This is an automated email from the ASF dual-hosted git repository.
sergeykamov pushed a commit to branch NLPCRAFT-468
in repository https://gitbox.apache.org/repos/asf/incubator-nlpcraft.git
The following commit(s) were added to refs/heads/NLPCRAFT-468 by this push:
new cdd23f0 WIP.
cdd23f0 is described below
commit cdd23f0fb21a73526b7789131343e17599f9dcab
Author: Sergey Kamov <[email protected]>
AuthorDate: Wed Oct 13 10:39:52 2021 +0300
WIP.
---
.../nlpcraft/model/curation+confirmation.txt | 12 ++++-----
.../scala/org/apache/nlpcraft/model/interfaces.txt | 29 +++++++++++-----------
2 files changed, 20 insertions(+), 21 deletions(-)
diff --git
a/nlpcraft/src/main/scala/org/apache/nlpcraft/model/curation+confirmation.txt
b/nlpcraft/src/main/scala/org/apache/nlpcraft/model/curation+confirmation.txt
index 156be4a..4a0fc84 100644
---
a/nlpcraft/src/main/scala/org/apache/nlpcraft/model/curation+confirmation.txt
+++
b/nlpcraft/src/main/scala/org/apache/nlpcraft/model/curation+confirmation.txt
@@ -20,8 +20,8 @@ a) add to intents property `need_confirm` (boolean, default
false)
b) add to models property `need_curation` (boolean, default false)
c) add to NCResult 3 fields:
- variants (sorted from best to worst, first is winner) List<List<NCToken>>,
mandatory.
-- `need_confirm` (boolean, optional, omitted processed as false)
-- `need_curation` (boolean, optional, omitted processed as false)
+- `need_confirm` (boolean, optional, if omitted - processed as false)
+- `need_curation` (boolean, optional, if omitted - processed as false)
1. Confirmation
- add 'need_confirm' = true flag to intents, which require confirmations.
@@ -31,12 +31,12 @@ c) add to NCResult 3 fields:
2. Curation.
- add 'need_curation' = true flag to model, which supports curation.
-- set `need_curation` = true in NCResult if there isn't intent found.
+- set `need_curation` = true in NCResult if there isn't any intent found.
- add new API method
- - `String curateText(String srvReqId)`; - which returns fixes request text
+ - `String curateText(String reqId)`; - which returns fixed request text
or
- - `List<NCToken> curateVariant(String srvReqId)`; - which returns created
winner variant.
+ - `List<NCToken> curateVariant(String reqId)`; - which returns created
winner variant.
-Curated request cached easiest way - by stems without stopwords.
+Curated requests cached the easiest way - by stems and without stopwords
(maybe with/without ordering)
diff --git a/nlpcraft/src/main/scala/org/apache/nlpcraft/model/interfaces.txt
b/nlpcraft/src/main/scala/org/apache/nlpcraft/model/interfaces.txt
index 654b30c..1f8fc28 100644
--- a/nlpcraft/src/main/scala/org/apache/nlpcraft/model/interfaces.txt
+++ b/nlpcraft/src/main/scala/org/apache/nlpcraft/model/interfaces.txt
@@ -15,37 +15,36 @@
# limitations under the License.
#
-Interfaces, which can be implemented by user.
-(All of them already have built-in implementations)
+Interfaces (pluggable components). All of them already have built-in
implementations
-1. org.apache.nlpcraft.model.nlp.NCNlpTokenizer. Text-to-words tokenizer.
+1. Text-to-words tokenizer - org.apache.nlpcraft.model.nlp.NCNlpTokenizer.
Delivered:
- org.apache.nlpcraft.model.components.tokenizer.NCOpenNlpTokenizer (not
configured)
- - Stanford impl.
+ - Stanford impl (not configured)
Default in config - NCOpenNlpTokenizer.
When user need to implement own:
- - own logic implementation
+ - own logic required (for example `opennlp` implementation is not satisfied,
and `stanford` license is not suitable)
- new languages support
-2. org.apache.nlpcraft.model.nlp.NCNlpNerParse. Ners finder.
+2. Ners finder - org.apache.nlpcraft.model.nlp.NCNlpNerParse.
Delivered:
- - org.apache.nlpcraft.model.components.ner.opennlp.NCOpenNlpNerParser
(configured types from predefined set of supported NERS: DATE, PERSON etc)
- - stanford impl (configured types from predefined set of supported NERS:
DATE, PERSON etc)
+ - org.apache.nlpcraft.model.components.ner.opennlp.NCOpenNlpNerParser
(configured from predefined set of supported NERS: DATE, PERSON etc)
+ - stanford impl (configured from predefined set of supported NERS: DATE,
PERSON etc)
- org.apache.nlpcraft.model.components.ner.synonyms.NCSynonymsNerParser
(configured by synonyms, macros ect)
Default in config - empty list.
When user need to implement own:
- - custom elements' logic support, which cannot be supported by
NCSynonymsNerParser
+ - custom elements' detection logic support, which cannot be defined by
NCSynonymsNerParser
- wrappers under existing NERS like Spacy
-3. org.apache.nlpcraft.model.nlp.NCNlpWordsDetector. Special words finder.
+3. Special words finder - org.apache.nlpcraft.model.nlp.NCNlpWordsDetector.
Delivered:
- org.apache.nlpcraft.model.components.detectors.NCDefaultStopWordsDetector
(EN, configured with additional and excluded words set)
- org.apache.nlpcraft.model.components.detectors.NCDefaultSwearWordsDetector
(EN, not configured)
- - org.apache.nlpcraft.model.components.detectors.NCConfiguredWordsDetector
(configured by simple words set)
+ - org.apache.nlpcraft.model.components.detectors.NCConfiguredWordsDetector
(configured simple way with words set)
Default in config - NCDefaultStopWordsDetector, NCDefaultSwearWordsDetector
-(`suspicious` doesn't have by default. Can be configured by
NCConfiguredWordsDetector)
+(`suspicious` detector is not set by default. Can be configured by
NCConfiguredWordsDetector)
When user need to implement own:
- - own logic implementation, which cannot be configured by
NCConfiguredWordsDetector
+ - own sophisticated logic implementation, which cannot be configured by
NCConfiguredWordsDetector
- new languages support
4. org.apache.nlpcraft.model.NCModelBehaviour
@@ -53,5 +52,5 @@ Delivered:
- No special implementation, this interface has all empty default methods.
Default in config - empty implementation.
When user need to implement own:
- - usage without intents case
- - some tricks even using intents
\ No newline at end of file
+ - when system used without intents
+ - some tricks, even using intents
\ No newline at end of file
