This package contains iWnn Middleware. You can use this package to provide full language-processing capabilities.
iWnn performs integrated language processing for the device in which it is embedded. It provides the following language processing capabilities.
・Multilingual prediction and conversion (Japanese, English, Chinese, Korean, etc.)
・Japanese kana-kanji conversion
・Japanese morphological analysis (delimited input, attached reading)
・Dictionary search
You can also use this package in your products.
2.1 Notes on Middleware
o Type names used by the external API (including structure and function types) have been standardized to IWNN_*. Preprocessor definitions are identified by NJ_*. o Make sure the file system does not access the dictionary file (*.A) simultaneously from multiple threads in a multi-thread environment (it is not thread-safe).
2.2 Notes on Sample Programs
o To run the sample demos, copy the folder named resources/sdcard/iWnn onto an SD Card. (Put the iWnn folder in the root directory of the SD Card.))
o Dictionary files are placed on the SD Card by the sample program. Separate provisions are required in order to place the dictionary in NAND memory.
o An operating log for the sample programs is created in UTF-16LE format under the names Result1_*.txt through Result5_*.txt in the iWinn folder on the SD Card.
o The following precautions are required when using the sample programs analyzeConv, searchAddWord, morpho, and otherAPIs. - The input text file is UTF-16BE format without a BOM symbol.
o The following precautions are required when using the multiDicSearch sample program. - The iWnn/rewrite folder (containing an empty learning dictionary and user dictionary) must be properly located ahead of time.
- Input text files are in UTF-16BE format without a BOM symbol. The only recognized newline code is a newline at the lowest row of the text.
- The fuzzy-match table for single-tap input supports the following languages.
English (US), English (GB), French, French (Quebec), German, Spanish, Portuguese, Portuguese (Brazil), Italian, Dutch
Chinese (Traditional-Pinyin), Chinese Language (Simplified-Pinyin)
- The fuzzy-match table for full keyboard input supports the following languages.
English (US), English (GB), French, French (Quebec), German, Spanish, Portuguese, Portuguese (Brazil), Italian, Dutch, Russian
Chinese (Traditional-Pinyin), Chinese (Simplified-Pinyin)
- Chinese (Simplified) is implemented to accept Pinyin input; Chinese (Traditional) is implemented to accept Pinyin and Zhuyin input; Cantonese is implemented to accept Cangjie input only.
- Four-tone input is not supported. (Only the current filter function is implemented. If required, you must separately implement a process for extracting four-tone information from the input character string.)
Conversion is not supported for Chinese (Hong Kong) Cangjie input. (Only prediction is supported.)
The following development environment is required in order to run this package.
This package has been confirmed to run in the following development environment versions.
Extract this package to any directory. Nintendo recommends that you put this package in a directory with a path that does not include spaces or double-byte characters. After extracting, run the setup.wsf file to configure the absolute path to the package's root directory in the CTRMW_IWNN_ROOT environment variable.
Link to the following header files and libraries to include this middleware in an application.
include/mw/iwnn/iwnnCTR.hlibrary/CTR-TS.Process.MPCore/release(or verbose or noopt)/libmw_iwnn.fast.a (or libmw_iwnn.small.a)
The components of this package are listed below.
Note: This does not list all included files.
CTR_MW-iWnn/
|
+- documents/
| |
| +- API/
| | |
| | +- index.html … API Reference
| | +- mw/
| |
| +- Manual/
| |
| +- iWnn_API_Manual.pdf … CTR iWnn Programming Manual
|
+- include/
| |
| +- mw/
| |
| +- iwnn/
| |
| +- iwnnCTR.h … iWnn class include file (include file for the application)
| +- iwnn.h … iWnn class include file
| +- engine/ … set of iWnn engine include files
|
+- libraries/
| |
| +- CTR-TS.Process.MPCore/
| |
| +- noopt/
| | |
| | +- libmw_iwnn.fast.a … Library file (fast version)
| | +- libmw_iwnn.small.a … Library file (small version)
| |
| +- verbose/
| | |
| | +- libmw_iwnn.fast.a … Library file (fast version)
| | +- libmw_iwnn.small.a … Library file (small version)
| |
| +- release/
| |
| +- libmw_iwnn.fast.a … Library file (fast version)
| +- libmw_iwnn.small.a … Library file (small version)
|
+- sampledemos/ … Sample demos (5 total)
| |
| +- analyzeConv/
| | |
| | +- fprintf.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- analyzeConv.rsf … rsf file
| | +- OMakefile … OMakefile
| |
| +- searchAddWord/
| | |
| | +- fprintf.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- searchAddWord.rsf … rsf File
| | +- OMakefile … OMakefile
| |
| +- morpho/
| | |
| | +- fprintf.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- morpho.rsf … rsf file
| | +- OMakefile … OMakefile
| |
| +- multiDicSearch/
| | |
| | +- fprintf.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- multiDicSearch.rsf … rsf file
| | +- OMakefile … OMakefile
| | +- multidic*.h … Header file
| | +- multidic*.cpp … Source file
| |
| +- otherAPIs/
| | |
| | +- fprintf.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- otherAPIs.rsf … rsf file
| | +- OMakefile … OMakefile
| |
| +- OMakefile
| +- OMakeroot
|
+- resources/
| |
| +- result/ … Set of output text files (output result files for all sample applications)
| | |
| | +- Result1.txt … analyzeConv sample demo output file
| | +- Result2.txt … searchAddWord sample demo output file
| | +- Result3.txt … morpho sample demo output file
| | +- Result4.txt … multiDicSearch sample demo output file
| | +- Result5.txt … otherAPIs sample demo output file
| +- sdcard/ … SD Card Root
| |
| +- iWnn/ … Set of input text files and dictionary folders
| |
| +- anlz.txt … analyzeConv, otherAPIs, sample demo input file
| +- anlz_giji.txt … analyzeConv sample demo input file
| +- anlz_giji2.txt … analyzeConv sample demo input file
| +- conv.txt … analyzeConv sample demo input file
| +- conv2.txt … analyzeConv sample demo input file
| +- conv3.txt … analyzeConv sample demo input file
| +- additional.txt … searchAddWord sample demo input file
| +- candidate.txt … searchAddWord sample demo input file
| +- reading.txt … searchAddWord sample demo input file
| +- input_str.txt … morpho sample demo input file
| +- input.txt … multiDicSearch sample demo input file
| +- merge1.txt … otherAPIs sample demo input file
| +- merge2.txt … otherAPIs sample demo input file
| +- state.txt … otherAPIs sample demo input file
| +- stroke_word.txt … otherAPIs sample demo input file
| +- dic/ … Set of dictionary files (excluding the learning dictionary and user dictionary)
| | |
| | +- DE/
| | +- EN/
| | +- ES/
| | +- FR/
| | +- IT/
| | +- JA/
| | +- JA_small/
| | +- KO/
| | +- NL/
| | +- PT/
| | +- RU/
| | +- ZH/
| +- rewrite/ … multiDicSearch sample demo learning dictionary and user dictionary
|
+- tools/
| |
| +- dtoa/ … Set of tools for creating customized dictionaries
| | |
| | +- dtoa.exe … The tool
| | +- Dtoa_Manual_en.txt … Tool operating manual
| | +- conf … Tool configuration file
| | +- *.ux … Sample demo input text file
| | +- *.a … Sample output dictionary files
| |
| +- roman/ … Set of Roman character input sample demos
| |
| +- Roman_Manual_en.txt … Sample demo operating manual
| +- cpp/
| | |
| | +- ResultLog/ … Output text file
| | | |
| | | +- kanaUtf16.txt … Roman character input sample demo output file
| | |
| | +- fprintf.h … Header file
| | +- RomanLetter.h … Header file
| | +- fprintf.cpp … Source file
| | +- main.cpp … Source file
| | +- RomanLetter.cpp … Source file
| | +- roman.rsf … rsf file
| | +- OMakefile … OMakefile
| |
| +- OMakefile
| +- OMakeroot
|
+- Readme-iwnn-OM-1_5_0.html … Readme file
+- setup.wsf
For more information about this middleware's specifications and how to use the tools, see the Programming Manual.
iWnn Programming Manual (PDF)
For more information about each function and to learn how to run the sample demos, see the API Reference.
CTR iWnn Middleware API Reference Manual
iWnn_OM_1_5_0.)Overview.html).InformationOfDic.html).iWnn_OM_1_4_1.)NjxCreateDic.html).InformationOfDic.html).iWnn_OM_1_4_0.)iWnn_OM_1_3_1.)analyzeConv, morpho, searchAddWord, and otherAPIs).
iWnn_1_3_1.)OMakefile/OMakeroot/*.rsf).InformationOfDic.html).iWnn_1_3_0.)flush argument (true) to the (FS) Write function. (CTR-SDK 4.0.0 modified specification support)nn::os::Initialize function from all sample programs.multiDicSearch sample program, made changes so that "demo end!" is output as the result.NN_USING_MIDDLEWARE macro in the iWnn library from the constructor to the NjxInit function.iWnn_1_2_7.)analyzeConv sample program.analyzeConv sample program input file (anlz_giji2.txt) and changed the output result file (Result1.txt).multiDicSearch sample program's output result file (Result4.txt).iWnn_1_2_6.)multiDicSearch sample program.multiDicSearch sample program input file (input.txt) and output result file (Result4.txt).InformationOfDic.html).iWnn_1_1_0.)InformationOfDic.html).InformationOfDic.html) simplified and traditional dictionary versions.iWnn_1_2_4.)Result4.txt).InformationOfDic.html).iWnn_1_2_3.)iWnn_1_2_2.)iWnn_1_2_1.)NJ_MAX_ADDITIONAL_LEN.iwnnCTR.h, and deleted CCFLAGS defined by OMakefile.Finalize process for ExpHeap class instances.releaseDic function used in the moprho sample program.nn::fs::FileStream class to the freeLanguage function used in the multiDicSearch sample program.internalLoadDictionary function used in the multiDicSearch sample program for when a supplemental information dictionary error occurs.internalLoadDictionary function used in the multiDicSearch sample program for when an error occurs opening a flash dictionary.analyzeConv sample program.extensionArea is not NULL.iwnnCTR.h.iWnn_1_1_2.)iWnn_1_1_1.)iWnn_1_1_0.)iWnn_1_0_11.)iWnn_1_0_10.)iWnn_1_0_9.)tools/dtoa/Dtoa_Manual_jp.txt). tools/dtoa/*.ux) used with the dtoa tool has been changed. MountSdmc function are checked for each sample code and roman tool code.iWnn_1_0_8.)NN_DEFINE_MIDDLEWARE and NN_USING_MIDDLEWARE definitions were removed from iwnn.h. NJ_STEM defined in iwnn.h was renamed IWNN_STEM. NN_LOG function. internfalFseek function has been removed from sample code (multiDicSearch).input.txt) for sample code (multiDicSearch) has been revised. tools/roman/Roman_Manual_jp.txt.iWnn_1_0_7.)multiDicSearch).nj_strlen function with the internalStrlen function in the sample code (multiDicSearch).multiDicSearch).input.txt) to "Account".KO_font and HK_font folders from the dictionary package.iWnn_1_0_6.)type argument to the NjxCreadDic function has been changed from ranging from 0 to 4 to accepting the values given below. NJ_CREATE_DIC_TYPE_USER definitionNJ_CREATE_DIC_TYPE_USER_ADDITIONAL definitionNJ_CREATE_DIC_TYPE_LEARN_AWNN definitionNJ_CREATE_DIC_TYPE_LEARN definitionNJ_CREATE_DIC_TYPE_LEARN_ADDITIONAL definitionnn::fs::MountSdmc function call was added to each sample code. NN_OS_DEVICE_MEMORY_SIZE found in each sample code has been replaced with the nn::os::GetDeviceMemorySize function.morpho). o A mistaken comment was revised in sample code (multiDicSearch). o An error in the return values of the internalCreateDic and internalCheckDic functions used in the sample code (multiDicSearch) has been revised. o Deleted libsample_cconv.*.a and changed specifications so each sample code log is output in UTF-16LE. iWnn_1_0_5.)libmw_iwnn_m.*.a) for expanding all dictionaries in memory.multiDicSearch). NjxCheckExtensionArea functionNjxGetExtensionAreaSize functionNjxInitExtensionArea functionNJ_ERR_PARAM_EXTENSION_AREA_NULL definitionNJ_ERR_PARAM_EXTENSION_AREA_INVALID_SIZE definitioniWnn_1_0_4.)libmw_iwnn_m.*.a) for expanding all dictionaries in memory. iWnn_1_0_3.)multiDicSearch).analyzeConv, morpho, multiDicSearch, and otherAPIs). o Added a process to sample code (analyzeConv) for checking the actual number of candidates retrieved and the candidate index numbers to be confirmed. o Revised a mistaken comment in sample code (multiDicSearch) (changed "WLF" to "MULTI_CLASS") and deleted unnecessary processing (nn::fs::FileStream::Seek). o An illegal frequency setting in the single kanji dictionary was corrected in sample code (multiDicSearch). o Added information on how to distinguish FLASH and non-FLASH dictionaries to Section 7.1 of the iWnn API Manual. o Revised the format of the descriptive text for NJ_MAX_USER_COUNT in Section 3.1 of the iWnn API Manual.
OMRONSW and the module name to iWnn_1_0_0.)CTR_MW-iWnn" and the environmental variable for the root folder was changed to "CTRMW_IWNN_ROOT".noopt/verbose/release and fast/small).multiDicSearch sample program.roman folder (for sample code that converts kana entered in roman characters) has been added to tools.iwnn.h.NJ_TYPE_FULLNUMERIC_READING→NJ_TYPE_FULL_NUMERIC_READINGNJ_CATEGORY_MONTH_FEBRUAY→NJ_CATEGORY_MONTH_FEBRUARYiwnn.h were changed. NJ_FUNC_ALL_CANDIDATES:NJ_FUNC_ALL_ZENKOUHO→NJ_FUNC_ZENKOUHOiwnn.h. NJ_GET_*(s):NJ_GET_*((NJ_WORD*)s)→NJ_GET_*(((NJ_WORD*)s))NJ_GET_*(f):NJ_GET_*((NJ_WORD*)f)→NJ_GET_*(((NJ_WORD*)f))NJ_SET_*(s,v):NJ_SET_*((NJ_WORD*)s,v)→NJ_SET_*(((NJ_WORD*)s),v)iwnn.h.NJ_PARTS_OF_SPEECH_*_*NJ_STATUS_SEARCH_*NJ_WORD to IWNN_WORD.NJ_SEARCH_LOCATION_SET to IWNN_SEARCH_LOCATION_SET.NJ_SEARCH_LOCATION to IWNN_SEARCH_LOCATION.NJ_FZK to IWNN_ANCILLARY.NJ_DIC_FREQ to IWNN_DIC_FREQUENCY.NJ_HINDO_IWNN_FREQUENCY to IWNN_FREQUENCY.IWNN_NUMERIC_FORECAST_DIVISION_INFO.NJ_GET_TYPE_FROM_STEM to NJ_GET_WORD_TYPE_FROM_STEM.NJ_SET_TYPE_TO_STEM to NJ_SET_WORD_TYPE_TO_STEM.otherAPIs as a sample of various other functions. MultiDicSearch provided in the previous release.CONFIDENTIAL