This is an automated email from the ASF dual-hosted git repository.
aradzinski pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-nlpcraft.git
The following commit(s) were added to refs/heads/master by this push:
new fe73434 WIP
fe73434 is described below
commit fe73434edfc02a949c723f752fb6044c85f74a27
Author: Aaron Radzinski <[email protected]>
AuthorDate: Tue Dec 28 10:16:44 2021 -0800
WIP
---
.../src/main/scala/org/apache/nlpcraft/NCLifecycle.java | 2 +-
.../main/scala/org/apache/nlpcraft/NCModelClient.java | 16 +++++++---------
.../nlp/token/enricher/NCEnBracketsTokenEnricher.java | 4 ++--
.../nlp/token/enricher/NCEnDictionaryTokenEnricher.java | 4 ++--
.../nlp/token/enricher/NCEnLanguageTokenEnricher.java | 4 ++--
.../nlp/token/enricher/NCEnQuotesTokenEnricher.java | 4 ++--
.../nlp/token/enricher/NCEnSwearWordsTokenEnricher.java | 4 ++--
.../nlp/token/enricher/impl/NCEnDictionaryImpl.scala | 2 +-
.../nlp/token/enricher/impl/NCEnSwearWordsImpl.scala | 2 +-
.../nlp/token/parser/opennlp/NCEnOpenNlpTokenParser.java | 4 ++--
.../nlp/token/parser/opennlp/impl/NCEnOpenNlpImpl.scala | 2 +-
.../parser/opennlp/NCEnOpenNlpTokenParserBenchmark.java | 2 +-
.../apache/nlpcraft/internal/nlp/util/NCTestUtils.scala | 2 +-
13 files changed, 25 insertions(+), 27 deletions(-)
diff --git a/nlpcraft/src/main/scala/org/apache/nlpcraft/NCLifecycle.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/NCLifecycle.java
index ac1f431..0f3774c 100644
--- a/nlpcraft/src/main/scala/org/apache/nlpcraft/NCLifecycle.java
+++ b/nlpcraft/src/main/scala/org/apache/nlpcraft/NCLifecycle.java
@@ -26,7 +26,7 @@ public interface NCLifecycle {
/**
*
*/
- default void start() {}
+ default void start(NCModelConfig cfg) {}
/**
*
diff --git a/nlpcraft/src/main/scala/org/apache/nlpcraft/NCModelClient.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/NCModelClient.java
index 8534528..0e4e4fb 100644
--- a/nlpcraft/src/main/scala/org/apache/nlpcraft/NCModelClient.java
+++ b/nlpcraft/src/main/scala/org/apache/nlpcraft/NCModelClient.java
@@ -43,9 +43,9 @@ public class NCModelClient implements NCLifecycle {
// TODO:
}
- private static void start(List<? extends NCLifecycle> list) {
+ private static void start(List<? extends NCLifecycle> list, NCModelConfig
cfg) {
if (list != null)
- list.forEach(p -> p.start());
+ list.forEach(p -> p.start(cfg));
}
private static void stop(List<? extends NCLifecycle> list) {
@@ -54,15 +54,13 @@ public class NCModelClient implements NCLifecycle {
}
@Override
- public void start() {
+ public void start(NCModelConfig cfg) {
verify();
- NCModelConfig cfg = mdl.getConfig();
-
- cfg.getTokenParser().start();
- start(cfg.getEntityParsers());
- start(cfg.getEntityEnrichers());
- start(cfg.getTokenEnrichers());
+ cfg.getTokenParser().start(cfg);
+ start(cfg.getEntityParsers(), cfg);
+ start(cfg.getEntityEnrichers(), cfg);
+ start(cfg.getTokenEnrichers(), cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnBracketsTokenEnricher.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnBracketsTokenEnricher.java
index 43920a6..bf7d0b9 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnBracketsTokenEnricher.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnBracketsTokenEnricher.java
@@ -29,8 +29,8 @@ public class NCEnBracketsTokenEnricher implements
NCTokenEnricher {
private final NCEnBracketsImpl impl = new NCEnBracketsImpl();
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnDictionaryTokenEnricher.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnDictionaryTokenEnricher.java
index a845eeb..30ed1e9 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnDictionaryTokenEnricher.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnDictionaryTokenEnricher.java
@@ -28,8 +28,8 @@ public class NCEnDictionaryTokenEnricher implements
NCTokenEnricher {
private final NCEnDictionaryImpl impl = new NCEnDictionaryImpl();
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnLanguageTokenEnricher.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnLanguageTokenEnricher.java
index 687f9a1..cdd765b 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnLanguageTokenEnricher.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnLanguageTokenEnricher.java
@@ -29,8 +29,8 @@ public class NCEnLanguageTokenEnricher implements
NCTokenEnricher {
private final NCEnLanguageWordsImpl impl = new NCEnLanguageWordsImpl();
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnQuotesTokenEnricher.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnQuotesTokenEnricher.java
index ec2931a..834732c 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnQuotesTokenEnricher.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnQuotesTokenEnricher.java
@@ -32,8 +32,8 @@ public class NCEnQuotesTokenEnricher implements
NCTokenEnricher {
private final NCEnQuotesImpl impl = new NCEnQuotesImpl();
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnSwearWordsTokenEnricher.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnSwearWordsTokenEnricher.java
index 3e9e192..654d446 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnSwearWordsTokenEnricher.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/NCEnSwearWordsTokenEnricher.java
@@ -54,8 +54,8 @@ public class NCEnSwearWordsTokenEnricher implements
NCTokenEnricher {
}
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnDictionaryImpl.scala
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnDictionaryImpl.scala
index 2c22323..0c48f46 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnDictionaryImpl.scala
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnDictionaryImpl.scala
@@ -26,7 +26,7 @@ import org.apache.nlpcraft.internal.util.NCUtils
class NCEnDictionaryImpl extends NCTokenEnricher:
@volatile private var dict: Set[String] = _
- override def start(): Unit = dict =
NCUtils.readResource("moby/354984si.ngl", "iso-8859-1").toSet
+ override def start(cfg: NCModelConfig): Unit = dict =
NCUtils.readResource("moby/354984si.ngl", "iso-8859-1").toSet
override def stop(): Unit = dict = null
override def enrich(req: NCRequest, cfg: NCModelConfig, toks:
java.util.List[NCToken]): Unit =
toks.forEach(t => t.put("dict:en", dict.contains(t.getLemma)))
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnSwearWordsImpl.scala
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnSwearWordsImpl.scala
index af79096..1551e21 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnSwearWordsImpl.scala
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/enricher/impl/NCEnSwearWordsImpl.scala
@@ -54,7 +54,7 @@ object NCEnSwearWordsImpl:
class NCEnSwearWordsImpl(is: InputStream, res: String) extends NCTokenEnricher
with LazyLogging:
@volatile private var swearWords: Set[String] = _
- override def start(): Unit =
+ override def start(cfg: NCModelConfig): Unit =
val stemmer = new PorterStemmer
swearWords = NCUtils.readTextStream(is,
"UTF-8").map(stemmer.stem).toSet
logger.trace(s"Loaded resource: $res")
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCEnOpenNlpTokenParser.java
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCEnOpenNlpTokenParser.java
index a5ff4a6..d0cd2fb 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCEnOpenNlpTokenParser.java
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCEnOpenNlpTokenParser.java
@@ -37,8 +37,8 @@ public class NCEnOpenNlpTokenParser implements NCTokenParser {
private final NCEnOpenNlpImpl impl;
@Override
- public void start() {
- impl.start();
+ public void start(NCModelConfig cfg) {
+ impl.start(cfg);
}
@Override
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCEnOpenNlpImpl.scala
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCEnOpenNlpImpl.scala
index 5f6483b..428f08a 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCEnOpenNlpImpl.scala
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCEnOpenNlpImpl.scala
@@ -75,7 +75,7 @@ class NCEnOpenNlpImpl(
private var addStopWords: JSet[String] = _
private var exclStopWords: JSet[String] = _
- override def start(): Unit =
+ override def start(cfg: NCModelConfig): Unit =
NCUtils.execPar(
() => tokenizer = new TokenizerME(new TokenizerModel(tokMdlIn)),
() => tagger = new POSTaggerME(new POSModel(posMdlIn)),
diff --git
a/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/benchmark/token/parser/opennlp/NCEnOpenNlpTokenParserBenchmark.java
b/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/benchmark/token/parser/opennlp/NCEnOpenNlpTokenParserBenchmark.java
index 59e01ee..193dd00 100644
---
a/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/benchmark/token/parser/opennlp/NCEnOpenNlpTokenParserBenchmark.java
+++
b/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/benchmark/token/parser/opennlp/NCEnOpenNlpTokenParserBenchmark.java
@@ -58,7 +58,7 @@ public class NCEnOpenNlpTokenParserBenchmark extends
NCBenchmarkAdapter {
"opennlp/en-lemmatizer.dict"
);
- p.start();
+ p.start(null); // TODO: fix it.
return p;
}
diff --git
a/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/util/NCTestUtils.scala
b/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/util/NCTestUtils.scala
index 2b6b270..f5d624f 100644
---
a/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/util/NCTestUtils.scala
+++
b/nlpcraft/src/test/java/org/apache/nlpcraft/internal/nlp/util/NCTestUtils.scala
@@ -81,6 +81,6 @@ object NCTestUtils:
val t = make
val started = now()
- t.start()
+ t.start(null) // TODO: fix it.
println(s"'${t.getClass.getSimpleName}' created with time=${started -
start} ms and started=${now() - started} ms.")
t
