JPS6268325A - Sentence compression and expansion system - Google Patents

Abstract

PURPOSE:To compress a coded sentence by extracting sequentially character code strings having the same number of characters from a stored character code strings and outputting a word code corresponding to the character code string when they are retrieved and coincident. CONSTITUTION:A character code string of a sentence mixed with Kanji (Chinese character) and Kana (Japanese syllabary) desired to be compressed is fed sequentially from an inlet 1 to an input register section 2. The length of the register is selected by 2, 3 or 4 characters by selecting the length of the unit of conversion word. The code string of 2 characters stored in the input register section 2 is fed to a word retrieval section 3 and when a word stored in an electronic dictionary storage section 9 is retrieved and when they are coincident, the content of a word code supplied to the word is read and outputted to an output register section 5 via a compression control section 4. The code stored in the compression code input register section 12 is fed to a code retrieval section 13, a word code stored in the electronic dictionary storage section 9 is retrieved and the 2-character word of the index corresponding to the code is read and fed to an original sentence output register section 15 via an expansion control section 14 from the path 16.

Description

Detailed Description of the Invention (Industrial Application Field) The present invention relates to a method for compressing and decompressing encoded sentences, and is particularly applicable to keyboards, OCR, or file devices in kanji data output equipment or data transmission/reception terminals that handle Japanese. Compress the character code string of the text input from an input device such as
The present invention also relates to a Japanese text compression/expansion method for restoring an already compressed code string to the original input character code string.

(Prior art) Concerning character data compression in character code strings, continuous character compression methods have conventionally been common (for example, "Showa 5
Collected papers from the 5th Annual Conference of the Four Electrical Engineers of Japan [5]”, 1972
Oct. 2015, Ikuma, Nishimura et al., Coding for realizing network protocols, P5-101~P5-102)
. This method uses a compression control code, and when the same character appears consecutively for p consecutive characters, a control character indicating that p characters are consecutive is inserted in front of the character. A method is adopted in which when the sequence ends and each character is changed into a different character string again, another control character is inserted to indicate this.

(Problem to be solved by the invention) However, as is easily understood, this conventional method is effective only when four or more identical characters occur in a row, so the probability that this condition will occur in ordinary documents is low. Therefore, the compression effect is extremely low.

SUMMARY OF THE INVENTION An object of the present invention is to provide a text compression/expansion method capable of efficiently compressing/expanding any text regardless of the continuity of characters, while eliminating the above drawbacks.

(Means for Solving the Problems) The invention set forth in claim 1 (hereinafter referred to as the first invention) includes the following four means.

The first means stores in advance a plurality of character code strings of words each having a predetermined number of characters in correspondence with a word code specific to the word.

The second means temporarily stores a character code string obtained by encoding the text to be compressed character by character.The third means encodes the same number of characters as the predetermined number of characters from the second means. character code strings are sequentially extracted, and a search is made to see if the character code strings match the character code strings of words stored in the tenth means.

The fourth means outputs a word code specific to the word of the character code string obtained from the first means if the search results in the third means match, and if there is no match, the character code The first character code in the string is output as is from the second means.

In addition, the invention set forth in claim 2 (hereinafter referred to as
(referred to as the invention) is constituted by having the following four means.

The first method is to encode a text character by character and, if there is a predetermined word consisting of a predetermined number of characters in the character code string obtained, replace this word with a word code specific to that word. Temporarily stores the obtained compressed character code string.

The second means stores in advance a plurality of character code strings of words each having the predetermined number of characters in correspondence with a word code specific to the word.

The third means extracts each code from the compressed character code string stored in the first means and searches for a match with the word code stored in the second means.

The fourth means outputs the character code string of the word corresponding to the code obtained from the second means if the search result in the third means matches, and if the search does not match, the fourth means outputs the character code string of the word corresponding to the code obtained from the second means. Output as is from the means of.

(effect) The first invention operates as follows.

The character code string obtained by encoding the text is temporarily stored in the second means. The third means sequentially extracts character code strings of the same number as the number of characters of the word stored in the first means from the character code strings stored in the second means. Then, the third means searches whether the word of the retrieved character code string matches a word stored in advance in the first means. The fourth means outputs the word code corresponding to the character code string from the second means if the search results in a match, and if there is no match, outputs the first character code of the character code string. Output as is. As a result, the encoded text is compressed.

The second invention mentioned above is as follows! It works.

The compressed character code string compressed according to the first invention is temporarily stored in the first means. The third means extracts each code from the compressed character code string stored in the first means, and searches for a match with the word code stored in the second means. If there is a match as a result of this search, the fourth means outputs the character code string of the word corresponding to the code obtained from the second means (therefore, the code in this case is a word code), and If not, the code (therefore, the code in this case is a character code) is output as is from the first means. As a result, the compression-encoded text is expanded.

(Example) Prior to a detailed explanation of the present invention, the background of the present invention will be explained.

There are over several thousand types of kanji used in Japanese. Therefore, one character is encoded by a code of 10 bits or more. For example, when encoding with 15 bits, a maximum of 215 (=32.768) types of characters can be encoded, and when using 16 bits, 2'6 (=65.536) types of characters can be encoded. characters can be encoded. Although the present invention can be configured regardless of the number of bits, for ease of explanation, a 16-bit configuration method will be described below.

There are at most 10,000 types of kanji in daily use, including kana and special symbols, so if we adopt the 16-bit construction method, the remaining 50,000 out of 65,536 types of codes will be used. You can use the leftover code for other purposes. On the other hand, in general Japanese documents, compared to the frequency of single-character independent words, adjunct words (e.g., house, water, ya, ga,...) and special symbols, the frequency of occurrence is higher than that of single-character independent words and adjunct words (e.g., house, water, ya, ga,...), and special symbols that are composed of two or more characters. In general, independent words and attached words (e.g., meeting, robot, beautiful, seems, deshita, etc.) appear extremely frequently.

If we take advantage of this property, we can extract a word of two or more characters from a character code string that is input sequentially from an input device, and assign the part of the character code string that makes up that word to a word in advance using an electronic dictionary. The characters are replaced with the assigned 16-pit word code, and the input character codes that are not determined to be - constituents of the word are output as they are without any conversion.
By implementing a conversion mechanism into a word-mixed code string, it is possible to compress sentences using a combination of words.

On the other hand, the decompressing side receives the character/word mixed code string, outputs the single character code as is, and outputs the word code using an electronic dictionary in which the same words as the transmitting side are registered. By replacing it with a character code string of two or more characters and outputting it, it is possible to expand it to the original input character string before compression conversion.

FIG. 1 is a block diagram showing respective embodiments of the first invention and the second invention. In the figure, the part surrounded by a dashed-dotted line A is a compression mechanism, and the part surrounded by a dashed-two dotted line B is an extension mechanism. First, the compression mechanism will be explained. Reference numeral 2 denotes an input register section, into which character code strings of kanji, kana, and kanji texts to be compressed are sequentially supplied to the input port.

The length of the register can be set to 2, 3, or 4 characters by selecting the length of each converted word. For ease of explanation, the length will be explained below as two characters.

The two-character code string placed in the input register section 2 is supplied to the word search section 3, which searches the words stored in the electronic dictionary storage section 9, and if a match is found, it is added to that word. The contents of the word code are read out and output to the output register section 5 via the compression control section 4. If there is no match, the input register section 2 is shifted one character to the left and supplies the next character from the supply port 1, and the leftmost character shifted out passes through the path 7 and is compressed. The signal is sent to the output register section 5 via the control section 5 without conversion and is output.

Now, let us consider a case where the sentence "In response to your request..." is compressed and converted. It is assumed that the electronic dictionary storage unit 9 stores a dictionary whose contents include word/word code correspondences as shown in FIG. The first two characters "Goyone" are placed in the input register section 2. The word search section 3 sequentially reads out the heading sections of the electronic dictionary storage section 9 and compares them with "Ryobin" of the input register section 2. As is clear from FIG. 2, since the word "control unit" is not found, a signal to the effect that it is not found is given to the compression control unit 4, and the first character "" of the input register unit 2 is
``control'' is output to the output register section 5 via the control section 4, and the next character ``desir'' is given from 1, and ``1'' is placed in the input register section 2. This time, as is clear from FIG. 2, ``°request'' is searched, so a signal to that effect is given to the compression control unit 4, and the word code @03# added to ``request'' in the dictionary is sent to route 6. is output to the output register section 5 via the compression control section 4.

Next, the input register section 2 is shifted two characters to the left, and the next two characters "answer" are supplied from the input port 1 and placed in the two-character mark "request" that has been placed so far. Conversion is performed sequentially using the same procedure.

Next, the extension mechanism will be explained. Reference numeral 12 denotes a 16-bit compressed code input register section, to which a compressed code string is sequentially supplied one code (16 bits) at a time from a compressed sentence supply port 11. The code placed in the compression code input register section 12 is supplied to a code search section 3, which searches for word codes stored in the electronic dictionary storage section 9. If there is a matching word code, the two characters of the heading word corresponding to that code are read out and sent from a path 16 to the original text output register section 5 via the expansion control section 4.

To explain the case in which compressed code strings related to the previous example sentence "Answer to both convenience..." are sequentially supplied from the input port 11, the word code for "Go" is not in the electronic dictionary storage unit 9, so the path is 17 to the original text output register section 15 as it is.
is output to. Next, the word code "03#" is in register 12.
supplied to The code search unit 13 searches for word codes stored in the electronic dictionary storage unit 9. If there is a matching word code, the two characters of the heading word corresponding to that code are read out and sent from path 16 to original text output register section 15 via expansion control section ]4. In this case, as is clear from FIG. 2, the word "03" is found, so a signal indicating this is given to the expansion control section 14, and the two characters "request" in the heading section of the electronic dictionary storage section 9 are changed. Read route 1
6 to the original text output register unit 15 via the expansion control unit 14.
The two characters “request” are output. Below, I will not repeat the same procedure, but instead will write down the original text before compression.
...” is extended.

The present invention has been described above based on examples. In the above explanation, the word length stored in the electronic dictionary storage unit 9 was explained as being a fixed length, but for example, if the maximum number of characters of a word stored in the electronic dictionary storage unit 9 is 4 characters, the length of the word stored in the electronic dictionary storage unit 9 is 2 characters, 3 characters, etc. , and 4
The same result can be obtained by accommodating words up to the length of a character, and using a widely known longest matching method to preferentially search for the longest word and replace it with a single 16-bit code. Regarding text compression, words that are not stored in the electronic dictionary storage section 9 are simply output without conversion, and therefore words to be stored in the Seiko dictionary storage section 9 can be freely selected. In particular, if words with high frequency of occurrence are preferentially accommodated, very effective compression can be achieved.

(Effects of the Invention) As described above, according to the present invention, it is possible to provide a text compression/expansion method that can efficiently compress/expand any text regardless of the continuity of characters.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, and FIG. 2 is a diagram showing an example of storage in an electronic dictionary storage unit. 2... Input register section Output... Word search section 4... Compression control section Output register section 9... Electronic dictionary storage section 12... Compression code input register section 13... Code search section 14... Extension control section 15.
・Original output register section

Claims (2)

[Claims] (1) A first means of storing in advance a plurality of character code strings of words consisting of a predetermined number of characters in correspondence with word codes specific to the word; and a method of encoding the text to be compressed character by character. a second means for temporarily storing the obtained character code string; and a character code string for the same number of characters as the predetermined number of characters is sequentially extracted from the second means, and the character code string is stored in the first means. a third means for searching whether the word code matches the word code of the word being searched; and if the search result in the third means matches, a code unique to the word of the character code string obtained from the first means; a fourth means for outputting the word code of the character code string, and outputting the first character code of the character code string as it is from the second means if the two do not match; (2) If there is a predetermined word consisting of a predetermined number of characters in the character code string obtained by encoding the text character by character, it is obtained by replacing this with a word code specific to that word. a first means for temporarily storing a compressed character code string; and a second means for storing in advance a plurality of character code strings of a word consisting of the predetermined number of characters in correspondence with a word code specific to the word.
a third means for extracting each code from the compressed character code string stored in the first means and searching for a match with a word code stored in the second means; If there is a match as a result of the search in the means, a character code string of the word corresponding to the code obtained from the second means is output, and if there is no match, the code is output as is from the first means. A compressed text decompression method characterized by having the following means.