This is an automated email from the ASF dual-hosted git repository.
mawiesne pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/opennlp.git
The following commit(s) were added to refs/heads/main by this push:
new 6a6f9e6d OPENNLP-1587 Reduce creation of prefix String objects in
SequenceCodec classes
6a6f9e6d is described below
commit 6a6f9e6d5003919f657320f97d5add74adffdb72
Author: Martin Wiesner <[email protected]>
AuthorDate: Mon Jul 1 14:19:29 2024 +0200
OPENNLP-1587 Reduce creation of prefix String objects in SequenceCodec
classes
---
.../java/opennlp/tools/namefind/BilouCodec.java | 22 +++++++++-------------
.../main/java/opennlp/tools/namefind/BioCodec.java | 12 +++++-------
.../java/opennlp/tools/util/SequenceCodec.java | 2 ++
3 files changed, 16 insertions(+), 20 deletions(-)
diff --git a/opennlp-tools/src/main/java/opennlp/tools/namefind/BilouCodec.java
b/opennlp-tools/src/main/java/opennlp/tools/namefind/BilouCodec.java
index 8b9a20bc..7a37e299 100644
--- a/opennlp-tools/src/main/java/opennlp/tools/namefind/BilouCodec.java
+++ b/opennlp-tools/src/main/java/opennlp/tools/namefind/BilouCodec.java
@@ -90,7 +90,7 @@ public class BilouCodec implements SequenceCodec<String> {
if (name.length() > 1) {
if (name.getType() == null) {
- outcomes[name.getStart()] = "default" + "-" + BilouCodec.START;
+ outcomes[name.getStart()] = DEFAULT_PREFIX + BilouCodec.START;
}
else {
outcomes[name.getStart()] = name.getType() + "-" + BilouCodec.START;
@@ -98,7 +98,7 @@ public class BilouCodec implements SequenceCodec<String> {
// now iterate from begin + 1 till end
for (int i = name.getStart() + 1; i < name.getEnd() - 1; i++) {
if (name.getType() == null) {
- outcomes[i] = "default" + "-" + BilouCodec.CONTINUE;
+ outcomes[i] = DEFAULT_PREFIX + BilouCodec.CONTINUE;
}
else {
outcomes[i] = name.getType() + "-" + BilouCodec.CONTINUE;
@@ -106,7 +106,7 @@ public class BilouCodec implements SequenceCodec<String> {
}
if (name.getType() == null) {
- outcomes[name.getEnd() - 1] = "default" + "-" + BilouCodec.LAST;
+ outcomes[name.getEnd() - 1] = DEFAULT_PREFIX + BilouCodec.LAST;
}
else {
outcomes[name.getEnd() - 1] = name.getType() + "-" + BilouCodec.LAST;
@@ -114,7 +114,7 @@ public class BilouCodec implements SequenceCodec<String> {
}
else {
if (name.getType() == null) {
- outcomes[name.getEnd() - 1] = "default" + "-" + BilouCodec.UNIT;
+ outcomes[name.getEnd() - 1] = DEFAULT_PREFIX + BilouCodec.UNIT;
}
else {
outcomes[name.getEnd() - 1] = name.getType() + "-" + BilouCodec.UNIT;
@@ -152,23 +152,19 @@ public class BilouCodec implements SequenceCodec<String> {
for (String outcome : outcomes) {
if (outcome.endsWith(BilouCodec.START)) {
- start.add(outcome.substring(0, outcome.length()
- - BilouCodec.START.length()));
+ start.add(outcome.substring(0, outcome.length() -
BilouCodec.START.length()));
} else if (outcome.endsWith(BilouCodec.CONTINUE)) {
- cont.add(outcome.substring(0, outcome.length()
- - BilouCodec.CONTINUE.length()));
+ cont.add(outcome.substring(0, outcome.length() -
BilouCodec.CONTINUE.length()));
} else if (outcome.endsWith(BilouCodec.LAST)) {
- last.add(outcome.substring(0, outcome.length()
- - BilouCodec.LAST.length()));
+ last.add(outcome.substring(0, outcome.length() -
BilouCodec.LAST.length()));
} else if (outcome.endsWith(BilouCodec.UNIT)) {
- unit.add(outcome.substring(0, outcome.length()
- - BilouCodec.UNIT.length()));
+ unit.add(outcome.substring(0, outcome.length() -
BilouCodec.UNIT.length()));
} else if (!outcome.equals(BilouCodec.OTHER)) {
return false;
}
}
- if (start.size() == 0 && unit.size() == 0) {
+ if (start.isEmpty() && unit.isEmpty()) {
return false;
} else {
// Start, must have matching Last
diff --git a/opennlp-tools/src/main/java/opennlp/tools/namefind/BioCodec.java
b/opennlp-tools/src/main/java/opennlp/tools/namefind/BioCodec.java
index 622b228a..f7d6c5e8 100644
--- a/opennlp-tools/src/main/java/opennlp/tools/namefind/BioCodec.java
+++ b/opennlp-tools/src/main/java/opennlp/tools/namefind/BioCodec.java
@@ -99,7 +99,7 @@ public class BioCodec implements SequenceCodec<String> {
Arrays.fill(outcomes, BioCodec.OTHER);
for (Span name : names) {
if (name.getType() == null) {
- outcomes[name.getStart()] = "default" + "-" + BioCodec.START;
+ outcomes[name.getStart()] = DEFAULT_PREFIX + BioCodec.START;
}
else {
outcomes[name.getStart()] = name.getType() + "-" + BioCodec.START;
@@ -107,7 +107,7 @@ public class BioCodec implements SequenceCodec<String> {
// now iterate from begin + 1 till end
for (int i = name.getStart() + 1; i < name.getEnd(); i++) {
if (name.getType() == null) {
- outcomes[i] = "default" + "-" + BioCodec.CONTINUE;
+ outcomes[i] = DEFAULT_PREFIX + BioCodec.CONTINUE;
}
else {
outcomes[i] = name.getType() + "-" + BioCodec.CONTINUE;
@@ -135,18 +135,16 @@ public class BioCodec implements SequenceCodec<String> {
for (String outcome : outcomes) {
if (outcome.endsWith(BioCodec.START)) {
- start.add(outcome.substring(0, outcome.length()
- - BioCodec.START.length()));
+ start.add(outcome.substring(0, outcome.length() -
BioCodec.START.length()));
} else if (outcome.endsWith(BioCodec.CONTINUE)) {
- cont.add(outcome.substring(0, outcome.length()
- - BioCodec.CONTINUE.length()));
+ cont.add(outcome.substring(0, outcome.length() -
BioCodec.CONTINUE.length()));
} else if (!outcome.equals(BioCodec.OTHER)) {
// got unexpected outcome
return false;
}
}
- if (start.size() == 0) {
+ if (start.isEmpty()) {
return false;
} else {
for (String contPreffix : cont) {
diff --git a/opennlp-tools/src/main/java/opennlp/tools/util/SequenceCodec.java
b/opennlp-tools/src/main/java/opennlp/tools/util/SequenceCodec.java
index 7f304f5a..b6d74c36 100644
--- a/opennlp-tools/src/main/java/opennlp/tools/util/SequenceCodec.java
+++ b/opennlp-tools/src/main/java/opennlp/tools/util/SequenceCodec.java
@@ -27,6 +27,8 @@ import java.util.List;
*/
public interface SequenceCodec<T> {
+ String DEFAULT_PREFIX = "default-";
+
/**
* Decodes a sequence of {@link T objects} into {@link Span} objects.
*
