |
293 |
if (status != null) { |
294 |
if (status.isCancellationRequested()) {
|
295 |
status.cancel(); |
296 |
return null; |
297 |
} |
298 |
// status.incrementCurrentValue();
|
299 |
if((lineno % x) == 0){ |
300 |
status.setCurValue(lineno); |
301 |
} |
302 |
} |
303 |
position = this.getCurrentPosition();
|
304 |
// line_idx.set(lineno++, position);
|
305 |
} |
306 |
status.setCurValue(lineno); |
307 |
} else {
|
308 |
// Use buffered reader, fast and unsafe calculate position.
|
309 |
StringBuilder builder = new StringBuilder(); |
310 |
MyBufferedReader breader = new MyBufferedReader(this, MAX_BUFFER_FOR_LINE); |
311 |
while (lineno < countLines) {
|
312 |
this.seek(position);
|
313 |
breader.clean(); |
314 |
if(numberOfLines == null){ |
315 |
line = breader.readLine(); |
316 |
} else {
|
317 |
breader.mark(MAX_BUFFER_FOR_LINE); |
318 |
Integer nextLine = numberOfLines.apply(breader);
|
319 |
breader.reset(); |
320 |
builder.setLength(0);
|
321 |
for (int i = 0; i < nextLine; i++) { |
322 |
String l = breader.readLine();
|
323 |
if(l != null){ |
324 |
builder.append(l); |
325 |
} else {
|
326 |
break;
|
327 |
} |
328 |
} |
329 |
line = StringUtils.defaultIfBlank(builder.toString(), null);
|
330 |
} |
331 |
if (line == null) { |
332 |
break;
|
333 |
} |
334 |
if (filter.test(line)) {
|
335 |
continue;
|
336 |
} |
337 |
line_idx.set(lineno++, position); |
338 |
if (status != null) { |
339 |
if (status.isCancellationRequested()) {
|
340 |
status.cancel(); |
341 |
return null; |
342 |
} |
343 |
status.incrementCurrentValue(); |
344 |
} |
345 |
CharBuffer charBuffer = null; |
346 |
// ? Y si hay un \r\n ?
|
347 |
if(breader.isSkipLf()){
|
348 |
charBuffer = CharBuffer.wrap(line + "\r\n"); |
349 |
} else {
|
350 |
charBuffer = CharBuffer.wrap(line + "\n"); |
351 |
} |
352 |
ByteBuffer byteBuffer = this.charset.encode(charBuffer); |
353 |
position += byteBuffer.limit(); |
354 |
|
355 |
// line_idx.set(lineno++, position);
|
356 |
} |
357 |
} |
358 |
long t2 = System.currentTimeMillis(); |
359 |
line_idx.setNumElements(lineno); |
360 |
line_idx.setHeader(INDEX_HEADER_FILESIZE, this.raf.length());
|
361 |
line_idx.setHeader(INDEX_HEADER_INDEXCREATIONCOST, t2 - t1); |
362 |
if (status != null) { |
363 |
status.message("");
|
364 |
status.pop(); |
365 |
} |
366 |
return line_idx;
|
367 |
} finally {
|
368 |
this.seek(savedpos);
|
369 |
} |
370 |
} |
371 |
|
372 |
public long getLastLinesIndexCreationCost(RandomAccessFileIndex index) { |
373 |
return index.getHeader(INDEX_HEADER_INDEXCREATIONCOST);
|
374 |
} |
375 |
|
376 |
public static void main(String[] args) throws Exception { |
377 |
new DefaultLibrariesInitializer().fullInitialize();
|
378 |
|
379 |
String fname;
|
380 |
fname = "/home/jjdelcerro/Descargas/test/origen_coordenadas.csv";
|
381 |
// fname = "/home/jjdelcerro/Descargas/test/esp_poblaciones.csv";
|
382 |
// fname = "/home/jjdelcerro/Descargas/test/esp_provincias.csv";
|
383 |
// fname = "/home/jjdelcerro/Descargas/test/sigpac.csv";
|
384 |
|
385 |
File data_file = new File(fname); |
386 |
File idx_file = new File(FilenameUtils.removeExtension(data_file.getAbsolutePath()) + ".idx"); |
387 |
|
388 |
final TaskStatusManager taskStatusManager = ToolsLocator.getTaskStatusManager();
|
389 |
taskStatusManager.addObserver((Observable observable, Object notification) -> { |
390 |
TaskStatus status = taskStatusManager.getRunningTaskStatusMostRecent(); |
391 |
// System.out.print("\033[?25l\r");
|
392 |
// if( status!=null && status.isRunning() ) {
|
393 |
// System.out.print("\033[?25l\r");
|
394 |
// System.out.print(status.getTitle()+ " - " + status.getLabel());
|
395 |
// System.out.print("\033[K\033[?12l\033[?25h");
|
396 |
// }
|
397 |
// System.out.flush();
|
398 |
}); |
399 |
SimpleTaskStatus status = taskStatusManager.createDefaultSimpleTaskStatus(data_file.getName()); |
400 |
status.add(); |
401 |
|
402 |
RandomAccessFileReader reader = new RandomAccessFileReader(data_file, "UTF-8"); |
403 |
System.out.println("Index '" + idx_file.getName() + "', is creation recomended: " + reader.isRecomemendedTheRecreationOfTheLinesIndex(idx_file)); |
404 |
RandomAccessFileIndex lines_idx = reader.createOrOpenIndexOfLines(idx_file, FILTER_NONE, status); |
405 |
|
406 |
for (int linenumber = 0; linenumber < lines_idx.size(); linenumber++) { |
407 |
long lineoffset = lines_idx.get(linenumber);
|
408 |
reader.seek(lineoffset); |
409 |
MyBufferedReader breader = new MyBufferedReader(reader, MAX_BUFFER_FOR_LINE);
|
410 |
String line = breader.readLine();
|
411 |
if (linenumber < 100) { |
412 |
System.out.println(String.format("%6d/%d: %s", lineoffset, linenumber, line)); |
413 |
} else if (linenumber == 100) { |
414 |
System.out.println("More records..."); |
415 |
} |
416 |
} |
417 |
|
418 |
System.out.println("------------------------------------"); |
419 |
|
420 |
for (int linenumber = lines_idx.size() - 1; linenumber >= 0; linenumber--) { |
421 |
long lineoffset = lines_idx.get(linenumber);
|
422 |
reader.seek(lineoffset); |
423 |
MyBufferedReader breader = new MyBufferedReader(reader, MAX_BUFFER_FOR_LINE);
|
424 |
String line = breader.readLine();
|
425 |
if (linenumber < 100) { |
426 |
System.out.println(String.format("%6d/%d: %s", lineoffset, linenumber, line)); |
427 |
} else if (linenumber == 100) { |
428 |
System.out.println("More records..."); |
429 |
} |
430 |
} |
431 |
|
432 |
} |
433 |
|
434 |
/*
|
435 |
Copy of java's BufferedReader adding clean and isSkipLf methods
|
436 |
*/
|
437 |
public static class MyBufferedReader extends BufferedReader { |
438 |
|
439 |
private Reader in; |
440 |
|
441 |
private char cb[]; |
442 |
private int nChars, nextChar; |
443 |
|
444 |
private static final int INVALIDATED = -2; |
445 |
private static final int UNMARKED = -1; |
446 |
private int markedChar = UNMARKED; |
447 |
private int readAheadLimit = 0; |
448 |
/* Valid only when markedChar > 0 */
|
449 |
|
450 |
/**
|
451 |
* If the next character is a line feed, skip it
|
452 |
*/
|
453 |
private boolean skipLF = false; |
454 |
|
455 |
/**
|
456 |
* The skipLF flag when the mark was set
|
457 |
*/
|
458 |
private boolean markedSkipLF = false; |
459 |
|
460 |
private static int defaultCharBufferSize = 8192; |
461 |
private static int defaultExpectedLineLength = 80; |
462 |
|
463 |
/**
|
464 |
* Creates a buffering character-input stream that uses an input buffer
|
465 |
* of the specified size.
|
466 |
*
|
467 |
* @param in A Reader
|
468 |
* @param sz Input-buffer size
|
469 |
*
|
470 |
* @exception IllegalArgumentException If {@code sz <= 0}
|
471 |
*/
|
472 |
public MyBufferedReader(Reader in, int sz) { |
473 |
super(in);
|
474 |
if (sz <= 0) { |
475 |
throw new IllegalArgumentException("Buffer size <= 0"); |
476 |
} |
477 |
this.in = in;
|
478 |
cb = new char[sz]; |
479 |
nextChar = nChars = 0;
|
480 |
} |
481 |
|
482 |
/**
|
483 |
* Creates a buffering character-input stream that uses a default-sized
|
484 |
* input buffer.
|
485 |
*
|
486 |
* @param in A Reader
|
487 |
*/
|
488 |
public MyBufferedReader(Reader in) { |
489 |
this(in, defaultCharBufferSize);
|
490 |
} |
491 |
|
492 |
/**
|
493 |
* Checks to make sure that the stream has not been closed
|
494 |
*/
|
495 |
private void ensureOpen() throws IOException { |
496 |
if (in == null) { |
497 |
throw new IOException("Stream closed"); |
498 |
} |
499 |
} |
500 |
|
501 |
/**
|
502 |
* Fills the input buffer, taking the mark into account if it is valid.
|
503 |
*/
|
504 |
private void fill() throws IOException { |
505 |
int dst;
|
506 |
if (markedChar <= UNMARKED) {
|
507 |
/* No mark */
|
508 |
dst = 0;
|
509 |
} else {
|
510 |
/* Marked */
|
511 |
int delta = nextChar - markedChar;
|
512 |
if (delta >= readAheadLimit) {
|
513 |
/* Gone past read-ahead limit: Invalidate mark */
|
514 |
markedChar = INVALIDATED; |
515 |
readAheadLimit = 0;
|
516 |
dst = 0;
|
517 |
} else {
|
518 |
if (readAheadLimit <= cb.length) {
|
519 |
/* Shuffle in the current buffer */
|
520 |
System.arraycopy(cb, markedChar, cb, 0, delta); |
521 |
markedChar = 0;
|
522 |
dst = delta; |
523 |
} else {
|
524 |
/* Reallocate buffer to accommodate read-ahead limit */
|
525 |
char ncb[] = new char[readAheadLimit]; |
526 |
System.arraycopy(cb, markedChar, ncb, 0, delta); |
527 |
cb = ncb; |
528 |
markedChar = 0;
|
529 |
dst = delta; |
530 |
} |
531 |
nextChar = nChars = delta; |
532 |
} |
533 |
} |
534 |
|
535 |
int n;
|
536 |
do {
|
537 |
n = in.read(cb, dst, cb.length - dst); |
538 |
} while (n == 0); |
539 |
if (n > 0) { |
540 |
nChars = dst + n; |
541 |
nextChar = dst; |
542 |
} |
543 |
} |
544 |
|
545 |
/**
|
546 |
* Reads a single character.
|
547 |
*
|
548 |
* @return The character read, as an integer in the range 0 to 65535
|
549 |
* (<tt>0x00-0xffff</tt>), or -1 if the end of the stream has been
|
550 |
* reached
|
551 |
* @exception IOException If an I/O error occurs
|
552 |
*/
|
553 |
@Override
|
554 |
public int read() throws IOException { |
555 |
synchronized (lock) {
|
556 |
ensureOpen(); |
557 |
for (;;) {
|
558 |
if (nextChar >= nChars) {
|
559 |
fill(); |
560 |
if (nextChar >= nChars) {
|
561 |
return -1; |
562 |
} |
563 |
} |
564 |
if (skipLF) {
|
565 |
skipLF = false;
|
566 |
if (cb[nextChar] == '\n') { |
567 |
nextChar++; |
568 |
continue;
|
569 |
} |
570 |
} |
571 |
return cb[nextChar++];
|
572 |
} |
573 |
} |
574 |
} |
575 |
|
576 |
/**
|
577 |
* Reads characters into a portion of an array, reading from the
|
578 |
* underlying stream if necessary.
|
579 |
*/
|
580 |
private int read1(char[] cbuf, int off, int len) throws IOException { |
581 |
if (nextChar >= nChars) {
|
582 |
/* If the requested length is at least as large as the buffer, and
|
583 |
if there is no mark/reset activity, and if line feeds are not
|
584 |
being skipped, do not bother to copy the characters into the
|
585 |
local buffer. In this way buffered streams will cascade
|
586 |
harmlessly. */
|
587 |
if (len >= cb.length && markedChar <= UNMARKED && !skipLF) {
|
588 |
return in.read(cbuf, off, len);
|
589 |
} |
590 |
fill(); |
591 |
} |
592 |
if (nextChar >= nChars) {
|
593 |
return -1; |
594 |
} |
595 |
if (skipLF) {
|
596 |
skipLF = false;
|
597 |
if (cb[nextChar] == '\n') { |
598 |
nextChar++; |
599 |
if (nextChar >= nChars) {
|
600 |
fill(); |
601 |
} |
602 |
if (nextChar >= nChars) {
|
603 |
return -1; |
604 |
} |
605 |
} |
606 |
} |
607 |
int n = Math.min(len, nChars - nextChar); |
608 |
System.arraycopy(cb, nextChar, cbuf, off, n);
|
609 |
nextChar += n; |
610 |
return n;
|
611 |
} |
612 |
|
613 |
/**
|
614 |
* Reads characters into a portion of an array.
|
615 |
*
|
616 |
* <p>
|
617 |
* This method implements the general contract of the corresponding
|
618 |
* <code>{@link Reader#read(char[], int, int) read}</code> method of the
|
619 |
* <code>{@link Reader}</code> class. As an additional convenience, it
|
620 |
* attempts to read as many characters as possible by repeatedly
|
621 |
* invoking the <code>read</code> method of the underlying stream. This
|
622 |
* iterated <code>read</code> continues until one of the following
|
623 |
* conditions becomes true: <ul>
|
624 |
*
|
625 |
* <li> The specified number of characters have been read,
|
626 |
*
|
627 |
* <li> The <code>read</code> method of the underlying stream returns
|
628 |
* <code>-1</code>, indicating end-of-file, or
|
629 |
*
|
630 |
* <li> The <code>ready</code> method of the underlying stream returns
|
631 |
* <code>false</code>, indicating that further input requests would
|
632 |
* block.
|
633 |
*
|
634 |
* </ul> If the first <code>read</code> on the underlying stream returns
|
635 |
* <code>-1</code> to indicate end-of-file then this method returns
|
636 |
* <code>-1</code>. Otherwise this method returns the number of
|
637 |
* characters actually read.
|
638 |
*
|
639 |
* <p>
|
640 |
* Subclasses of this class are encouraged, but not required, to attempt
|
641 |
* to read as many characters as possible in the same fashion.
|
642 |
*
|
643 |
* <p>
|
644 |
* Ordinarily this method takes characters from this stream's character
|
645 |
* buffer, filling it from the underlying stream as necessary. If,
|
646 |
* however, the buffer is empty, the mark is not valid, and the
|
647 |
* requested length is at least as large as the buffer, then this method
|
648 |
* will read characters directly from the underlying stream into the
|
649 |
* given array. Thus redundant <code>BufferedReader</code>s will not
|
650 |
* copy data unnecessarily.
|
651 |
*
|
652 |
* @param cbuf Destination buffer
|
653 |
* @param off Offset at which to start storing characters
|
654 |
* @param len Maximum number of characters to read
|
655 |
*
|
656 |
* @return The number of characters read, or -1 if the end of the stream
|
657 |
* has been reached
|
658 |
*
|
659 |
* @exception IOException If an I/O error occurs
|
660 |
*/
|
661 |
@Override
|
662 |
public int read(char cbuf[], int off, int len) throws IOException { |
663 |
synchronized (lock) {
|
664 |
ensureOpen(); |
665 |
if ((off < 0) || (off > cbuf.length) || (len < 0) |
666 |
|| ((off + len) > cbuf.length) || ((off + len) < 0)) {
|
667 |
throw new IndexOutOfBoundsException(); |
668 |
} else if (len == 0) { |
669 |
return 0; |
670 |
} |
671 |
|
672 |
int n = read1(cbuf, off, len);
|
673 |
if (n <= 0) { |
674 |
return n;
|
675 |
} |
676 |
while ((n < len) && in.ready()) {
|
677 |
int n1 = read1(cbuf, off + n, len - n);
|
678 |
if (n1 <= 0) { |
679 |
break;
|
680 |
} |
681 |
n += n1; |
682 |
} |
683 |
return n;
|
684 |
} |
685 |
} |
686 |
|
687 |
/**
|
688 |
* Reads a line of text. A line is considered to be terminated by any
|
689 |
* one of a line feed ('\n'), a carriage return ('\r'), or a carriage
|
690 |
* return followed immediately by a linefeed.
|
691 |
*
|
692 |
* @param ignoreLF If true, the next '\n' will be skipped
|
693 |
*
|
694 |
* @return A String containing the contents of the line, not including
|
695 |
* any line-termination characters, or null if the end of the stream has
|
696 |
* been reached
|
697 |
*
|
698 |
* @see java.io.LineNumberReader#readLine()
|
699 |
*
|
700 |
* @exception IOException If an I/O error occurs
|
701 |
*/
|
702 |
String readLine(boolean ignoreLF) throws IOException { |
703 |
StringBuilder s = null; |
704 |
int startChar;
|
705 |
|
706 |
synchronized (lock) {
|
707 |
ensureOpen(); |
708 |
boolean omitLF = ignoreLF || skipLF;
|
709 |
|
710 |
bufferLoop: |
711 |
for (;;) {
|
712 |
|
713 |
if (nextChar >= nChars) {
|
714 |
fill(); |
715 |
} |
716 |
if (nextChar >= nChars) {
|
717 |
/* EOF */
|
718 |
if (s != null && s.length() > 0) { |
719 |
return s.toString();
|
720 |
} else {
|
721 |
return null; |
722 |
} |
723 |
} |
724 |
boolean eol = false; |
725 |
char c = 0; |
726 |
int i;
|
727 |
|
728 |
/* Skip a leftover '\n', if necessary */
|
729 |
if (omitLF && (cb[nextChar] == '\n')) { |
730 |
nextChar++; |
731 |
} |
732 |
skipLF = false;
|
733 |
omitLF = false;
|
734 |
|
735 |
charLoop: |
736 |
for (i = nextChar; i < nChars; i++) {
|
737 |
c = cb[i]; |
738 |
if ((c == '\n') || (c == '\r')) { |
739 |
eol = true;
|
740 |
break charLoop;
|
741 |
} |
742 |
} |
743 |
|
744 |
startChar = nextChar; |
745 |
nextChar = i; |
746 |
|
747 |
if (eol) {
|
748 |
String str;
|
749 |
if (s == null) { |
750 |
str = new String(cb, startChar, i - startChar); |
751 |
} else {
|
752 |
s.append(cb, startChar, i - startChar); |
753 |
str = s.toString(); |
754 |
} |
755 |
nextChar++; |
756 |
if (c == '\r') { |
757 |
skipLF = true;
|
758 |
} |
759 |
return str;
|
760 |
} |
761 |
|
762 |
if (s == null) { |
763 |
s = new StringBuilder(defaultExpectedLineLength); |
764 |
} |
765 |
s.append(cb, startChar, i - startChar); |
766 |
} |
767 |
} |
768 |
} |
769 |
|
770 |
/**
|
771 |
* Reads a line of text. A line is considered to be terminated by any
|
772 |
* one of a line feed ('\n'), a carriage return ('\r'), or a carriage
|
773 |
* return followed immediately by a linefeed.
|
774 |
*
|
775 |
* @return A String containing the contents of the line, not including
|
776 |
* any line-termination characters, or null if the end of the stream has
|
777 |
* been reached
|
778 |
*
|
779 |
* @exception IOException If an I/O error occurs
|
780 |
*
|
781 |
* @see java.nio.file.Files#readAllLines
|
782 |
*/
|
783 |
@Override
|
784 |
public String readLine() throws IOException { |
785 |
return readLine(false); |
786 |
} |
787 |
|
788 |
/**
|
789 |
* Skips characters.
|
790 |
*
|
791 |
* @param n The number of characters to skip
|
792 |
*
|
793 |
* @return The number of characters actually skipped
|
794 |
*
|
795 |
* @exception IllegalArgumentException If <code>n</code> is negative.
|
796 |
* @exception IOException If an I/O error occurs
|
797 |
*/
|
798 |
@Override
|
799 |
public long skip(long n) throws IOException { |
800 |
if (n < 0L) { |
801 |
throw new IllegalArgumentException("skip value is negative"); |
802 |
} |
803 |
synchronized (lock) {
|
804 |
ensureOpen(); |
805 |
long r = n;
|
806 |
while (r > 0) { |
807 |
if (nextChar >= nChars) {
|
808 |
fill(); |
809 |
} |
810 |
if (nextChar >= nChars) /* EOF */ { |
811 |
break;
|
812 |
} |
813 |
if (skipLF) {
|
814 |
skipLF = false;
|
815 |
if (cb[nextChar] == '\n') { |
816 |
nextChar++; |
817 |
} |
818 |
} |
819 |
long d = nChars - nextChar;
|
820 |
if (r <= d) {
|
821 |
nextChar += r; |
822 |
r = 0;
|
823 |
break;
|
824 |
} else {
|
825 |
r -= d; |
826 |
nextChar = nChars; |
827 |
} |
828 |
} |
829 |
return n - r;
|
830 |
} |
831 |
} |
832 |
|
833 |
/**
|
834 |
* Tells whether this stream is ready to be read. A buffered character
|
835 |
* stream is ready if the buffer is not empty, or if the underlying
|
836 |
* character stream is ready.
|
837 |
*
|
838 |
* @exception IOException If an I/O error occurs
|
839 |
*/
|
840 |
@Override
|
841 |
public boolean ready() throws IOException { |
842 |
synchronized (lock) {
|
843 |
ensureOpen(); |
844 |
|
845 |
/*
|
846 |
* If newline needs to be skipped and the next char to be read
|
847 |
* is a newline character, then just skip it right away.
|
848 |
*/
|
849 |
if (skipLF) {
|
850 |
/* Note that in.ready() will return true if and only if the next
|
851 |
* read on the stream will not block.
|
852 |
*/
|
853 |
if (nextChar >= nChars && in.ready()) {
|
854 |
fill(); |
855 |
} |
856 |
if (nextChar < nChars) {
|
857 |
if (cb[nextChar] == '\n') { |
858 |
nextChar++; |
859 |
} |
860 |
skipLF = false;
|
861 |
} |
862 |
} |
863 |
return (nextChar < nChars) || in.ready();
|
864 |
} |
865 |
} |
866 |
|
867 |
/**
|
868 |
* Tells whether this stream supports the mark() operation, which it
|
869 |
* does.
|
870 |
*/
|
871 |
@Override
|
872 |
public boolean markSupported() { |
873 |
return true; |
874 |
} |
875 |
|
876 |
/**
|
877 |
* Marks the present position in the stream. Subsequent calls to reset()
|
878 |
* will attempt to reposition the stream to this point.
|
879 |
*
|
880 |
* @param readAheadLimit Limit on the number of characters that may be
|
881 |
* read while still preserving the mark. An attempt to reset the stream
|
882 |
* after reading characters up to this limit or beyond may fail. A limit
|
883 |
* value larger than the size of the input buffer will cause a new
|
884 |
* buffer to be allocated whose size is no smaller than limit. Therefore
|
885 |
* large values should be used with care.
|
886 |
*
|
887 |
* @exception IllegalArgumentException If {@code readAheadLimit < 0}
|
888 |
* @exception IOException If an I/O error occurs
|
889 |
*/
|
890 |
@Override
|
891 |
public void mark(int readAheadLimit) throws IOException { |
892 |
if (readAheadLimit < 0) { |
893 |
throw new IllegalArgumentException("Read-ahead limit < 0"); |
894 |
} |
895 |
synchronized (lock) {
|
896 |
ensureOpen(); |
897 |
this.readAheadLimit = readAheadLimit;
|
898 |
markedChar = nextChar; |
899 |
markedSkipLF = skipLF; |
900 |
} |
901 |
} |
902 |
|
903 |
/**
|
904 |
* Resets the stream to the most recent mark.
|
905 |
*
|
906 |
* @exception IOException If the stream has never been marked, or if the
|
907 |
* mark has been invalidated
|
908 |
*/
|
909 |
@Override
|
910 |
public void reset() throws IOException { |
911 |
synchronized (lock) {
|
912 |
ensureOpen(); |
913 |
if (markedChar < 0) { |
914 |
throw new IOException((markedChar == INVALIDATED) |
915 |
? "Mark invalid"
|
916 |
: "Stream not marked");
|
917 |
} |
918 |
nextChar = markedChar; |
919 |
skipLF = markedSkipLF; |
920 |
} |
921 |
} |
922 |
|
923 |
@Override
|
924 |
public void close() throws IOException { |
925 |
synchronized (lock) {
|
926 |
if (in == null) { |
927 |
return;
|
928 |
} |
929 |
try {
|
930 |
in.close(); |
931 |
} finally {
|
932 |
in = null;
|
933 |
cb = null;
|
934 |
} |
935 |
} |
936 |
} |
937 |
|
938 |
/**
|
939 |
* Returns a {@code Stream}, the elements of which are lines read from
|
940 |
* this {@code BufferedReader}. The {@link Stream} is lazily populated,
|
941 |
* i.e., read only occurs during the
|
942 |
* <a href="../util/stream/package-summary.html#StreamOps">terminal
|
943 |
* stream operation</a>.
|
944 |
*
|
945 |
* <p>
|
946 |
* The reader must not be operated on during the execution of the
|
947 |
* terminal stream operation. Otherwise, the result of the terminal
|
948 |
* stream operation is undefined.
|
949 |
*
|
950 |
* <p>
|
951 |
* After execution of the terminal stream operation there are no
|
952 |
* guarantees that the reader will be at a specific position from which
|
953 |
* to read the next character or line.
|
954 |
*
|
955 |
* <p>
|
956 |
* If an {@link IOException} is thrown when accessing the underlying
|
957 |
* {@code BufferedReader}, it is wrapped in an {@link
|
958 |
* UncheckedIOException} which will be thrown from the {@code Stream}
|
959 |
* method that caused the read to take place. This method will return a
|
960 |
* Stream if invoked on a BufferedReader that is closed. Any operation
|
961 |
* on that stream that requires reading from the BufferedReader after it
|
962 |
* is closed, will cause an UncheckedIOException to be thrown.
|
963 |
*
|
964 |
* @return a {@code Stream<String>} providing the lines of text
|
965 |
* described by this {@code BufferedReader}
|
966 |
*
|
967 |
* @since 1.8
|
968 |
*/
|
969 |
@Override
|
970 |
public Stream<String> lines() { |
971 |
Iterator<String> iter = new Iterator<String>() { |
972 |
String nextLine = null; |
973 |
|
974 |
@Override
|
975 |
public boolean hasNext() { |
976 |
if (nextLine != null) { |
977 |
return true; |
978 |
} else {
|
979 |
try {
|
980 |
nextLine = readLine(); |
981 |
return (nextLine != null); |
982 |
} catch (IOException e) { |
983 |
throw new UncheckedIOException(e); |
984 |
} |
985 |
} |
986 |
} |
987 |
|
988 |
@Override
|
989 |
public String next() { |
990 |
if (nextLine != null || hasNext()) { |
991 |
String line = nextLine;
|
992 |
nextLine = null;
|
993 |
return line;
|
994 |
} else {
|
995 |
throw new NoSuchElementException(); |
996 |
} |
997 |
} |
998 |
}; |
999 |
return StreamSupport.stream(Spliterators.spliteratorUnknownSize(
|
1000 |
iter, Spliterator.ORDERED | Spliterator.NONNULL), false);
|
1001 |
} |
1002 |
|
1003 |
public boolean isSkipLf() { |
1004 |
return this.skipLF; |
1005 |
} |
1006 |
|
1007 |
public void clean() { |
1008 |
nextChar = nChars = 0;
|
1009 |
markedChar = UNMARKED; |
1010 |
readAheadLimit = 0;
|
1011 |
skipLF = false;
|
1012 |
markedSkipLF = false;
|
1013 |
|
1014 |
} |
1015 |
} |
1016 |
|
1017 |
} |