JPH0797369B2 - Kana-Kanji conversion device - Google Patents

Description

The present invention relates to a kana-kanji conversion device, and more particularly to a kana-kanji conversion device used in, for example, a Japanese word processor for performing phrase conversion or the like.

[Conventional technology]

In a Japanese word processor or the like, a idiom dictionary in which the reading data and each kanji data are registered in association with each other is used when converting the input reading into kanji.
The reading data registered in the dictionary is usually "aiueo ...
・ In order to speed up dictionary search, a reading index indicating the start address in the dictionary by the first character of reading etc. is generally provided.

FIG. 7 shows a configuration example of a conventional phrase dictionary.

A conventional reading index is, for example, as shown in FIG. 7 (a), a reading start address starting with "a",
A pointer to reading and kanji data is registered with a character or a character string including the first character of reading as an item, such as a reading start address starting with “i”.

The contents of the dictionary are as shown in FIG. 7 (b), for example. Following the reading, the number of kanji and the idiom corresponding to the reading are arranged by the number of homonyms, and the end code indicates the end. After that, the data of the next reading follows.

[Problems to be solved by the invention]

According to the conventional method shown in FIG. 7, the reading length in the dictionary is variable, so when searching for reading data that matches the input reading, the end code is found and You need to find the reading. That is, it is necessary to check the end code one byte at a time and follow the reading, and there is a problem that it takes time to retrieve the reading.

It may be possible to provide a pointer to the next reading for each reading data in the dictionary so that the end code does not need to be checked, but the storage capacity of the dictionary increases,
There is also a problem that the process of registering a new reading in the dictionary becomes complicated.

[Means for solving problems]

If the reading length is fixed and readings of the same length are arranged consecutively, it is not necessary to search for the end code when searching for the next reading from one reading, and it is simpler. It is considered that reading can be sequentially searched at high speed by the address addition processing. The present invention has been made with this point in mind. First, the dictionary is configured to have reading data and kanji data separately,
And the reading data is grouped by reading length. Then, the items of the reading index are classified according to the characters or character strings including the first character and the reading length, and the corresponding reading data is pointed to by the reading index table for each classified item.

When searching for a reading that matches the input pronunciation and obtaining conversion candidate Kanji data, refer to the reading index table based on the character or character string that contains the first character of the input reading and the reading length, Get start address of read data grouped by read length. Then, while updating the start address according to the reading length, the reading data is extracted, the readings are compared, and desired kanji data can be selected at high speed.
An embodiment will be described below with reference to the drawings.

〔Example〕

FIG. 1 is a configuration of an embodiment of the present invention, FIG. 2 is a configuration example of a reading index, FIG. 3 is a configuration example of an idiom index, FIG. 4 is a configuration example of a reading table, and FIG. An example, FIG. 6 is a process explanatory view of an embodiment of the present invention.

In FIG. 1, 10 is an idiom dictionary, 11 is a reading index table, 12 is a reading table in which reading data is registered, 13 is an idiom index table, 14 is an idiom block in which kanji data is registered, and 20 is a reading. Keyboard, 21 is an input control unit, 22 is a reading input buffer, 23 is a reading character string extraction unit that extracts a reading character string from the data of the reading input buffer 22, and 24 is a reading length control that indicates the reading length to be converted. Section,
25 is a Yomi index table search unit that searches the Yomi index table 11, 26 is an idiom index table 1
Jukugo index table search unit for searching 3; 27 is a read table address management unit for managing a search pointer to the reading table 12; 28 is a Jukugo block address management unit for managing a pointer to the Jukugo block 14; 29 is a reading table 1
A reading table retrieval unit that extracts reading data from 2, a reading comparison unit that compares and collates the input reading with the reading in the reading table 12, and 31 is a kanji character in the idiom block 14 that corresponds to a matching reading. A compound word candidate selection unit that uses data as a compound word candidate, 32 is a compound word determination unit, 33 is a display control unit that performs display control on the display, and 34 is a display.

The detailed structure of the reading index table 11 is as shown in FIG. In the reading index table 11, for example, not only the start address for reading the first character "a" but also the start address in which the reading length is taken into consideration are registered. That is, in this embodiment, the maximum reading length is 6 (the number of reading characters is 6 digits), and the first character is "A".
Then, the start address of the read data with a read length of 6, the first character is "A", the start address of a read length of 5, ...,
Start address with the first character "a" and a reading length of 1,
The start address classified by the reading length is set like this. The maximum value of the reading length does not have to be 6, and two or more characters such as "Oh" and "Ai" may be used as the key information of the index, not just the first character.

Similarly, the detailed structure of the idiom index table 13 is as shown in FIG. 3, which also corresponds to the reading index table 11 and is classified in consideration of the reading length. The Yomi index table 11 and the idiom index table 13 may be combined into one table.

The reading table 12 is constructed, for example, as shown in FIG. That is, a set of a reading with a reading length of 6 and its idiom block length is arranged by the number of reading data with a reading length of 6, and then a reading with a reading length of 5 and a set of the idiom block length. Will be continued in sequence. It should be noted that the first character of the reading is fixed by the retrieval from the reading index table 11, and therefore is omitted from the reading table 12 and registered. The idiom block length is the length of the Kanji data group for the homophones and the different words in the idiom block 14.

Since the reading length is known in advance, the reading search in the reading table 12 can be performed by adding a constant to the address.

The phrase block 14 is configured as shown in FIG. That is, the kanji data and the terminator indicating the end of one kanji data are registered in the reading order in the reading table 12.

Next, the processing operation of this embodiment will be described with reference to FIG.
The following numbers ~ correspond to the process numbers shown in FIG.

The illustrated reading character string extraction unit 23 in FIG. 1 extracts a character string having a maximum reading length (6 in this embodiment) or less in the idiom dictionary 10 from the beginning of the reading input buffer 22 and sets it as the first reading.
By sequentially searching from a long character string, the candidate with the longest match is selected first.

From the reading index table 11, the start address of the reading table 12 corresponding to the first character of reading and the corresponding reading length is obtained.

Similarly, the start address of the phrase block 14 corresponding to the leading character of the reading and the corresponding reading length is found from the phrase index table 13 and stored.

The reading data code corresponding to the reading length is extracted from the corresponding address of the reading table 12.

Readings extracted in the above process and reading table
Determine whether the 12 readings match. If they do not match, control is passed to the process.

When the readings match, a group of idiom candidates corresponding to the corresponding idiom block length in the reading table 12 is read from the idiom block 14.

 The read idioms are sequentially displayed on the display 34.

It is judged by the input from the keyboard 20 whether or not it is a correct idiom. When the idiom is confirmed, one conversion process is terminated and the data is passed to the document editing unit (not shown). Or continue the next conversion process or reading input.

When the idiom is not confirmed, it is determined whether or not the display of all idioms corresponding to the reading has been completed. If not, control is returned to the process and the next candidate is displayed. When the display of all idioms is completed, it means that there is no idiom corresponding to the current reading length, so control is transferred to the process.

When the readings do not match in the processing judgment, the address of the phrase block 14 is saved to the pointer.
The idiom block length in the reading table 12 is added to obtain the block address of the next idiom block 14.

The increment value corresponding to the reading length is added to the pointer that has saved the address of reading table 12, and the reading table
Find the next data address in 12.

Check whether the data for the corresponding reading length is complete.

If the data has not ended, control is returned to the process, the read data is similarly extracted, and the process of comparing the read is continued. When the data is complete, perform the following processing.

The reading length control unit 24 shown in FIG. 1 shortens the reading length by one character.

Determine whether the reading length has reached "0". If the reading length is not "0", control is returned to process and the process is repeated.

When the reading length is "0", control is transferred to a processing unit (not shown) when there is no corresponding Chinese character, and the conversion processing is ended.

〔The invention's effect〕

As described above, according to the present invention, the phrase block in which the kanji data is stored and the reading table in which the reading data is stored are separated, and the items of the reading index are classified according to the reading length. The address to retrieve the read data can be updated with a fixed increment value related to the read length, and the conversion target can be retrieved at high speed. In addition, the capacity of the dictionary is relatively small.

[Brief description of drawings]

FIG. 1 is a configuration of an embodiment of the present invention, FIG. 2 is a configuration example of a reading index, FIG. 3 is a configuration example of an idiom index, FIG. 4 is a configuration example of a reading table, and FIG. An example, FIG. 6 is a diagram for explaining the processing of one embodiment of the present invention, and FIG. 7 shows an example of a conventional idiom dictionary. In the figure, 10 is a idiom dictionary, 11 is a reading index table, and 12
Is a reading table, 13 is an idiom index table, 14 is an idiom block, 20 is a keyboard, 21 is an input control unit, 22 is a reading input buffer, 23 is a reading character string extraction unit, 24 is a reading length control unit, and 25 is a reading index. Table search unit, 26 is a compound word index table search unit, 27 is a reading table address management unit, 28 is a compound word block address management unit, 29 is a reading table search unit, 30 is a reading comparison unit, 31 is a compound word candidate selection unit, 32 is a The idiom determination unit, 33 is a display control unit, and 34 is a display.

Claims (1)

[Claims] 1. A kana-kanji that has a dictionary having a reading index whose characters or character strings include at least the first character of the reading, searches the dictionary by inputting the reading, and converts the reading into the corresponding kanji. In the conversion device, the items of the reading index are classified by a character or a character string including the first character and the reading length, and a reading index table (11) that points the corresponding reading data for each classified item. ， The reading data corresponding to the above reading index item is
The kanji data is stored in the variable area in correspondence with the reading table (12) in which the reading length is grouped and arranged for each item and the individual reading data registered in the reading table (12). The searched idiom block (14) and the reading index table (11) are searched based on the character or character string including the first character of reading and the reading length, and the corresponding reading data in the reading table (12) is searched. (25) for acquiring the storage position information of the reading data and means for sequentially comparing the reading data pointed from the reading index table (11) with the input reading while updating the address based on the reading length ( 27,2
9, 30) and a means (31) for extracting kanji data corresponding to the reading data that matches the input reading from the idiom block (14) and selecting it as a conversion candidate (31). Converter.