JPH04167051A - Document editing device - Google Patents

Abstract

PURPOSE:To eliminate the need for similar kinds of repetitive process by an operator by comparing document information at a 1st editing point with document information at a 2nd editing point which is obtained by an operator's process and utilizing the result for subsequent editing operation. CONSTITUTION:The document editing device is equipped with a storage means 3 stored with the document information at the 1st editing point as to document information and an extracting means 4 which compares the document information at the 1st editing point with the document information at the 2nd editing point obtained by the operator's process and extracts information for converting the document information at the 1st editing point into the document information at the 2nd editing point. The device is further equipped with a storage means 3 stored with the conversion information extracted by the extracting means 4 and the extracted information is utilized for subsequent editing operation. Consequently, the repetition of similar kinds of process performed by the operator can be avoided and the efficiency of the operation is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention (Field of Industrial Application) The present invention relates to a document editing device used for editing texts.

(Prior Art) Recently, computer technology has made remarkable progress, and word processors that output document information by inputting reading information have become widespread.

By the way, document editing devices are used to edit documents created by such word processors, such as correcting and deleting sentences.

In conventional document editing, the desired process is executed by specifying the operation target and operation command.
Since each of these processes is independent,
For example, when correcting many identical words in a sentence, it takes a lot of time and effort, which is impractical.

Therefore, recent word processors have methods such as deleting strings,
A ``full text'' editing method is being considered in which operations such as replacement are memorized and the same operation is applied to the entire document.

However, since this method applies one operation to the entire text, it cannot be applied to multiple operations. Furthermore, such editing methods that target the entire text are inappropriate for processing that is not suitable for applying one operation to the entire text (for example, when converting the character Ω to the character 1). Ta.

On the other hand, in addition to this "full text" editing method, there is also a query method in which the user specifies the string to be replaced with the string to be replaced, and the system searches for the string and queries the user. There is an editing method that involves string replacement of the type. However, even in this case, it was necessary for the user to appropriately specify one operation.

However, when using such an editing method to proofread a text read by an OCR (optical character reader) used for machine translation, the operator must
Every time CR discovered a word that was misread, it was necessary to correct the word or make a batch correction. In particular, due to the characteristics of OCR, if the letter m is misread as the letter rn,
“systemn+” to “3y6tem”, and “
Like converting ttrne - to "tiIIle-,"
The same process (replacing "rn" with "m") had to be done for many words.

Further, some document editing devices have a function of storing a series of commad strings and combining the processing into one command. This type of function can improve operational efficiency when multiple processes are combined into one and the same process is repeated.

However, this function still stores a series of processing commands for one object, so if there are multiple desired processes, it is necessary to define multiple operation commands, and the operator It was very difficult for people to remember those commands and apply them properly. In particular, it is difficult to apply this method when editing from the beginning according to the text. In addition, the operator had to specify not only the editing command but also the conditions for applying the operation, which required a high degree of skill.

(Problem to be Solved by the Invention) As described above, with conventional document editing devices, it is not possible to appropriately store and use multiple processes that the operator wants to perform, so the operator has to repeat the same type of process over and over again. It was also difficult to edit from scratch along with the text. Furthermore, even with the function of defining multiple commands in one, they require a high degree of skill to use, and have drawbacks such as inefficient editing and complicated operations. .

The present invention has been made in view of the above circumstances, and it is an object of the present invention to provide a document editing device that can be easily operated by an operator and that can improve the efficiency of editing work.

[Structure of the Invention] (Means for Solving the Problems) The present invention is a document editing device that executes editing corresponding to an editing command on text information, the text information at the time of first editing on the text information. A storage means for storing the above-mentioned text information, compares the text information at the first editing time and the text information at the second editing time, and converts the text information at the first editing time to the text information at the second editing time. The apparatus includes an extraction means for extracting information to be converted into an image, and a storage means for storing the conversion information extracted by the extraction means.

The present invention also provides a document editing device that executes editing corresponding to an editing command on text information, comprising a storage means for storing text information at the time of first editing the text information;
The text information at the time of first editing and the text information at the time of second editing are compared, and the information for converting the text information at the time of first editing into the text information at the time of second editing is extracted. It is equipped with an extraction means, a storage means for storing the conversion information extracted by the extraction means, and a search means for making it possible to search for information in a text, and the time when the information is retrieved by the search means is defined as the first editing time and the first editing time. This is configured to be set at at least one of the second editing time points.

Furthermore, the present invention provides a document editing device that executes editing corresponding to an editing command on text information, comprising a storage means for storing text information at the time of first editing of the text information; Extraction means for comparing text information at the time of editing with text information at a second editing time and extracting information for converting text information at the first editing time into text information at the second editing time; a storage means for storing the conversion information extracted by the means; and a search means for searching for text information to which conversion from text information at the first editing time to text information at the second editing time is applied; The system is configured to enable verification of the conversion results of text information retrieved by the search means.

(Operation) As a result, according to the present invention, the text information at the first editing time is stored, and then the text information at the second editing time obtained as a result of the processing by the operator is stored at the first editing time. By comparing with the editing information, information for converting text information at the first editing point to text information at the second editing point can be extracted, and this information can be used in subsequent editing work. Therefore, the operator can avoid repeatedly performing the same type of processing, and work efficiency can be improved. In other words, the processing performed by the operator (even the result of executing a plurality of commands) can be used for subsequent processing, which enables efficient editing work.

(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 the same embodiment. In the figure, reference numeral 1 denotes an input section consisting of a keyboard, etc., and this input section 1 is designed to manually input sentences and various commands to be edited.

Input data input from this input section 1 is sent to a control section 2. The control section 2 is for controlling the entire apparatus, and here, the control section 2 is configured to give control commands to the storage section 3, the verification section 4, the display control section, etc.

The storage unit 3 stores sentences and the like inputted from the input unit 1. The verification unit 4 stores knowledge used in verification processing. Then, the display control unit 5 controls the display unit 6
It controls the display on the input section 1, and displays sentences etc. input from the input section 1.

FIG. 2 is a diagram showing an example of a key arrangement of a keyboard used in the input section 1. As shown in FIG.

In this case, the input unit 1 includes, in addition to character keys (not shown) for inputting text, a search instruction key 100 and edit keys 101 to 1.
04, Cursor movement key 105, Cursor enlargement key: 1
06, cursor reduction key 107, confirmation key 108, cancel key 109, end key 110, learning start key 111
, learning end key 112, next candidate key 113, conversion key 1
14 and a mode change key 115.

Next, the operation of the embodiment configured as described above will be explained.

First, the operation of the control section 2 for controlling the entire apparatus will be explained using the flowchart shown in FIG.

In this case, in the control section 2, step AI, step A4
．． Step A6. Step A8. Step A10. Step A12. In step A14, the key input from the input section 1 is determined.

In step A1, it is determined whether a character key (not shown) on the keyboard of the input unit 1 shown in FIG. 2 has been pressed. Here, if a character key is pressed, it is determined that the operator is inputting the original text. Then, proceeding to step A2, the character code corresponding to the pressed character key is stored in the corresponding storage area of the storage unit 3, and
In step A3, the display control unit 5 displays the information on the display unit 6.

In step A4, it is determined whether the learning start key 111 on the keyboard of the input unit 1 has been input.

If the learning start key 111 is input, step A
In step 5, text information is stored. The text information storage process will be explained in detail later.

In step A6, it is determined whether the learning end key 112 has been input. When the learning end key 112 is input, learning processing is performed in step A7. This learning process will also be explained in detail later.

Here, after the learning start key 111 is input until the learning end key 112 is input, the operator uses other editing functions to correct the text.

In step A8, it is determined whether the conversion key 114 on the keyboard of the input unit 1 has been pressed manually. If the conversion key 114 is input, conversion processing is performed in step A9. The conversion process will also be explained in detail later.

In step AIO, it is determined whether the search instruction key 100 on the keyboard of the input unit 1 has been pressed manually.

If the search instruction key 100 is input, step A
Search processing is performed in ll. The search process will also be explained in detail later.

In step A12, it is determined whether the edit keys 101 to 104 on the keyboard of the human power section 1 have been input. If the edit keys 101 to 104 are input, step A13
, performs processing corresponding to the edit key. For example, if editing such as correction, insertion, or deletion is required during text input, use the cursor movement key 106 to move the cursor to the desired editing location, and then press the insert key among the editing keys 101 to 104. 101), a delete key (102), a back delete key (103), and other keys to perform editing work.

Next, in step A14, it is determined whether or not to end the editing process based on whether or not the end key 110 is input. When the end key 110 is input, in step A15, the text input in step A1 is stored in the storage unit 3, and the editing process is ended.

If any other key is manually operated, the process corresponding to that key is performed in step A16.

Note that other keys include the keys mentioned above and the end key 1.
Refers to something other than 10.

Next, the operation of the control unit 2 in response to inputs from the learning start key 111 and the learning end key 112 will be explained according to the flowcharts of FIGS. 4 and 5.

First, FIG. 4 shows a flowchart of step A5 shown in FIG. 3, that is, text information storage processing.

In this case, in step B1, the character strings before and after the cursor position at the time when the learning start key 111 is input are
Take each character and store them in the consummation string area. This X
The value of is an appropriate string length. In step B2, the cursor position information at the time when the learning start key 111 is input is stored in the original cursor position holding area.

Thereby, the position of the character/word indicated by the cursor in the document can be fixed.

Further, FIG. 5 shows a flowchart of step A7 shown in FIG. 3, that is, the learning process. In this case, in step C1, the position stored in the original cursor position holding area set in step B2 in FIG. A string of X characters before and after the position in the document corresponding to the cursor position at the time of the cursor position) is stored in the current character string area. In step C2, an original word list is created from the character strings in the fill-and-cross character string area stored in step B1 of FIG. 4, and a current word list is created from the character strings in the current character string area stored in step C1. Create.

Here, creation of a word list will be explained.

FIG. 6 shows an example of creating a word string list from character strings. FIG. 6(a) is an example of an English sentence string, and based on this English sentence string, a word list as shown in FIG. 6(b) is created.

For English text, use the word delimiter (blank,
Words are extracted based on tabs, breaks, lines, etc.).

On the other hand, FIG. 6(C) shows an example of a string of Japanese sentences. In this case as well, a word list as shown in FIG. 6(d) is created. In the case of this Japanese text, words are extracted based on character type information (kanji, hiragana, katakana, symbols, etc.). Of course, it is also possible to perform morphological analysis of Japanese to extract words, but in this embodiment, an example of a simple method is used.

Returning to FIG. 5 again, in step c3, both the original word list and the current word list are compared, and the original word group and the current word group are extracted. Here, the source word group and the current word group indicate the difference between the source word list and the current word list, and when the original word group in the source word list is replaced with the current word group, the result is the current word list. have characteristics.

FIG. 7 shows an example of the original word list and the current word list. In this case, the third element in the original word list is “5ystern”, but in the current word list, “s
system”.In this case, the original word group is “
5system” and the current word group is “system”
It is.

Further, FIG. 8 is a flowchart for explaining a method for extracting the original word group and the current word group. In this case, in step D1, the original word group and the current word group are compared from the beginning, and the position where a different word is found is stored as a different word start position. At this time, instead of simply comparing the original word group and the current word group from the beginning, the word matching starting point is searched for while alternately shifting the word comparison positions. For example, if the original word group is r 1:ese, 2:runational, 3:d
description.

4: Language, l, and the current word group is rL:
nction&1.2: descriptionOn, 3:
Language,, J, the first word of the source word group and the word in the current word group, the first word of the current word group and the word in the source word group, the second word of the source word group and the word in the current word group, the second word in the current word group, and the word in the source word group, and so on. The second word “description” is recognized as the starting point for word matching. Then “descrtpt
ton'', the position of the word that differs for the first time between the original word group and the current word group becomes the start position of each different word.

Then, in step D2, it is determined whether or not a word matching point was found in step D1.

Here, if a word match cannot be found, the extraction process ends as a failure. On the other hand, if a -scattering point is found, the process proceeds to step D3, where the original word group and the current word group are compared backwards, and the position where the different word is found is stored as the different word end position. In this process, as in step D1, instead of simply comparing from the end of the word group, the word matching starting point is searched for while alternately shifting the comparison position of the words. Then, the process advances to step D4.

In step D4, the character string from the different word start position to the different word end position of the original word group is stored in the dead character string area. Furthermore, the character string from the different word start position to the different word end position of the current word group is stored in the current character string area.

Through this series of processing, for example, the original word group or r
l:Th1s, 2:operating, 3:5yst
ern, 4:has.

5: multilevel, 6: tirne, 7: sh
aring, 8: mode J, current word group, r
1: Th1s, 2: operating, 3: sys
tem, 4:has, 5:1ulti
1 level, 6: tirne, 7: shar
ing.

8: In the case of model, "5ystern" is stored in the confluence string area, and "system" is stored in the current character string area. When the process of step D4 is completed, the extraction is deemed successful and the process ends.

Returning to the explanation of FIG. 5 again, in step C4 it is determined whether the extraction of the dead character string and the current character string was successful. Here, if the extraction fails, the process ends. on the other hand,
If successful, the process advances to Step C5 and registration in the learning area is performed based on the dead character string and the current character string.

FIG. 9 is a diagram showing the data structure of the learning area.

In this case, the learning area is made up of a set of items consisting of a set of four elements: a matching character string, a target character string, a rule type, and an application condition.

Here, the matching character string is a character string to be replaced. The target string is the string to be replaced. The rule type indicates the type of the rule, for example, “Word”, “char”, “chc
t” takes three values. “word” represents a word replacement rule, “char” represents a partial string replacement rule, and “chet” represents a partial string replacement rule. Indicates that it is a substring replacement rule.The applicable condition is "query""auto"
O"query", which takes two values, queries the operator about the execution (word replacement) of the result of applying the rule, and "auto" automatically executes the result without asking. . The rule type and application conditions will be explained later.

Next, FIG. 10 is a flowchart illustrating in detail the processing in step C5 described in FIG.

In this case, in step E1, the dead string is a ma soting string, the current string is a tarp sot string, the rule type is "word", and the application condition is "query".
Register as a learning dictionary. Here, if it is already in the learning dictionary, duplicate registration will not be performed. Death string is “5ys”
tern" and the current character string is "system", the first item in FIG. 9 is registered.

Next, in step E2, it is checked whether the previous context partial string replacement rule for the dead string and the current string can be extracted. This compares the dead string and the current string, and the string replacement required to convert the dead string to the current string is 1
If the character is a character, and the character before that character exists (is not blank), the character before that character and the replacement character are included in the matching character string and the target character string. For example, if the dead character string is "culture" and the current character string is "culture", the current character string can be obtained by converting "1" in the dead character string to "l". At this time, the previous context string is “u1
'', and the pre-context substring conversion rule is as shown in Figure 4 in Figure 9.
This is the second item. However, for example, if the death string is “vt
” and the current string is “utility”.
In this case, since there is no character before the character “V” to be replaced, the previous context character string conversion rule cannot be extracted.

That is, if the previous context partial character string conversion rule can be extracted in step E2, the process proceeds to step E3, where it is registered in the learning dictionary. On the other hand, if extraction is not possible, proceed to step E4, and check whether or not the post-context string conversion rule can be extracted; if extraction is possible, step E5
Register with. The post-context sub-character string conversion rule corresponding to the example of the pre-context sub-character string conversion rule is the fifth item in FIG.

Then, in step E6, a partial character string replacement rule necessary for converting the dead character string to the current character string is extracted and registered in the learning dictionary. If the required replacement is one character, set the rule type to "char"; otherwise, set it to "chct". Also, if it is a replacement or an entire word, registration is not performed.

Returning to FIG. 5 again, when the registration process in step C5 is finished, the process for the learning end key 112 is finished.

Next, the conversion process (process when the conversion key 114 is input) in step A9 shown in FIG. 3 will be explained.

FIG. 11 shows a flowchart for explaining the conversion process in step A9. In step F1,
The character string from the cursor position at the time when the conversion key 114 is input is extracted as the target character string. Next, in step F2, the rule type wo of the learning dictionary is applied to the character string.
Matching is performed with the matching character string of the rule rd, and if a match is found, the target character string of that rule is set in the candidate list.

In step F3, it is determined whether or not there is an element in the candidate list, and if there is a candidate, the process advances to step F4.

In step F4, the application conditions of the conversion items in the learning dictionary that generated the elements of the candidate list are checked. Applicable conditions are out
Unless there is only one element that is o, the process proceeds to step F5.

In step F5, the elements of the candidate list are presented to the operator, and the operator is asked to select one appropriate element. Then, in step F6, the element selected in step F5 or the element whose application condition is auto recognized in step F4 is replaced with the target character string in the text.

Thereby, a character string in a sentence is converted into another character string using the conversion knowledge registered in the learning dictionary.

On the other hand, if it is determined in step F3 that there is no element in the candidate list, the process advances to step F7. In step F7, the items of rule types chct and char in the learning dictionary are applied to the target character string. For example, if the target string is “tirne”
”, the second rule in Figure 9 is applicable, replacing “rn” with “m” and “tiIIle
” is generated. If there are any other applicable rules, they are applied, and candidates other than “tia+e” are also generated. These candidates are set in the candidate list.

In step F8, it is determined whether the mode is for performing verification processing. This is determined by the operator and
The operator can set whether or not to perform the verification process using a command. When performing verification processing, step F
Proceed to step 9. In step F9, the elements in the candidate list are verified and unqualified elements are deleted from the candidate list.

The verification process may be any verification, but here, a dictionary check for elements in the candidate list will be taken as an example. That is, the dictionary stored in the storage unit is referred to, and it is assumed that the dictionary contains the elements in the candidate list. This can be achieved by the method used for ordinary subrigcinics.

In step F10, it is determined whether the element in the candidate list exists. If there are candidates, step F
If not, a bell sound is emitted in step F11 to let the operator know that no conversion candidate exists, and the process ends.

Through the above processing, for example, when the contents of the learning dictionary are as shown in FIG. 9, when the operator moves the cursor to ttrne' and inputs the conversion key
It becomes possible to convert "ne" into "time".

Next, the search process in step All in FIG. 3 will be explained.

In this search process, replacement targets are sequentially searched for in the text, and replacement candidates are designated/selected for the searched replacement targets using the above-mentioned learning dictionary. Any method or condition may be used to search for replacement targets, but in this example, matching character strings of unknown words (words not registered in existing word dictionaries) and learning dictionary items are used as search targets. (However, if the rule type is word
items) are used.

FIG. 12 shows a flowchart of the search process in step All. In this case, in step G1, the position in the text indicated by the cursor at the time when the search key 100 is input is set as a position pointer. Then step G
In step 2, it is determined whether or not the sentence has ended.

Here, if the end of the sentence has been reached, the operator is presented with "end of search" in step G3, and the process is ended. If not, in step G4, it is determined whether or not the word at the position indicated by the position pointer is an unknown word using the dictionary stored in the storage unit. Matching is performed between the sentence character string from the position and the matching character string of the item in the learning dictionary (items whose rule type is word).

If the word is not an unknown word and the matching fails, the process advances to step G5, advances the position pointer to the next word position, and returns to step G2.

If the matching character string of the unknown word or the learning dictionary item is successfully matched, in step G6,
The learning end process, conversion process, and learning start process that have already been explained are performed. Through this process, when entering step G6, the learning information extracted from the results of operator operations (which may be the result of multiple commands) performed before that point is stored, and the text at that point is stored. The state of is saved by the learning start process. Furthermore, the conversion process performs presentation/selection of conversion candidates, etc.

After this, the process advances to step G7. In step G7, it is checked whether the application condition of the conversion rule (item of the learning dictionary) applied in step G6 is auto. This is au
If it is to, proceed to 51205E. If not, proceed to 81208.

At 51208, it is determined whether the operator has inputted the mode change key 115, and if so, the process advances to step G9. In step G9, the application condition of the learning dictionary item applied in step G6 and whose conversion result was selected by the operator is set to auto. This results in
In subsequent processing, the rules are automatically applied without asking the operator. This completes the process.

Next, a specific example in which the present invention is applied to English sentences will be described with reference to the drawings.

FIG. 13(a) is an initial screen. The cursor is “
Th1s". Nothing is registered in the learning dictionary. When the operator inputs the search key 100 in this state, an unknown word is searched.

The first unknown word is “5ystern”, so the same figure (
The cursor moves to the "5ystern" position as shown in b).

Here, in step G6 described in FIG. 12, learning completion processing is performed, but since nothing is originally set in the character string area, nothing is registered in the learning dictionary. Furthermore, by the learning start process, the character strings before and after the cursor at that time are originally stored in the character string area. In this case, “Th1s is
a 5ystern foroffice vork
, The 5ystern provides var.
ius' is stored. The position of "5ystern" is also stored. In the conversion process, since the learning dictionary is empty, no conversion candidates are generated. The operator uses normal editing commands to change "5ystern' to "sys
Convert to "tem".

Thereafter, when the search key 100 is input again, the next unknown word "5ystern" is searched, and the cursor moves to the second "5ystern" as shown in FIG. 13(c). At this time, learning completion processing is performed in step G6 described in FIG. 12. Originally in the character string area ~”Th1s i
s a 5ystern for of'fice w
ork.

The 5ystern provides varj
ous'. Here, if you extract the current string from the original memorized cursor position, "Th1s is a
system foroffice vork, Th
e 5ystern provides various
s' is obtained. From the filled intersection string and the current string, Figure 5,
Through the processes shown in FIGS. 8 and 11, the first and second items shown in FIG. 9 are registered in the learning dictionary.

In the conversion process of step G6 in FIG. 12, the first conversion rule of the learning dictionary can be applied, and the operator can
system" is presented as a conversion candidate (step F5 in FIG. 11). When the operator selects this, the screen becomes as shown in FIG. 13(d).

At this point, when the operator inputs the mode change key 115, a determination is made in step G8 in FIG. 12, and the process in step G9 is performed.

As a result, the application condition for the first item in the learning dictionary is au
is set to. When the operator inputs a search key in this state, an unknown word is searched. The next unknown word is “d
Therefore, the cursor moves to the position of "docurn'ents" as shown in FIG. 13(e). In the conversion process in step F6 of FIG. , by applying the substring conversion rules, "ctocuments" is presented to the operator as a candidate. When the operator selects this and enters the search key again, the next unknown word is searched and "5ystern" is detected. At this point, the learning dictionary should include “documents” as “docun+en”.
An item to be converted to ts' is registered, and "5ystern" is converted to "system" by the conversion process. At this time, since the application condition of the conversion rule to be applied is AUTO, the search for the next unknown word is started without inquiring the operator, resulting in a state as shown in FIG. 13(f).

Next, a specific example in which the present invention is applied to Japanese text will be described with reference to the drawings.

In this case, FIG. 14(a) is the initial screen. The cursor is in its "natural" position. This document states, "Natural language processing includes morphological analysis, syntactic analysis, semantic analysis, and context analysis. Of these, context analysis is most closely related to reasoning using knowledge." It was input using OCR. Japanese OCR calculates information indicating the validity of the characters it reads, and attaches the information ``uncertain'' to characters or character strings with low validity. Here, the search target is this "uncertain" character or character string. Of course, it is also possible to search for unknown words in Japanese (or search for both "uncertain" characters or character strings and unknown words), just as we did for English sentences. It is. In this example, "ori" in "morphological analysis", "ori" in "context analysis" in the first sentence, and "ori" in "context analysis" in the second sentence are each "uncertain" characters. The information is given. In the initial state, shown in FIG. 14(a), nothing is registered in the learning dictionary.

In this state, when the operator manually presses the search key 100, an uncertain character is searched. Since the first uncertain character is "ori" in "morphological analysis", the cursor moves to the position of "ori" as shown in FIG. 14(b). In this case, the first
In step G6 of FIG. 2, learning completion processing is performed, but since nothing is set in the dead character string area, nothing is registered in the learning dictionary. Also, by the learning start process, the character strings before and after the cursor at that point are saved in the dead character string area. In this case, "natural language processing includes morphological analysis, syntactic analysis, semantic analysis, and context analysis" is stored. The position of the "fold" is also stored.

In the conversion process of step G6 in FIG. 12, no conversion candidates are generated because the learning dictionary is empty.

The operator uses normal editing commands to change the "fold" to "
The state shown in FIG. 14(C) is obtained. after that,
When the search key 100 is input again, the next uncertain character, that is, "fold" of "context analysis" in the first sentence, is searched, and the cursor moves to the second "fold" as shown in FIG. 14(d). Move to ``Fold''. At this time, learning end processing is performed in step G6 of FIG. 12.

The dead string area includes "Natural language processing includes morphological analysis, syntactic analysis, semantic analysis, and context analysis." Extracting the current string from the original memorized cursor position,
“Natural language processing consists of morphological analysis, syntactic analysis, semantic analysis,
Includes contextual analysis. ” is obtained.

Then, from the dead character string and the current character string, the items shown in FIG. 15 are registered in the learning dictionary through the processes shown in FIGS. 5, 8, and 11 described above. In this case, the first item shown in FIG. 15 is a conversion rule for the word list item shown in FIG.

The second item is a partial conversion rule for converting "ori" into "ori". Further, the third item is a pre-context partial conversion rule. In the conversion process in step G6 of FIG. 12, the second and third conversion rules of the learning dictionary can be applied. These rules produce the same result, so
Present "analysis" to the operator as a conversion candidate. When the operator selects this, the screen appears as shown in FIG. 14(e). At this point, the operator presses the mode change key 1.
If 15 is input, a determination is made in step G8 of FIG. 12, and the process of step G9 is performed.

As a result, the application condition of the rule with the strictest application condition, that is, the previous context conversion rule (the third item in the learning dictionary in FIG. 15) is set to auto. When the operator inputs the search key 100 in this state, the next uncertain character is searched. The next uncertain character is "context analysis" in the second sentence.
Therefore, the cursor moves to the position of "fold" in the second sentence "context analysis" as shown in Figure 14 f). 1st
In the conversion process at step G6 in FIG. 2, "Ori" is converted to "Ai" by applying the partial character string conversion rule as before. At this time, since the application condition of the conversion rule to be applied is auto due to the previous processing, "ori" is converted to "analysis" without asking the operator, and the search for the next uncertain word is started.

Note that the present invention is not limited to the above-mentioned embodiments, but can be implemented with appropriate modifications within the scope without changing the gist.

For example, in the above-described embodiment, the present invention is applied to texts in English and Japanese, but it can also be applied to texts in languages other than these.

[Effects of the Invention] According to the document editing device of the present invention, the text information at the first editing time is stored, and then the text information at the second editing time obtained as a result of processing by the operator is stored. Compare with the first editing information, extract information for converting text information at the first editing time to text information at the second editing time,
Since this information can be used in subsequent editing operations, it is possible to avoid the operator from repeatedly performing the same type of processing, and it becomes possible to increase the effectiveness and complexity of the editing operations. In other words, the processing performed by the operator (even the result of executing a plurality of commands) can be used for subsequent processing, which enables efficient editing work. This is particularly effective when the same type of repetitive editing processing is performed, such as in correction processing of OCR-input texts.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention;
FIG. 2 is a diagram showing an example of the key arrangement of the input unit used in the same embodiment, FIG. 3 is a flowchart for explaining an overview of the operation of the control unit in the same embodiment, and FIG. Flowchart showing the flow of text information storage processing in the embodiment, No. 5
Fig. 6 is a flowchart showing the flow of learning processing in the same embodiment, Fig. 6 is a diagram showing an example of a character string/word string list in the same embodiment, and Fig. 7 shows an original word list/current word list in the same embodiment. A diagram showing an example of a word list, FIG. 8 is a flowchart showing an example of processing for extracting dead character strings and current character strings in the same embodiment, and FIG. 9 shows an example of a learning dictionary in the same embodiment. Figure 10 is
FIG. 11 is a flowchart showing an example of the flow of the conversion process in the same embodiment. FIG. 12 is a flowchart showing an example of the flow of the search process in the same embodiment. Flowcharts showing examples, Figures 13 and 1
FIG. 4 is a diagram showing an example of a screen for explaining the operation of the same embodiment, and FIG. 15 is a diagram showing an example of a learning dictionary used for explaining FIG. 14. 1... Input section, 2... Control section, 3... Storage section, 4
... Verification section, 5... Display control section, 6... Display section. Applicant's Representative Patent Attorney Takehiko Suzue Figure 1 Figure 2 Figure 4 Figure 5 Figure 7 Figure 8 Figure 9 Figure 10 ffl Figure 13

Claims (3)

[Claims] (1) In a document editing device that executes editing corresponding to an editing command on text information, a storage means for storing text information at a first editing time point for the text information; and a storage means for storing text information at a first editing time point for the text information; an extraction means for extracting information that compares the text information at the second editing time with the text information at the second editing time and converts the text information at the first editing time into text information at the second editing time; 1. A document editing device comprising: storage means for storing converted information. (2) In a document editing device that executes editing corresponding to an editing command on text information, a storage means for storing text information at a first editing time point for the text information; and a storage means for storing text information at a first editing time point for the text information; an extraction means for extracting information that compares the text information at the second editing time with the text information at the second editing time and converts the text information at the first editing time into text information at the second editing time; a storage means for storing converted information, and a search means for making the text information searchable; A document editing device characterized in that settings are made at the time of editing. (3) In a document editing device that executes editing corresponding to an editing command on text information, a storage means for storing text information at a first editing time point for the text information; and a storage means for storing text information at a first editing time point for the text information; an extraction means for extracting information that compares the text information at the second editing time with the text information at the second editing time and converts the text information at the first editing time into text information at the second editing time; a storage means for storing converted information; and a search means for searching for text information to which the conversion from the text information at the first editing time to the text information at the second editing time is applied; A document editing device characterized in that it is possible to verify the conversion result of text information retrieved by means.