Skip to content

Commit bd565ee

Browse files
brandstaetterbrandstaetter
committed
implemented a new Fastq Reader/Writer for the new Illumina Fastq Formatting (according to Wikipedia it's version 1.8, new this month
extended the Fastq class to be able to generate DNASequence representations with the Quality (as Phred Numbers) added as Feature extended the Fastq class to have a contructor that accepts a DNASequence some test cases code formatting git-svn-id: http://code.open-bio.org/repos/biojava/biojava-live/trunk@9545 7c6358e6-4a41-0410-a743-a5b2a554c398
1 parent d3dc5ab commit bd565ee

28 files changed

Lines changed: 1241 additions & 1317 deletions

‎biojava3-sequencing/pom.xml‎

Lines changed: 31 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -1,19 +1,35 @@
11
<?xml version="1.0" encoding="UTF-8"?>
22
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/maven-v4_0_0.xsd">
3-
<parent>
4-
<artifactId>biojava</artifactId>
5-
<groupId>org.biojava</groupId>
6-
<version>3.0.3-SNAPSHOT</version>
7-
</parent>
8-
<modelVersion>4.0.0</modelVersion>
9-
<artifactId>biojava3-sequencing</artifactId>
10-
<name>biojava3-sequencing</name>
11-
<dependencies>
12-
<dependency>
13-
<groupId>junit</groupId>
14-
<artifactId>junit</artifactId>
15-
<scope>test</scope>
16-
</dependency>
17-
</dependencies>
3+
<parent>
4+
<artifactId>biojava</artifactId>
5+
<groupId>org.biojava</groupId>
6+
<version>3.0.3-SNAPSHOT</version>
7+
</parent>
8+
<modelVersion>4.0.0</modelVersion>
9+
<artifactId>biojava3-sequencing</artifactId>
10+
<name>biojava3-sequencing</name>
11+
<distributionManagement>
12+
<downloadUrl>http://www.biojava.org/download/maven/</downloadUrl>
13+
<repository>
14+
<uniqueVersion>false</uniqueVersion>
15+
<id>biojava-maven-repo</id>
16+
<name>BioJava Maven Repository</name>
17+
<url>scp://portal.open-bio.org/home/websites/biojava.org/html/static/download/maven/</url>
18+
<layout>default</layout>
19+
</repository>
20+
</distributionManagement>
21+
<dependencies>
22+
<dependency>
23+
<groupId>junit</groupId>
24+
<artifactId>junit</artifactId>
25+
<scope>test</scope>
26+
</dependency>
27+
<dependency>
28+
<groupId>org.biojava</groupId>
29+
<artifactId>biojava3-core</artifactId>
30+
<version>3.0.3-SNAPSHOT</version>
31+
<scope>compile</scope>
32+
</dependency>
33+
</dependencies>
1834
</project>
1935

‎biojava3-sequencing/src/main/java/org/biojava3/sequencing/io/fastq/AbstractFastqReader.java‎

Lines changed: 102 additions & 139 deletions
Original file line numberDiff line numberDiff line change
@@ -20,15 +20,8 @@
2020
*/
2121
package org.biojava3.sequencing.io.fastq;
2222

23+
import java.io.*;
2324
import java.net.URL;
24-
25-
import java.io.BufferedReader;
26-
import java.io.File;
27-
import java.io.FileInputStream;
28-
import java.io.InputStream;
29-
import java.io.InputStreamReader;
30-
import java.io.IOException;
31-
3225
import java.util.ArrayList;
3326
import java.util.List;
3427

@@ -37,26 +30,32 @@
3730
*
3831
* @since 3.0.3
3932
*/
40-
abstract class AbstractFastqReader
41-
implements FastqReader
42-
{
33+
abstract class AbstractFastqReader implements FastqReader {
4334

44-
/** Parser state. */
45-
private static enum State
46-
{
47-
/** Description parser state. */
48-
DESCRIPTION,
35+
/**
36+
* Parser state.
37+
*/
38+
private static enum State {
4939

50-
/** Sequence parser state. */
40+
/**
41+
* Description parser state.
42+
*/
43+
DESCRIPTION,
44+
/**
45+
* Sequence parser state.
46+
*/
5147
SEQUENCE,
52-
53-
/** Repeat description parser state. */
48+
/**
49+
* Repeat description parser state.
50+
*/
5451
REPEAT_DESCRIPTION,
55-
56-
/** Quality score parser state. */
52+
/**
53+
* Quality score parser state.
54+
*/
5755
QUALITY,
58-
59-
/** Complete parser state. */
56+
/**
57+
* Complete parser state.
58+
*/
6059
COMPLETE;
6160
};
6261

@@ -75,8 +74,7 @@ private static enum State
7574
* @param lineNumber current line number in input stream
7675
* @throws IOException if the specified description is not valid
7776
*/
78-
protected abstract void validateDescription(FastqBuilder builder, String description, int lineNumber)
79-
throws IOException;
77+
protected abstract void validateDescription(FastqBuilder builder, String description, int lineNumber) throws IOException;
8078

8179
/**
8280
* Validate the specified sequence.
@@ -86,8 +84,7 @@ protected abstract void validateDescription(FastqBuilder builder, String descrip
8684
* @param lineNumber current line number in input stream
8785
* @throws IOException if the specified sequence is not valid
8886
*/
89-
protected abstract void validateSequence(FastqBuilder builder, String sequence, int lineNumber)
90-
throws IOException;
87+
protected abstract void validateSequence(FastqBuilder builder, String sequence, int lineNumber) throws IOException;
9188

9289
/**
9390
* Validate the specified repeat description.
@@ -97,8 +94,7 @@ protected abstract void validateSequence(FastqBuilder builder, String sequence,
9794
* @param lineNumber current line number in input stream
9895
* @throws IOException if the specified repeat description is not valid
9996
*/
100-
protected abstract void validateRepeatDescription(FastqBuilder builder, String repeatDescription, int lineNumber)
101-
throws IOException;
97+
protected abstract void validateRepeatDescription(FastqBuilder builder, String repeatDescription, int lineNumber) throws IOException;
10298

10399
/**
104100
* Validate the specified quality scores.
@@ -108,160 +104,127 @@ protected abstract void validateRepeatDescription(FastqBuilder builder, String r
108104
* @param lineNumber current line number in input stream
109105
* @throws IOException if the specified quality scores are not valid
110106
*/
111-
protected abstract void validateQuality(FastqBuilder builder, String quality, int lineNumber)
112-
throws IOException;
113-
107+
protected abstract void validateQuality(FastqBuilder builder, String quality, int lineNumber) throws IOException;
114108

115-
/** {@inheritDoc} */
116-
public final Iterable<Fastq> read(final File file) throws IOException
117-
{
118-
if (file == null)
119-
{
109+
/**
110+
* {@inheritDoc}
111+
*/
112+
@Override
113+
public final Iterable<Fastq> read(final File file) throws IOException {
114+
if (file == null) {
120115
throw new IllegalArgumentException("file must not be null");
121116
}
122117
InputStream inputStream = null;
123-
try
124-
{
118+
try {
125119
inputStream = new FileInputStream(file);
126120
return read(inputStream);
127-
}
128-
catch (IOException e)
129-
{
121+
} catch (IOException e) {
130122
throw e;
131-
}
132-
finally
133-
{
134-
if (inputStream != null)
135-
{
136-
try
137-
{
123+
} finally {
124+
if (inputStream != null) {
125+
try {
138126
inputStream.close();
139-
}
140-
catch (IOException e)
141-
{
127+
} catch (IOException e) {
142128
// ignore
143129
}
144130
}
145131
}
146132
}
147133

148-
/** {@inheritDoc} */
149-
public final Iterable<Fastq> read(final URL url) throws IOException
150-
{
151-
if (url == null)
152-
{
134+
/**
135+
* {@inheritDoc}
136+
*/
137+
@Override
138+
public final Iterable<Fastq> read(final URL url) throws IOException {
139+
if (url == null) {
153140
throw new IllegalArgumentException("url must not be null");
154141
}
155142
InputStream inputStream = null;
156-
try
157-
{
143+
try {
158144
inputStream = url.openStream();
159145
return read(inputStream);
160-
}
161-
catch (IOException e)
162-
{
146+
} catch (IOException e) {
163147
throw e;
164-
}
165-
finally
166-
{
167-
if (inputStream != null)
168-
{
169-
try
170-
{
148+
} finally {
149+
if (inputStream != null) {
150+
try {
171151
inputStream.close();
172-
}
173-
catch (IOException e)
174-
{
152+
} catch (IOException e) {
175153
// ignore
176154
}
177155
}
178156
}
179157
}
180158

181-
/** {@inheritDoc} */
182-
public final Iterable<Fastq> read(final InputStream inputStream) throws IOException
183-
{
184-
if (inputStream == null)
185-
{
186-
throw new IllegalArgumentException ("inputStream must not be null");
159+
/**
160+
* {@inheritDoc}
161+
*/
162+
@Override
163+
public final Iterable<Fastq> read(final InputStream inputStream) throws IOException {
164+
if (inputStream == null) {
165+
throw new IllegalArgumentException("inputStream must not be null");
187166
}
188167
int lineNumber = 0;
189168
State state = State.DESCRIPTION;
190169
List<Fastq> result = new ArrayList<Fastq>();
191170
FastqBuilder builder = new FastqBuilder().withVariant(getVariant());
192171
BufferedReader reader = new BufferedReader(new InputStreamReader(inputStream));
193-
while (reader.ready())
194-
{
172+
while (reader.ready()) {
195173
String line = reader.readLine();
196-
switch (state)
197-
{
198-
case DESCRIPTION:
199-
validateDescription(builder, line, lineNumber);
200-
builder.withDescription(line.substring(1).trim());
201-
state = State.SEQUENCE;
202-
break;
203-
case SEQUENCE:
204-
validateSequence(builder, line, lineNumber);
205-
builder.withSequence(line.trim());
206-
state = State.REPEAT_DESCRIPTION;
207-
break;
208-
case REPEAT_DESCRIPTION:
209-
if (!line.startsWith("+"))
210-
{
211-
builder.appendSequence(line.trim());
212-
}
213-
else
214-
{
215-
validateRepeatDescription(builder, line, lineNumber);
216-
state = State.QUALITY;
217-
}
218-
break;
219-
case QUALITY:
220-
validateQuality(builder, line, lineNumber);
221-
builder.withQuality(line.trim());
222-
state = State.COMPLETE;
223-
break;
224-
case COMPLETE:
225-
if (!builder.sequenceAndQualityLengthsMatch())
226-
{
227-
builder.appendQuality(line.trim());
228-
}
229-
else
230-
{
231-
try
232-
{
233-
result.add(builder.build());
234-
}
235-
catch (IllegalStateException e)
236-
{
237-
throw new IOException("caught an IllegalStateException at line " + lineNumber + " " + e.getMessage());
238-
//throw new IOException("caught an IllegalStateException at line " + lineNumber, e); jdk 1.6+
239-
}
174+
switch (state) {
175+
case DESCRIPTION:
240176
validateDescription(builder, line, lineNumber);
241177
builder.withDescription(line.substring(1).trim());
242178
state = State.SEQUENCE;
243-
}
244-
break;
245-
default:
246-
break;
179+
break;
180+
case SEQUENCE:
181+
validateSequence(builder, line, lineNumber);
182+
builder.withSequence(line.trim());
183+
state = State.REPEAT_DESCRIPTION;
184+
break;
185+
case REPEAT_DESCRIPTION:
186+
if (!line.startsWith("+")) {
187+
builder.appendSequence(line.trim());
188+
} else {
189+
validateRepeatDescription(builder, line, lineNumber);
190+
state = State.QUALITY;
191+
}
192+
break;
193+
case QUALITY:
194+
validateQuality(builder, line, lineNumber);
195+
builder.withQuality(line.trim());
196+
state = State.COMPLETE;
197+
break;
198+
case COMPLETE:
199+
if (!builder.sequenceAndQualityLengthsMatch()) {
200+
builder.appendQuality(line.trim());
201+
} else {
202+
try {
203+
result.add(builder.build());
204+
} catch (IllegalStateException e) {
205+
throw new IOException("caught an IllegalStateException at line " + lineNumber + " " + e.getMessage());
206+
//throw new IOException("caught an IllegalStateException at line " + lineNumber, e); jdk 1.6+
207+
}
208+
validateDescription(builder, line, lineNumber);
209+
builder.withDescription(line.substring(1).trim());
210+
state = State.SEQUENCE;
211+
}
212+
break;
213+
default:
214+
break;
247215
}
248216
lineNumber++;
249217
}
250-
if (state == State.COMPLETE)
251-
{
252-
try
253-
{
218+
if (state == State.COMPLETE) {
219+
try {
254220
result.add(builder.build());
255221
state = State.DESCRIPTION;
256-
}
257-
catch (IllegalStateException e)
258-
{
259-
throw new IOException("caught an IllegalStateException at line " + lineNumber + " " + e.getMessage());
260-
//throw new IOException("caught an IllegalStateException at line " + lineNumber, e); jdk 1.6+
222+
} catch (IllegalStateException e) {
223+
//throw new IOException("caught an IllegalStateException at line " + lineNumber + " " + e.getMessage());
224+
throw new IOException("caught an IllegalStateException at line " + lineNumber, e);
261225
}
262226
}
263-
if (state != State.DESCRIPTION)
264-
{
227+
if (state != State.DESCRIPTION) {
265228
throw new IOException("truncated sequence at line " + lineNumber);
266229
}
267230
return result;

0 commit comments

Comments
 (0)