Page History
Starting the WebSpellChecker Server With package version 5.6.3.0, we added the ability to enable n-gram data sets to sets in the grammar engine to detect errors with words that are often confused, like their and there. This option is available for the English. The n-gram data sets are available for the group of English dialects, German, French, Spanish, and Dutch languages.
Due to a considerable increase in minimal hardware requirements and a moderate increase in quality, n-grams are not included by default in the standard packages. However, if you still prefer to make use of this option, this guide will help you to setup n-grams for your on-premises installation.
Tip |
---|
|
...
|
1. Download and unzip n-gram data sets
Contact our support team to get the links to download the latest n-gram data sets for required language(s).
Note |
---|
The n-gram data sets are huge and may take from 2.4-14.3GB of SSD depending on the chosen language or their combination. Make sure you have a SSD that fits the space requirements and has at least 10% free space left. |
Details about file sizes by languages are represented in the table below.
Language | Zipped size, GB | Unzipped size, GB |
---|---|---|
English | 8.75 | 14.3 |
German | 1.58 | 3.06 |
Spanish | 1.68 | 3.03 |
French | 1.78 | 3.17 |
Dutch | 1.2 | 2.4 |
2. Stop AppServer
It is recommended to stop AppServer before
...
Note |
---|
N-gram data sets take 8-16 GB space depending on the chosen language. Please make sure you have a fast SSD. |
1. Stop AppServer
Before making any changes to the AppServerX.xml file, it is recommended to stop AppServer.
...
3. Specify path to n-gram data sets in AppServer configuration file
- Open the AppServerX.xml configuration file for editing.
Info |
---|
The default path to the AppServerX.xml file: <WebSpellChecker_Installation_Path>/AppServer/AppServerX.xml |
- Find section with the PathToNgramData parameter which is responsible for enabling and configuration of n-gram data sets:. PathToNgramData.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<!-- Path to n-gram data sets. Can be used to improve grammar quality. --> <!-- <PathToNgramData></PathToNgramData>--> |
- Uncomment the PathToNgramData parameter and set a path to unzipped folder of ngramsn-grams.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<PathToNgramData>your_path_to_ngrams</PathToNgramData> |
Info |
---|
<!-- Path example For example, the path for Windows: <PathToNgramData>C <PathToNgramData>C:/Program Files/WebSpellChecker/AppServer/NgramData/</PathToNgramData> →PathToNgramData> The path <!-- Path example for Linux: <PathToNgramData>/homeopt/spellcheckWSC/svcAppServer/NgramData/</PathToNgramData> →PathToNgramData> |
- Add the EnableNgramData parameter inside the Language tag for each language you want to use nlanguage(s) where n-grams should be enabled.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<EnableNgramData>true</EnableNgramData> |
This is an example of the for added the EnableNgramData parameter enabled for American English. You can find the list of language short code (used as Language Id) with the approprialte language in the Default Language guide.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<Language Id="en_US"> <Alias>en</Alias> <Alias>am</Alias> <GrammarCheckProviderOptions>en-US</GrammarCheckProviderOptions> <EnableNgramData>true</EnableNgramData> <ThesaurusEnabled>true</ThesaurusEnabled> <SpellEngineOptions> <Locale>am</Locale> <SpellCheckProvider>ssce</SpellCheckProvider> <Dictionary FullPath="ssceam2.clx"> <ForSuggest>no</ForSuggest> </Dictionary> <Dictionary FullPath="ssceam2s.clx"> <ForSuggest>yes</ForSuggest> </Dictionary> <Dictionary FullPath="sscema2.clx"/> <Dictionary FullPath="keywords.clx"/> <Dictionary FullPath="ssceam.tlx"/> </SpellEngineOptions> </Language> |
...
4. Start AppServer
As soon as a correct path to ngrams and enabke it for the nesessary lanagues you made the nesessary actions to enable n-gram data sets in AppServerX.xml, start AppServer for the changes to take effect.
...