JPH04116764A - Document processor - Google Patents

Abstract

PURPOSE:To alter the break of a clause becoming the object of reconversion by instructing a reconversion instruction means in a state where a clause display index is displayed on a decided clause and altering the reading of the clause which the clause display index shows. CONSTITUTION:A character string and an instruction, which are inputted from an input part 1, are given to a main control part 2 and are transmitted to a KANJI (Chinese character) conversion processing part 4, are written into a text memory 13 through a text processing part 12 and are outputted to an output part 15 through a display control part 14. The KANJI conversion processing part 4 stores the input character in an input buffer 9, executes a KANA (Japanese syllabary)-KANJI conversion preprocessing such as the retrieval of the dictionary 11 and the check of grammar and writes a conversion result in an output buffer 10. The content of the output buffer is written into the text memory 13 through the text processing part 12 and is outputted to the output part 15 through the display control part 12 and is outputted to the output part 15 through the display control part 14 as the conversion result. Thus, reconversion can be executed based on the alteration of the break of the clause.

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Field of Industrial Application The present invention relates to a document processing device such as a Japanese word processor.

(b) Conventional technology In conventional document processing devices, the pronunciation of the character to be converted into a character string containing kana and kanji as shown in Figures 1 and 3 (a) is inputted, and the reading of the character is input as shown in Figures 1 and 3 (b). Obtains the display of the conversion result (unconfirmed state) after receiving the conversion instruction to a character string,
Next, when creating a document by the process of determining the conversion result as shown in (d), in the undetermined state of (b),
As shown in Japanese Unexamined Patent Application Publication No. 63-89969, it is possible to output the correct conversion result by using a re-conversion function ((C) in the same figure) by changing the reading divisions of phrases. Ta.

However, after the conversion result is confirmed as shown in (d), the characters are retyped, the reading is developed again for the conversion result, and the conversion is performed again, as shown in (e) to (h) of the same figure. or the cursor NC (hereinafter referred to as "cursor NC" used for normal character input)
The correct answer was output by converting the homophones of the word containing the character indicated by the cursor (usually called the cursor).

(c) Problems to be Solved by the Invention Therefore, when reconverting a determined character string as shown in FIG. 13(c), reconversion cannot be performed by changing the bunsetsu break. Therefore, if a conversion error in Kana-Kanji conversion is not limited to an error in homophones, but also a conversion error that spans several phrases due to a misrecognition of phrase breaks, conventional reconversion methods can reduce the amount of operations required for correction. The burden on Ta was heavy. The present invention aims to solve the problem that when conversion errors are made over several phrases due to erroneous recognition of phrase breaks, conventional reconversion methods impose a heavy burden on the operator for correction.

(d) Means for Solving the Problems In order to solve the above problems, the present invention provides an input means for inputting a kana character string having reading information, and a conversion method for converting the input kana character string into a character string containing kanji and kana. means, a document storage unit that stores the character code of the character string created by the conversion means and an attribute corresponding to the character string, and a display means that displays the kana character string and text input by the input means; a display indicator that indicates the position, etc., of displaying characters input on the display means; a kanji/reading storage unit that stores portions of character strings on the text that have been converted into kanji and their pronunciations; a recognition means for recognizing a phrase containing a confirmed character on a text to be recognized, a phrase display indicator indicating a character string that is a recognized phrase, and instructions for displaying phrase recognition processing by the recognition means and the phrase display indicator; display instructing means for instructing the reconversion, means for reconverting by changing the division of pronunciations of the phrase indicated by the phrase display indicator in the mode in which the phrase display indicator is displayed; and instructing means for instructing the reconversion. and when the re-conversion instruction means is instructed while the bunsetsu display indicator is displayed on a fixed bunsetsu, the reading of the clause indicated by the bunsetsu display indicator is changed, It is characterized in that several phrases below the phrase display index are converted as conversion targets.

(E) Effect When the reconversion instruction means is instructed while the phrase display indicator is displayed on the phrase determined as described above, the reading of the phrase indicated by the phrase display index is changed. As a result, the break between clauses to be reconverted is changed.

(f) Example An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a document processing device that implements the present invention. In the same figure, an input unit 1 is composed of a keyboard as shown in FIG. 2 through which characters are input and instructions such as kana-kanji conversion, phrase cursor display, reconversion, etc. are input. The character string and instructions input from the input unit 1 are given to the main control unit 2, and if the mode determination unit 3 determines that the character input is valid, the main control unit 2 converts the given character string to kana-kanji conversion. Send to section 4. The given character string is written into a text memory 13 via a text processing section 12, and outputted via a display control section 14 to an output section 15 such as a CRT display.

The kana-kanji conversion processing unit 4 stores the input characters in an input buffer 9 at the same time as a character string (a kana character string to be converted into a kana-kanji mixed character string) is input, and performs searches in a dictionary 11, grammar checks, etc. Pre-conversion processing is performed, and when a conversion instruction such as kanji conversion is issued from the input unit 1, the conversion result is written to the output buffer 1°. Alternatively, if characters are input consecutively and it is determined that the characters can be converted and outputted by the pre-processing for Kana-Kanji conversion, the conversion results are written to the output buffer 10 even if there is no conversion instruction. The contents of the output buffer are written to the text memory 13 through the text processing section 12, and are written to the output section 1 through the display control section 14.
5 as the conversion result. In the present invention, the above-described kana-kanji conversion process is referred to as "natural conversion", and an input mode that performs natural conversion is referred to as natural conversion mode. In addition, for words registered in the dictionary 11 and written to the text memory 13 as a conversion result, the kanji/reading table processing unit 16 registers the word notation and its pronunciation in the kanji/reading table 19 through the registration unit 17. The first character of the registered word has a character attribute indicating the starting position of the reading information (the flag of the attribute ATT in the text memory is turned on).

[Kana-Kanji conversion and Kanji/pronunciation registration processing in natural conversion mode] Figure 4 shows the case where continuous input of kana characters such as ``Tekaki mo ninshiki sutemu no kai hatsu ni jujisuru'' is performed. The process of natural transformation of is shown as an example of output to the output unit 15. Further, FIG. 5 shows the corresponding character code and attribute ATT on the text memory 13. First, (a) schematically shows the display on the output section 15 and the state of the text memory 13 at the time when "Tegakimojininshiki" is input. For input character strings that have not been subjected to kana-kanji conversion, an underline is displayed to indicate that the character strings are subject to kanji conversion. Continuing on, “Sistem”
When you input , the existing conversion processing means is used to convert the part of the input kana character string to ``tekakimo'' and it is determined that it can be output, and ``handwritten mo'' is output as the conversion result. Determined by At this time, the word "handwritten" registered in the dictionary 11 is registered in the kanji/pronunciation table 19 (FIG. 6(a)), and the word "handwritten" in the text memory 13 is
The flag of the attribute ATT for "hand" is turned on as shown in FIG. 5(b).

Continue to input "nokaihatsu ni" (Figure 4 (c))
Then, the part ``jininshi'' is converted into kanji, and it is determined that it can be output, and ``resignation'' is determined as the conversion result and output. As a result, the word "resignation" is registered in the kanji/pronunciation table 19 as shown in FIG. 6(b), and the attribute ATT flag for "hand" on the text memory 13 is turned on (see FIG. 5(C)). ).

Similarly, when you enter "Jiyuushisuru" (Figure 4 (d)), the "kisistem" part is converted to "period", "discard", and so on.
"Mu" is registered in the kanji/reading table 19 (FIG. 6(C)). Finally, enter the period “.” in Figure 4 (e).
The part ``gaihani jujisuru'' is converted, and ``development'' and ``engagement'' are registered in the kanji/reading table 19 (Figure 5 (d)). Therefore, the final conversion result, text memo January 3, kanji/reading table 19 is 4.5.6
(e), (e), and (d) in the figure, respectively.

[Phrase cursor display function for determined character string] When an instruction for displaying the phrase cursor TC is input from the input unit 1, that is, when the “Phrase cursor” key 1c in FIG. It is determined whether the instruction is valid or not, and if it is determined that the instruction is valid, the phrase recognition processing unit 5
Start. The phrase recognition processing section 5 searches the text for the character indicated by the normal cursor NC and the characters preceding it, for which the flag check section 7 determines that the flag of the attribute ATT is on. At this time, the number of characters 'countl' from the position of the normal cursor NC to the character position where the flag is on is counted. Also, the specified number of characters (for example, 30 characters) is read from the text memo January 3 and stored in the text buffer in the various buffer tables 2° (the characters whose flag is on are If the search fails or the corresponding character is before the specified number of characters (for example, 15 characters), the normal cursor NC
Starting from the position, the specified number of characters are read from the text memo January 3 and stored in the text buffer). Furthermore, the reading for the character string stored in the text buffer is developed by searching the kanji/reading table 19 through the search unit 18 and stored in the reading buffer. At this time, attribute 1- of the first character of the text buffer
If the ATT flag is on and the word is retrieved from the Kanji/Yomi table 19, the number of characters in the word is
Count unt2'.

At the same time as developing readings, we also create a word break table. The word division table indicates the word notations of character strings stored in the text buffer and the divisions in their pronunciations. For example, for the contents of the text buffer shown in FIG. ” and “idea” are registered), and the word break table is as shown in (C). In other words, words with pronunciations are considered as one unit in their character string, and other characters are considered as one unit, and the word division is counted as the number of characters (hexadecimal number) from the beginning of the text buffer. This is a break table.

Next, the clause recognition processing unit 5 recognizes a character string that becomes a clause based on the contents of the text buffer, instructs the cursor processing unit 8 to display a clause cursor, and changes the input mode from "natural conversion mode" to "passage cursor". Change to mode j. Also, store the number of characters and readings of the recognized phrase in the phrase index buffer.

A flowchart of phrase recognition is shown in FIG.

1-13 Is the normal cursor NC on a word whose pronunciation is registered (is course+tl<=count2)?
.

1-2. r-yomi registered word” + “attached word (hiragana character string with attribute ATT flag off)”
” is a phrase.

1-31 Is the normal cursor NC located on a hiragana whose attribute ATT flag is off?

1-48 Is there a word registered in reading immediately before the hiragana string containing the hiragana at the normal cursor NC position?

1-5. Only hiragana strings are used as clauses.

1-69 Is the normal cursor NC placed on katakana or alphanumeric characters?

1-7. r Katakana string” or “alphanumeric string” + “adjunct word” as a clause.

1-83 Only the characters (symbols) indicated by the normal cursor NC are designated as clauses.

If we take the example of FIG. 7, the recognized phrases will be as shown in FIG. 8.

[For the character string after the phrase cursor display function has been determined (Fig. 4 (e)), input an instruction to display the phrase cursor TC by moving the normal cursor NC over "hand" (n(f)). If you go, the character ``
``hand'' is searched, and the text is read from that position and stored in the text buffer. The text buffer has a character code and an attribute ATT like text (FIG. 9(a)). Next, write the characters whose attribute ATT flag is on to the kanji/reading table 19 (9th
For characters that are not registered in the kanji/reading table 19, that is, kana characters, the reading is developed from the character code on the text buffer (“mo” → "Mo", "shi" → "shi", "no" → "
",""ni" → "ni", "suru." → "suru."). The result of reading and expanding this example is shown in FIG. 9(c). Further, the word division table is as shown in FIG. 9(d).

Here, the number of returned characters 'countl' from the normal cursor NC position to the character position where the flag is on is O, and the number of characters 'count2' of the word whose reading is registered at the beginning of the text buffer is 3. , it is determined that the normal cursor NC is on a word with reading,
Based on the flowchart in Figure 3, it recognizes that "words whose pronunciation has been registered" + "hiragana"("handwritten" + "mo") is a phrase, displays the phrase cursor TC, and changes the mode ( (Fig. 4(g)) - [Re-conversion function by changing the reading break of a clause in the bunsetsu cursor mode] When an instruction for conversion by changing the reading break of a clause is input from the input unit 1, the main control unit 2 The mode judgment section 3 is instructed to judge the mode, and if it is the bunsetsu cursor mode, the re-conversion clause information setting processing section 6 is activated, and the characters and pronunciations of the clause indicated by the clause cursor TC and several clauses after it are read. Recognizes based on the information stored in the text buffer, reading buffer, and word break table, and
It is stored in the clause information buffer in table 20. The clause information setting processing unit 6 sets the clause information according to the following rules.

(]) Refer to the text buffer, reading buffer, and phrase index buffer to set the character code, reading code, and number of characters of the character string indicated by the phrase cursor TC.

(2) For the second and subsequent clauses, if the attribute ATT flag of the first character of the corresponding clause is on, "one unit of the string containing that character and indicated by the bunsetsu break table" + "and The following hiragana character string is considered to be one phrase.

(3) If the flag of attribute ATT is not on, that one character is treated as one clause.

Feng then instructs Kana-Kanji conversion process 4 to re-convert by changing the division in the pronunciation of the clause, and refers to the content of the change in division, and performs the kana-kanji conversion according to the dDi & clause division of the first clause within that range. conduct. After conversion - is
Displays the phrase cursor TC on the conversion result of the first phrase (the conversion result for the reading that is extended or shortened by one character), updates the contents of the phrase index buffer to correspond to the phrase cursor movement key after conversion, The reading of the new conversion result is registered in the kanji/reading table 19.

[Re-conversion function by changing the pronunciation break of a clause] Suppose that from the state shown in Fig. 4 (g), a re-conversion is instructed in the direction of shortening the pronunciation of the clause by one character (in Fig. 2, keys 1b and 1d are pressed at the same time). ). Text buffer, Kanji/
The reading table 19, reading buffer, and word break table are as shown in FIG. 9, and the phrase information setting processing section 6 first sets the information of the first phrase. Next, since the attribute ATT flag for the character "ji" at the beginning of the second clause is on, one unit of the character string "resignation" determined from the word break table and the continuous hiragana character string "shi" are on. is determined to be the second clause. Next, the attribute ATT for “period” which is the first character of the third clause.
flag is also on, it is determined that the character string unit ``period'' and the continuous hiragana string ``shi'' are the second clause, and the clause information buffer is set as shown in Figure 11 (in this example In this case, a total of three clauses, the clause at the clause cursor TC and the last two clauses, are to be converted).

The kana-kanji conversion processing unit 4 calculates the following from the bunsetsu information shown in FIG.
Kana-kanji conversion is performed assuming that the bunsetsu break of the first bunsetsu is between "tekaki" and "moshi ninshikishi", and the conversion result "handwritten character recognition" is stored in the output buffer and displayed ( Figure 4(h)). Newly register "character" and "recognition" in the kanji/reading table 19. Final text memory 13, kanji/reading table 19, Figure 5 (f
), as shown in FIG. 6(f).

[Phrase cursor TC movement function in phrase cursor mode] When a phrase cursor movement key instruction is input from the input unit 1 (when either cursor movement key 1d or 1e in FIG. 2 is pressed), the main control unit 2 instructs the mode determination section 3 to determine the mode, and if it is the phrase cursor mode, gives an instruction to the phrase recognition processing section 5 to recognize the phrase in the specified direction of movement. The phrase recognition processing unit 5 performs the same processing as described for the phrase cursor TC display function, assuming that the normal cursor NC is located at the position immediately before the beginning or immediately after the end of the currently displayed phrase cursor TC. Perform processing.

However, if a character is recognized according to the recognition rule and falls within the range indicated by the previous clause cursor TC, the display range of the clause cursor TC after movement is set to a part that does not include that character. At the same time, the contents of the clause index buffer are updated, the specified number of characters after the clause cursor TC are stored in the text buffer, the pronunciation is expanded to the pronunciation table, and the word break table is created.

In the example of FIG. 4(h), if an instruction is given to move the cursor to the right, the clause cursor TC display function is executed assuming that the normal cursor NC is on "sentence".

In this example, since the normal cursor NC is on a character whose attribute ATT flag is on, the flow shown in Figure 3 determines that the "character" part is a clause, and the previous clause cursor TC is Since there are no characters that overlap with the characters in the indicated range, display the phrase cursor TC at the character J. The text buffer, reading buffer, and word break table are as shown in FIG.

FIGS. 4(j) and 4(k) show the results of moving the clause cursor TC and converting the reading of the clause.

[Phrase reading expansion function in phrase cursor mode When an instruction for reading and expanding a phrase is input from the input section 1,
The main control unit 2 instructs the mode determination unit 3 to determine the mode,
If it is in the bunsetsu cursor mode, the bunsetsu cursor TC displays the readings for the number of bunsetsu readings stored in the bunsetsu index buffer among the readings expanded in the reading buffer.
is displayed from the beginning and sent to the kana-kanji conversion processing section 4. Also, exit the phrase cursor mode and change to the conventional reconversion mode (``Phrase cursor mode'' → ``Successive conversion mode'').

In the example in Figure 4, the clause cursor T
When C is above "Recognition", when you instruct the pronunciation development of a clause (by pressing key 1a in Figure 2), the number of pronunciations stored in the clause index buffer among the pronunciation codes in the pronunciation buffer (in this example Now, just 4), get the reading code and display it. Therefore, the result after reading and development is as shown in FIG. 4 (1).

(g) Effects of the Invention As described in detail above, according to the present invention, it is possible to change the segmentation of phrases to be reconverted even for character strings that have been finalized. Retransformation can be performed based on changes in the break.

[Brief explanation of the drawing]

1 to 12 show one embodiment of the present invention, FIG. 1 is a block diagram showing the configuration, FIG. 2 is a schematic diagram showing an example of an input section, FIG. 3 is a flowchart, and FIG. 4 is a block diagram showing the configuration. Fig. 5 is a schematic diagram showing changes in the display state. Fig. 5 is a schematic diagram showing changes in the text memory state. Fig. 6 is a schematic diagram showing changes in the state of the kanji/yomi table. Fig. 7 is a schematic diagram showing changes in the text buffer and reading bar. Figure 8 is a schematic diagram to explain the normal cursor state and clause cursor state. Figure 9 is the text buffer, reading buffer, and word division buffer. Figures 10 and 11 are schematic diagrams to explain the clause information buffer, and Figure 12 shows the states of the text buffer, reading buffer, and word break buffer after changes. FIG. 713 shows a schematic diagram for explaining a conventional example. DESCRIPTION OF SYMBOLS 1... Input unit, 2... Main control unit, 4... Kana-Kanji conversion processing unit, 5... Clause recognition processing unit, 6... Re-conversion clause information setting processing unit, 7... Flag check Part 8...
- Cursor processing section, 11... Dictionary, 12... Text processing section, 13... Text memory, 15... Output section, 16... Kanji/reading table processing section, 19...
Kanji/reading table, 20...Various buffer tables,
NC: Normal cursor, TC: Clause cursor.

Claims (1)

[Claims] (1) An input means for inputting a kana character string with reading information, a conversion means for converting the input kana character string into a character string mixed with kanji and kana, and the character code of the character string created by the conversion means and its characters. a document storage unit that stores attributes corresponding to the columns; a display unit that displays the kana character string and text input by the input unit; and a display index that indicates the position, etc., where the input characters are displayed on the display unit. a kanji/reading storage unit that stores a part of a character string on a text that has been converted into a kanji and its pronunciation; and a recognition means that recognizes a clause that includes a confirmed character on the text indicated by the display indicator; a phrase display indicator indicating a character string that is a recognized phrase; a display instruction means for instructing the recognition means to perform phrase recognition processing and display the phrase display index; means for reconverting by changing the reading division of the phrase indicated by the phrase display indicator;
and instruction means for instructing the reconversion, and when the reconversion instruction means is instructed in a state where the phrase display index is displayed on a confirmed phrase, the phrase display index indicates A document processing device characterized in that a reading of a phrase is changed and several phrases below the phrase display index are converted as conversion targets.