WO2008141583A1

WO2008141583A1 - Character input method, input system and method for updating word lexicon

Info

Publication number: WO2008141583A1
Application number: PCT/CN2008/071027
Authority: WO
Inventors: Zhankai Ma; Lei Yang
Original assignee: Beijing Sogou Technology Development Co., Ltd.
Priority date: 2007-05-22
Filing date: 2008-05-21
Publication date: 2008-11-27
Also published as: CN101051323A; CN100483416C

Abstract

A character input system has an input interface unit, an information transforming unit and a display and output unit, a system lexicon recording a basal word and correlative information, and a cell word set collection to record an expanding word and correlative information. The cell word set is received from at least one of the multiple cell word lexica that are stored by a server end. Word in each cell word set has at least one common attribute. The application provides a dynamic cell word lexicon based on lexicon capability of existing input system, the user can add small word lexicon in a hand or in an automatic manner, so that the words of the user can be covered through individuation selection or a tailor and associated use of an automatic updating and system lexicon, the covering surface of the word lexica is expanded to the most for the user. As the great improvement of the correct radio of the first selection of words for the input system, and the covering surface of the individual word lexicon is expanded to the most in theory, the correct radio of typing is greatly upgraded.

Description

Method for inputting characters, input method system and method for updating lexicon

This application claims priority to Chinese Patent Application No. 200710099474.6, entitled "A Method of Character Entry, Method of Input Method, and Method of Updating the Thesaurus" submitted by the Chinese Patent Office on May 22, 2007. The entire contents are incorporated herein by reference. Technical field

The present invention relates to the field of inputting character information, and more particularly to a method for inputting characters, an input method system, and a method for updating a thesaurus and a thesaurus publishing system.

Background technique

With the popularization and development of computer technology and Internet technology, users of different professional fields, different interests and usage habits are increasingly demanding the intelligence of the input method system.

When evaluating the intelligence of input methods, the accuracy of preferred words is a very important evaluation criterion. At the same time, the ordering of candidates is also very important. In the input method system, the input method lexicon with the term information and word frequency information is one of the important factors affecting both. If the target word required by the user exists in the thesaurus, and the corresponding word frequency information is in line with the user's usage habits, the preferred word accuracy and candidate ordering for the user will be more in line with the demand.

However, the current lexicon of the input method can only cover a part of the vocabulary used by people, usually including some common vocabulary commonly used, and a considerable part of the vocabulary input lexicon cannot be included. Because if all the vocabulary used by the user is added, the lexical capacity of the input method will be on the order of millions. The vocabulary is too large, too many homophones (or too many reweights), candidates are increased, users who do not need to use these words will be seriously interfered; and such a large vocabulary will inevitably occupy CPU, memory and other computing device resources. It is unacceptable for personal computers.

Specifically, when using the input method, everyone enters many common words (such as "now", "time", "how much", etc.), and also enters some niche vocabulary, for example: Some game nouns "Azela ", "Druid", the latest movie "Clouds" and so on. These vocabulary are often entered for small groups, such as: World of Warcraft players, chemistry engineers, biology teachers, and more. However, the use of these terms in the overall user is particularly low, and the input method lexicon in the current model excludes these terms in order to save space and improve efficiency. Therefore, under the prior art, the accuracy of the preferred word when the user inputs the above-mentioned niche vocabulary is very low, which seriously affects the user's use experience and the expression of his thought. In summary, a technical problem that needs to be solved urgently by those skilled in the art is: how to improve the input method vocabulary so that it can satisfy the resource allocation of existing computing devices and greatly improve the input efficiency of each user.

Summary of the invention

The technical problem to be solved by the present invention is to provide a novel input method lexicon mode and a complete set of input solutions, which can meet the resource allocation of existing computing devices, does not occupy more computing resources, and can significantly improve the users' Input efficiency.

In order to solve the above problem, the present invention discloses an input method system, including an input interface unit, an information conversion unit, and a display output unit, and further includes:

System vocabulary for recording basic words and related information;

a set of cell words for recording extended words and related information; the cell word set is obtained by at least one cell vocabulary obtained from a plurality of cell lexicons stored on the server side; words in each cell lexicon The word has at least one common attribute;

The information conversion unit performs a search query in the system vocabulary and the cell word set to obtain corresponding candidates.

Preferably, the input method system further includes: an automatic update module, configured to obtain, by using a comparison, the required update data from the server according to the local existing cell vocabulary list.

Preferably, the type of related information stored in the cell word set is less than or equal to the type of related information stored in the system vocabulary; at least one cell vocabulary exists in the plurality of cell lexicons stored by the server end by artificial Generated manually.

Further, the input method system may further include: a user vocabulary.

Further, the input method system may further include: adding a module, configured to add cell lexicon information acquired from the server end to the cell word set; the cell word set is an independent vocabulary or A collection of thesaurus that exists in parallel for multiple thesaurus. Preferably, the adding process is performed in a separate cache vocabulary.

Further, the input method system may further include: a cell lexicon deactivation module, configured to receive a user instruction, and remove a term record belonging to the cell vocabulary selected by the user from the cell word set.

According to an embodiment of the present invention, a method for character input is also disclosed, including:

Loading a system vocabulary and a cell word set; the cell word set is used to record extended words and related letters The cell word set is obtained from at least one cell vocabulary obtained from a plurality of cell lexicons stored on the server side; the words in each cell lexicon have at least one common attribute;

Receiving input information of the user;

Searching in the system lexicon and cell word set according to the received input information, and obtaining corresponding candidates;

The user's selection information is received, and the specified candidate is output on the screen.

The loading is: combining the cell word set and the system vocabulary into a vocabulary and placing it in a cache; or, the loading is: using the cell word set and the system vocabulary as two or more independent lexicons Placed in the cache and set the thesaurus priority according to the preset rules; the priorities are used for the display ordering of the candidates.

Preferably, the cell word set records the cell vocabulary to which each term belongs and the corresponding cell lexicon priority; the priority is used for display ordering of the candidates.

Further, the method may further include: dynamically adjusting the cell lexicon priority according to the usage environment of the input method during the loading process.

According to another embodiment of the present invention, there is also disclosed a method for updating a thesaurus, the updated thesaurus relating to a set of cell words for recording extended words and related information, the cell word set being from the server side Obtaining at least one cell vocabulary selected from the plurality of cell lexicons stored; each word in the cell lexicon has at least one common attribute; the method comprising:

Receive triggers, compare existing cell lexicon lists and server-side cell vocabulary lists to get a list of vocabularies that need to be updated;

Download the updated cell lexicon entry information and add it to the cell word set.

Further, the method may further include: manually or automatically upgrading the cell vocabulary stored on the server side, and changing the corresponding version information. Preferably, the adding process is performed in a separate cache vocabulary.

According to another embodiment of the present invention, a thesaurus publishing system is further disclosed, including:

The cell vocabulary generating unit comprises: an interface module, configured to receive input information; a generating module, configured to generate a cell vocabulary according to the received information; and an identifying module, configured to specify identifier and version information for each cell vocabulary; , the words in each cell lexicon have at least one common attribute;

The communication unit is configured to receive the trigger and transmit the corresponding cell lexicon entry information to the client.

Further, the cell vocabulary generating unit may further include: a modified update module, configured to modify The new cell lexicon stores information, and notifies the identification module to generate a new version letter for the cell vocabulary. The lexicon publishing system may further include: an identification module, configured to compare the server-side cell vocabulary list and The client's cell vocabulary list, according to the obtained comparison result, transmits the required update data to the client through the communication unit.

Preferably, the cell vocabulary obtained according to the received information stores a plurality of vocabulary information; or the cell vocabulary obtained according to the received information stores index information, and the index information corresponds to other cell lexicons.

Further, the the thesaurus publishing system may further include: a merging module, configured to merge the plurality of cell lexicon entry information into one download vocabulary, and notify the communication unit to transmit the downloaded vocabulary to the client.

Compared with the prior art, the present invention has the following advantages: and the cell word set is composed of two parts, wherein the system vocabulary is still oriented to all users, and the general vocabulary is mainly used, and the cell word set part provides multiple cells through the server side. The vocabulary, where the user selects one or more cell lexicons that best suits them and then combines them. Therefore, it can be guaranteed that the input method lexicon used by the user still belongs to the existing lexicon capacity level, but by the individual selection and use of each person, it can basically cover almost all the vocabulary of the current user, and has relative More accurate word frequency information, which can greatly improve the current user's preferred word accuracy rate, and can also achieve candidate sorting that is more in line with current user usage habits.

The invention realizes a dynamic cell-type vocabulary on the lexicon capacity level of the existing input method, and the user can automatically add a small vocabulary by manual or by computer, through individualized selection or customization of each person, through cell words. The combination of the set and the system vocabulary can cover almost all the words of the current user. This allows the user to enter almost any vocabulary or sentence they need, which greatly increases the accuracy of the preferred word for the input method. In theory, the coverage of the individual's thesaurus is expanded to the maximum, so that the accuracy of typing has a greater improvement.

The invention uses the use of a plurality of cell lexicons, and can update the cell vocabulary by means of automatic upgrade, so that the individual vocabulary can be synchronized with the times. Individuals can maintain the freshness of their vocabulary without hands-on, so that the accuracy of the preferred word for typing can be more obvious in the context of the rapid development of the Internet. Improve the typing speed, reduce the appearance of new words, and reduce the number of page flips.

At the same time, the present invention also provides a thesaurus publishing system, which is used to help each user manually generate a cell vocabulary of a group to which they belong, and update and modify the cell vocabulary; an automatic update function is added on the client side, thereby helping Users get accurate classification of cell lexicon and automatic update of cell vocabulary to keep users consistent with the world and never fall behind.

DRAWINGS

1 is a structural block diagram of an embodiment of an input method system;

2 is a flow chart showing the steps of an embodiment of a method for performing character input;

3 is a structural block diagram of an embodiment of a thesaurus publishing system;

4 is a flow chart showing the steps of an embodiment of a method for automatically updating a thesaurus.

detailed description

The present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.

The invention can be applied to input method platforms of various input modes, including keyboard symbols, handwritten information, voice input and the like. That is, the input information may include an encoded character string, and may also include information such as handwritten input information and voice input, because these input methods also require the use of the thesaurus for candidate sorting. Since the information conversion process in these input methods is a well-known technique, it will not be described in detail herein. The following only takes the code string input as an example for detailed description.

Referring to FIG. 1, an embodiment of an input method system of the present invention is shown, which may specifically include: an input interface unit 101, configured to receive input information input by a user;

The information conversion unit 102 is configured to perform a query in the system vocabulary 104 and the cell word set 105 according to the input information input by the user (for example, receiving keyboard characters), complete the code conversion, and obtain corresponding candidates;

The display output unit 103 is configured to display candidates, and receive user selection, upper screen output.

System vocabulary 104, used to record basic words and related information;

a cell word set 105, representing a set of cell lexicons for recording extended words and related information; the cell word set is obtained by at least one cell vocabulary selected from a plurality of cell lexicons stored at the server side; Words in each cell's thesaurus have at least one common attribute.

In the process of character input, the preset system is used to retrieve the system lexicon and cell word set for encoding. Convert, you can complete the input process that meets the user's individual needs.

The cell lexicon, the specific meaning is a certain common group, a certain person or a part of people using a certain common vocabulary (ie, the words in each cell lexicon have at least one common attribute), for example: Movie vocabulary, the latest song noun library, World of Warcraft lexicon, biological lexicon, Tsinghua University owner noun library, a certain company noun library, Haidian District noun library. The method of obtaining the cell lexicon can be: automatically classifying and parsing the cell vocabulary through a management organization or a server group; or providing a server platform, and manually generating the cell vocabulary of the group described by the user. To better meet the needs of individualized groups. That is, preferably, at least one cell vocabulary of the plurality of cell lexicons in the embodiment is manually generated by a user.

In the prior art, the input method platform can be run on a variety of computing devices, such as personal computers, personal digital assistants, mobile terminal devices, etc., and the present invention can also be applied to various computing devices described above, and its operating environment. There is no need to limit it.

The following is a brief introduction to the process of inputting characters such as Chinese characters, Korean characters, and Japanese, which require encoding conversion. Take Chinese input as an example:

In Chinese, Chinese characters as the basic language unit do not directly correspond to the keys on the keyboard. Therefore, input conversion is required by the input method. First of all, it is necessary to convert Chinese characters into letters, numbers, etc. that can be directly input by Chinese character encoding; usually, the encoding used is pinyin (including various forms such as simple spelling, double spelling, and fuzzy sounding). The user inputs the encoded character string of the Chinese character into the computer through the keyboard (in some cases, the mouse may also be used, such as a soft keyboard); the keyboard input of the user is passed to the input method through the operating system, and the input method is used for decoding; due to different Chinese character sequences (words) , sentences) may have the same encoding, so the input method usually provides a candidate list for the user to choose from.

For example, the Pinyin input method might include the following steps:

a, Pinyin analysis: Split the input string to get pinyin, such as zhuanli today [zhuan] [li]. Of course, sometimes this cut is not unique, such as fangan today [313] [311] or 11] 3⁄4311] (respectively "program" "disgust"). Preferably, the input method can support Jane spelling, allowing the user to input in the following forms: zl, zhl, zhuanl, zhli, ..., etc. Considering that some users are not standard in pronunciation, they can also support fuzzy sounds: zuanli. In addition, it is also possible to adopt a form such as double spelling.

b. Chinese character decoding. The pinyin sequence obtained by segmentation is searched for the corresponding word in the lexicon, or a corresponding sentence is generated by a certain algorithm. c, the user selects the required content, the screen (may also have the process of making words, making sentences). Since different Chinese character sequences may correspond to the same encoding, for a particular encoded string, the input method needs to guess the user's true intent. And to this point, the support of the thesaurus is needed.

For the purposes of the present invention, the thesaurus can contain various language information, such as:

(1) entry

Although the input method can also be constructed on the basis of words, since the word is the smallest nominal unit in Chinese, the modern input method makes extensive use of the term information. For example, when the user inputs the "zhuan" pinyin separately, it is difficult to determine which one of the words "want to earn bricks...". Similarly, when the user enters "li", it is difficult to determine which word in "Li Lili..." he wants to enter. However, if the user continuously inputs the two syllables "zhuanli", it is basically determined that the user wants to input the word "patent". This can greatly improve the accuracy of the input method.

(2) word frequency

There are a lot of homophones, and homophones still exist. In this case, all options can only be listed for the user to choose. However, candidate locations have a large impact on the ease of use of the input method. In general, it is more advantageous for the user to put the more commonly used words in the front position, that is, the word frequency is an important basis for candidate ranking.

In addition, many of the existing input methods integrate the function of automatically constructing sentences. At this time, word frequency information is also an important basis for sentence construction.

The above two language information is indispensable in the input method vocabulary, and the input method vocabulary of the present invention may also include other information that is advantageous for improving the accuracy of the input method, for example:

Language connection. In the process of constructing sentences, in addition to the need to consider word frequency, the input method also needs to consider the connection between words and words. For example, "of" often appears after adjectives, nouns, pronouns, etc., while "ground" often appears after adverbs. In this case, if the user enters "de", it is not possible to see only "of" "ground" which word is higher. The language information required by the input method is stored in the thesaurus, and the user can complete the character input. However, the language information required by different users is not the same. For example:

(1) The terms are different. Almost every industry has its own special vocabulary, these words are in other areas. It is rarely used, and you don't have to worry about constructing the input method lexicon. For example, computer vocabulary "cache" and so on.

(2) The degree of importance of the terms is different. Different users may need to use the same words, but their importance varies from user to user. For example, the homonyms "study" and "smoke and tobacco", the former is used more in the academic field, while the latter is used more in daily life. But both are possible, so when the user enters the pinyin "yanjiu", it will appear in the user's candidate list. Due to their different importance, the relative size of the candidate locations can affect the user's intuitive perception.

The importance of the terms relative to the user can be used in a variety of ways, either alone or in combination, in the thesaurus, for example:

Word frequency information. The word frequency information is usually represented by a number, which is used to indicate the frequency of use of the word; the more frequently the word is used, the higher the frequency.

Word order information. The word order information is usually also a number, but is only used to indicate the relative meaning of the importance of the item.

Or, location information. For convenience, this data can also be omitted, and the relative position of the terms in the thesaurus is used to express the importance of the terms. For example, you can think of words that are in front of the thesaurus more important than words that are listed later, so that the former is placed before the candidate list.

Since the input method lexicon cannot generate a dedicated vocabulary for each user, the present invention proposes to divide the input lexicon into two parts: a systematic lexicon and a cell vocabulary. The system vocabulary is used to record common vocabulary to meet the input needs of most people in most cases, while the individualized needs of a user are recorded by the cell word set. In order to improve the fit of the cell word set to each user, the present invention can generate a large number of cell lexicons manually or automatically, and then each user can select his own cell vocabulary to obtain a cell word set. The cell word set is very good for each user's fit, because the personalized part is chosen by itself.

In the case where the user selects a cell lexicon, the cell lexicon can directly constitute a cell vocabulary.

When a user selects multiple cell lexicons, the cell word set can have multiple representations. For example: (1) on the client side, the plurality of cell lexicons are combined into one lexicon, that is, the cell word set exists in the form of an independent lexicon; the lexicon can store the source of each vocabulary (ie, The cell lexicon) information may or may not be stored. (2) on the client side, storing the plurality of cell lexicons in parallel, that is, The cell word set exists in the form of coexistence of multiple independent lexicons, and the plurality of cell lexicons may be sequentially scanned during the coding conversion. (3) at the client side, merging some of the plurality of cell lexicons (for example, some similar lexicons), that is, the cell word set exists in the form of multiple independent lexicons. But some of these independent lexicons are obtained by combining multiple cell lexicons.

For cell word sets, because some language information is more complicated, for example, language connection, etc., one is difficult to obtain, and the second is difficult to store, so it is preferable for cell word sets (actually including individual cell words). Library), wherein the type of language information stored may be less than the type of language information stored in the system lexicon. Of course, the type of linguistic information stored in the cell word set may also be more than the type of linguistic information stored in the system lexicon. For example, for word order information or location information, it is generally stored in the cell lexicon, and in the system lexicon. Generally not.

Further, the input method system of this embodiment may further include a user vocabulary 106 for recording the user's input habits to better meet the personalized needs of the user.

On the platform provided by the server, there are a large number of cell lexicons, and a large number of users will modify and update these cell lexicons. Therefore, how to provide the latest and best cell lexicon to choose The use of the input method user of the cell vocabulary is also one of the technical problems that the present invention further needs to solve.

Preferably, the embodiment may further include: an automatic update module 107, configured to receive the trigger, and download the required update data from the server according to the existing cell vocabulary list. For example, the user's input method system stores a list of information of the cell vocabulary being applied, and then compares it with the information on the server side. If an update is required, the download update is completed according to the preset update policy. The update data may be the entire cell lexicon. For example, if the cell lexicon needs to be updated, all the vocabulary information of the cell lexicon may be directly downloaded; the updated data may also be in a cell vocabulary. Part of the term information, for example, knowing that the cell lexicon needs to be updated, only the changed term information is downloaded through the lexical comparison. Of course, the server side can also merge the changed term information in multiple cell thesaurus into a new thesaurus as the update data for download.

If the user selects multiple cell lexicons, the server side can merge the multiple cell lexicons into one vocabulary and then send them to the client as a cell word set, that is, the data addition task of the cell vocabulary is completed by the server.

If the user selects multiple cell lexicons, the data for the cell lexicon can also be entered by input. The method is completed by itself. In this case, the embodiment may further include: an adding module 108, configured to add the downloaded cell lexicon information to the cell word set. The adding module 108 can adopt various feasible adding policies, for example, the adding manner is: completing the update to download a cell vocabulary, adding the cell lexicon to the cell word set; or, the adding manner is: After all downloads of the cell lexicon to be updated are completed, they are merged and added to the cell word set.

The add module 108 can be used when the cell word set is first formed, or when the vocabulary is updated. The adding module 108 can be used to download the entire cell lexicon or to download a partial vocabulary information in a cell vocabulary.

Preferably, if the thesaurus adding process can be completed in a short time (for example, no more than 1 second), since the influence is not large, the adding process can be directly inserted into the user's input process. However, if it cannot be completed in a short period of time and may affect the user's experience, the thesaurus addition process should be performed in a separate cached lexicon. In this process, the original vocabulary of the input method is not affected and the user can use it normally. When the cached thesaurus is created, you can directly replace the original thesaurus of the input method. Since this replacement process can be fast, interference to the normal use of the user can be minimized.

Preferably, in order to further improve user management of the thesaurus, the embodiment may further include: a cell lexicon deactivation module 109, configured to receive user instructions (eg, by clicking a menu item, etc.), to remove from the cell word set. A term record belonging to the cell vocabulary selected by the user, for the purpose of deactivating one or some cell lexicons.

The removing process may be: receiving a user instruction, deleting the cell vocabulary selected by the user from the list, and re-adding the cell vocabulary in the list to obtain a new cell word set. Since the deleted cell lexicon is no longer in the list, the newly obtained cell word set will not contain the words therein, and the effect is equivalent to the lexicon has been deleted. For a cell vocabulary that exists independently in a cell's vocabulary, it can be deactivated by directly deleting or adding a delete tag.

Alternatively, the removing process may be: receiving a user instruction to delete a term record belonging to the cell vocabulary selected by the user from the cell word set, wherein the cell word set records the cell lexicon information to which each term belongs. Alternatively, the removing process may be: receiving a user instruction, in the cell word set, adding a deletion mark to the entry record belonging to the cell vocabulary selected by the user, where the cell word set records the cell to which the term belongs Thesaurus information.

That is, the source of each entry is recorded in the large lexicon of the cell word set, when the user specifies to delete a certain The cell lexicon informs the input method system (or it actively) to remove the terms from the cell lexicon from the lexicon. This removal can be done by directly deleting the term from the data structure and releasing its corresponding space, or by deleting the tag. Terms with delete tags will be ignored in subsequent use (no space is freed, but it will be easier to implement). The advantage of this approach is that the system overhead caused by deleting a small number of lexicon entries when the cell vocabulary is large will be relatively small.

Referring to FIG. 2, an embodiment of a method for performing character input is shown. Specifically, the method may include: Step 201: loading a system vocabulary and a cell word set; and the cell word set is composed of a plurality of cells stored from a server end. At least one cell lexicon selected in the lexicon is obtained; the words in each cell lexicon have at least one common attribute;

Step 202: Receive input information of a user.

Step 203: Perform a search in the system vocabulary and the cell word set according to the received input information, and obtain corresponding candidates;

Step 204: Receive selection information of the user, and output the specified candidate on the screen.

One of the more important issues in this embodiment is how to check out candidates when multiple lexicons coexist. The loading process described in step 201 may be: merging the cell word set and the system vocabulary into a vocabulary, and placing it in the cache.

When the input method is started, the system vocabulary and cell word set in the input method system are scanned, and the two are combined into one lexicon and then loaded into the cache, so that the user can follow the system lexicon in subsequent operations. Use directly. The loading of the system vocabulary and the loading of the cell word set can be performed separately. For example, in a simple case, the user only needs to load the system vocabulary, and in some cases, the user selection or input method system is automatically started (for example, In the case of a preset policy, the triggering of the loading of the cell word set is triggered, and then the cell word collection is merged into the system vocabulary and placed in the cache for retrieval upon user input.

Further, the loading process described in step 201 may also be: placing the cell word set and the system vocabulary as two or more independent vocabularies in a cache, and setting the lexicon priority according to the preset rule; The priority is used for display ordering of candidates.

That is, during the loading process, the cell word set is placed in a space other than the system vocabulary, and the cell word set is also retrieved while the system vocabulary is retrieved. Preferably, the priority of the system vocabulary and the cell word set needs to be specified at this time. For example, if the default cell word set has a higher priority than the system vocabulary, when the candidate is output, Words that belong to a cell word set are forced to be placed in front of words belonging to the system lexicon.

For a cell word set to be a large word inventory, there are two separate thesaurus in the cache. While the cell word set is also composed of multiple cell lexicons, there may be multiple independent lexicons in the cache. Of course, it is necessary to set the priority of each lexicon at this time; the priority is used for the display order of the candidates.

Preferably, when the cell word set is a large word stock, in order to reflect the difference of each cell lexicon, the cell vocabulary to which each term belongs and the corresponding cell lexicon priority may also be recorded in the cell word set.

For the case where priority is set for each cell lexicon (including the case where each cell vocabulary is independently existed and merged into one big word inventory), it is preferable that during the loading process, it can be dynamically adjusted according to the usage environment of the input method. Cell lexicon priority. For example, a cell word set includes two cell lexicons, "office language" and "network terminology", which normally have the same priority. However, when the input method system recognizes that the current application is a Word word processing program, the "office language" cell vocabulary can be weighted, and when the user switches to the QQ chat program, the "network term" cell lexicon can be weighted.

Referring to Figure 3, there is shown an embodiment of a thesaurus publishing system suitable for use in the aforementioned input method system (described by the term "input client" for clarity of description), the thesaurus publishing system can be used for input methods The process by which the client first downloads the cell vocabulary from the server side to obtain the cell word set can also be used to update the existing cell vocabulary.

The thesaurus publishing system shown in FIG. 3 may specifically include:

The cell lexicon generating unit 301 includes an interface module 3011 for receiving input information, a generating module 3012 for obtaining a cell vocabulary based on the received information, and an identifier for specifying identification and version information for each cell vocabulary. Module 3013; words in each cell lexicon have at least one common attribute;

The communication unit 302, generally located at the server end, is configured to receive trigger information and transmit corresponding cell lexicon information to the client.

The cell lexicon generating unit 301 is generally located at the server side for uniformly managing and maintaining the cell vocabulary. Certainly, some or all of the modules in the cell lexicon generating unit 301 may also be located in the client (which may be other clients independent of the input method client). For example, the interface module 3011 and the generating module 3012 are located at the client, and the user may directly Send the generated cell vocabulary file to the server Yes, the work of specifying the identification and version information is done by the server.

The triggering information may be a user's selection operation, or may be a trigger information automatically sent by the input method system client, or may be an automatic detection trigger of the server. For example, the server or the client retrieves the user IP address or the current input environment, and automatically recommends the corresponding cell vocabulary to the user; or, the update message sent by the client is also a type of trigger information.

The generation of cell lexicon can be done manually, automatically, etc. The following is a brief description of the process of manually generating cell lexicon:

The thesaurus generator needs to provide the following information through the interface module 3011 (for example, including the lexicon edit page): name, category, number of articles, version, description, thesaurus author, entry examples, terms (including pronunciation information), etc. Wait. When the submit button is clicked, the information is saved to the database; then the thesaurus generator is enabled. At its simplest, the thesaurus program directly saves this information in a file to a file for users to download.

The cell vocabulary is a file containing data that may have:

In order to improve the efficiency of cell lexicon addition, the format of the cell lexicon can also be processed as necessary. For example, the internal terms are sorted. Of course, these tasks can be completed in the generating module 3012, and then the data files sorted by the terms are provided to the user as a cell lexicon file for downloading.

The cell lexicon can also be encrypted for purposes such as copyright protection. Corresponding To decrypt the cell lexicon when it is installed. That is, preferably, the server side may further include an encryption module, and the input method client may further include a decryption module.

For ease of updating, the identification module 3013 also assigns a unique ID and a version number to each cell vocabulary.

The cell vocabulary in the embodiment shown in FIG. 3 can have various representations, for example, generally, multiple vocabulary information is directly stored in the cell vocabulary; or, the cell vocabulary can only store index information. The index information corresponds to other cell lexicons. The cell lexicon storing index information can be generally applied to: The server side stores a plurality of cell vocabularies obtained according to the received information, and then generates a new cell vocabulary according to a certain commonality of the cell lexicons (ie, indirect Using the received information), in order to achieve simplicity, the index information may be stored only in the new cell vocabulary. When the user needs the vocabulary, the server then merges the corresponding lexicons and transmits them.

Further, in order to satisfy the rapid update of the cell vocabulary, the cell vocabulary generating unit 301 of the vocabulary publishing system in this embodiment may further include: a modify updating module 3014, configured to modify the updated information of the updated cell vocabulary, and notify the The identification module 3013 generates new version information for the cell vocabulary. The modification may be done manually, or may be performed by adjusting the cell vocabulary according to a certain preset strategy, for example: other users add a new entry to a cell lexicon; or, according to a preset policy, The words in the two cell lexicons are combined into one cell lexicon; or, according to the Internet word frequency statistics result, the words in the cell vocabulary whose Internet word frequency does not meet the preset condition are deleted or sorted and adjusted.

First, download the cell vocabulary to the local, and then open the file by double-clicking to complete the data addition. A cell vocabulary is a file with a specific suffix name, such as a scd suffix. When the input method system is installed, the .scd suffix is associated with a specific application through the registry. When the user double-clicks the suffix to the .scd file, the operating system starts the corresponding application module according to the association rule (for example, the adding module in the embodiment shown in FIG. 1), and completes the addition of the cell lexicon data.

The second is to directly add the cell lexicon data online by clicking the link on the page. After the user clicks on the cell lexicon link on the page, there are two ways: save and execute. If the user saves the cell lexicon file, the same as the previous method. If the user chooses to execute, the system saves the cell thesaurus file In the system's temporary folder, then run it. The internal implementation mechanism is the same as the first one, except that the file is downloaded to the system temporary folder, so the user is not required to specify the download location. At the same time, the system will clean up the temporary directory when necessary, so although the cell vocabulary has been downloaded to the temporary directory, it is actually invisible to the user.

Preferably, the process of adding the downloaded cell vocabulary to the cell word set may also include a conversion step, such as sorting the original unordered terms in the lexicon to improve the efficiency of the addition. If there is a conversion step, the converted thesaurus file will be used; otherwise the original thesaurus file will be used directly. Of course, if the server has completed the conversion sorting process during the thesaurus generation process, the client does not need to repeat the work when the data is added.

During the data addition process, the input method system (ie, the input method client) needs to maintain a list of currently applied cell lexicons. The list of cell lexicons can take various feasible forms, for example, copy all active cell lexicons into a specified directory, or save a list of file names (this list can be placed in a local disk file) , can also be stored in the registry, or stored in a remote, such as on the network).

The process of adding the data of the cell lexicon to the cell word set can be operated immediately after the download is completed (for example, notifying the input method client to start the add operation); or waiting for the input method to actively discover the update (for example, the user starts the input next time) When the method is started, add the operation again: Scan the cell vocabulary list, read in and add each cell lexicon to the cell word set.

The following is an example of the expression of the cell word set for an independent large vocabulary. The specific addition process can be in two ways: incremental, batch.

The batch method combines the words in all cell lexicons into one large temporary vocabulary at once, and then adds the cell word set at once. This approach is simpler to implement, but the user must wait until all the lexicons have been merged before using the newly added cell lexicon. The incremental mode is: When several entries are read, they are added to the cell word set. If the merge time is long, the user can use it while merging, but this requires higher system design.

For the incremental merge mode, it can be used during the merge process, so there is no need to notify the input method system when the merge is completed. However, for the batch merge method, it is necessary to notify the input method system that the new thesaurus is ready for use after the merge is completed. An alternative approach is to directly access the storage space of the input method and update the data so that although the input method is not notified, the data has been updated, so the actual New data is ready to use.

Preferably, in the process of data addition, an optimization step may be further included for optimizing the repeated words in the thesaurus, for example, combining the repeated terms. Of course, in order to accurately record the word, information such as the identifier of the plurality of cell lexicons described therein may be recorded in its source attribute. Further, different priorities of the plurality of cell lexicons described in the word can also be recorded for different candidate environments to prioritize candidate cell rankings for different input environments.

In order to help the input method client to better complete the update task, the thesaurus publishing system in this embodiment can set the update identification work to be completed on the server side. Preferably, the vocabulary publishing system in this embodiment may further include: an identifying module 303, configured to compare the cell vocabulary list saved by the server and the cell vocabulary list sent by the client, and transmit according to the obtained comparison result. The required update data to the client. For example, the changed cell vocabulary formation list may be sent to the client, and the client may determine and initiate a download request; or, the server may directly push the changed cell vocabulary to the client to complete the update. The update data may be an entire cell vocabulary, for example, identifying that the cell lexicon needs to be updated, and transmitting all the vocabulary information of the cell lexicon; the updated data may also be in a cell vocabulary. Part of the term information, for example, identifying that the cell lexicon needs to be updated, further through the lexical comparison, only transmitting the changed term information.

In order to further improve the efficiency of the thesaurus publishing, the embodiment may further include: a merging module 304, configured to merge the plurality of cell lexicon terms information into one download vocabulary, and notify the communication unit 302 to transmit the downloaded vocabulary to the Client. The merging unit can be used for various possible scenarios, for example, combining a plurality of cell vocabularies selected by a user into one lexicon for transmission; or changing a plurality of cell lexicons that need to be updated. The term information is merged to obtain a new thesaurus, and then transmitted; or, the cell lexicon corresponding to the index information in the cell lexicon is combined to obtain a new vocabulary, and then transmitted.

Referring to FIG. 4, an embodiment of a method for updating a thesaurus is described. The required updated thesaurus relates to a set of cell words for recording extended words and related information in an input method system. At least one cell lexicon selected from a plurality of cell lexicons stored on the server side; the words in each cell lexicon have at least one common attribute;

The method embodiment may specifically include:

Step 401: Receive a trigger, compare an existing cell vocabulary list and a server-side cell vocabulary list, Obtaining a list of the thesaurus that is required to be updated; the triggering may be a manual triggering or an automatic triggering; Step 402: Download the updated cell lexicon entry information and add to the cell word set. Preferably, the method embodiment may further include the step 403: manually or automatically upgrading the cell vocabulary stored on the server side, and changing the corresponding version information. The upgrade may be done manually, or may be performed by adjusting the cell vocabulary according to a certain preset strategy, for example: other users add a new entry to a cell vocabulary; or, according to a preset policy, The words in the two cell lexicons are merged into one cell lexicon; or, according to the Internet word frequency statistical result, the words in the cell vocabulary whose Internet word frequency does not meet the preset condition are deleted or sorted and adjusted.

For easy updating, each cell lexicon can have a unique ID, which can be a naturally growing integer, or a network address or other information (as long as two different cell lexicons have different IDs) can). Each cell vocabulary can also have a version information, which can be a serial number or the time of the last modification. A change in the version information indicates that the thesaurus file needs to be updated. For example, if the last update time of the client is used as the version information, if there is a change in the file update time saved on the server, the thesaurus file needs to be updated.

For the implementation of the comparison process in step 401, various implementation manners may be used, for example: (1) The input method client sends the existing cell vocabulary list to the server, which may be sent through the TCP/IP protocol or sent through the HTTP protocol. ; Aligned by the server: Whether the cell lexicon corresponding to the ID in the list needs to be updated.

(2) The input method client initiates an update request, and the server sends back all the list information of the cell lexicon (or the changed cell vocabulary list), and the input method client determines which existing vocabulary needs to be updated.

(3) The input method client sends the existing cell vocabulary list to the server, and the server sends back the version information of the cell vocabulary corresponding to the ID in the list, and the input method client determines which existing vocabulary needs to be updated.

The above several methods have different network transmission bandwidths and device computing pressures, and those skilled in the art may select them according to actual needs.

For the case where the identification process is completed by the server, the server may send the changed cell vocabulary formation list to the client, and the client determines and initiates the download request (for example, selecting a part from the server) The lexicon is updated.) Alternatively, the cell vocabulary that has changed can be directly pushed to the client by the server to complete the update.

The data downloaded in step 402 may be the entire thesaurus or a part of the term information in a cell vocabulary, for example, changed term information.

For the data addition process in step 402, an incremental mode, a batch mode, or a combination of both may be employed. For example, the adding manner is: completing the update to download a cell vocabulary, adding the cell lexicon information to the cell word set; or, adding the method: after completing downloading of all the cell vocabularies to be updated , is added to the cell word set.

For incremental mode, you can install a thesaurus by updating a thesaurus. The advantage is that the downloaded thesaurus is not affected by the undownloaded thesaurus and can take effect immediately. However, when downloading a large number of thesaurus, it may lead to frequent addition of the thesaurus, which increases the burden on the system. Batch mode requires all the thesaurus to be downloaded locally before adding. Due to fewer add operations, the system load is lower. However, when the download process is long, especially if the download failure may occur in the middle, there will be a problem that the downloaded thesaurus cannot be used for a long time. In actual use, the two modes can be combined. For example, if a vocabulary is successfully downloaded, it is checked whether the last time the operation has been added has passed a predefined time interval; if it is exceeded, the vocabulary is added.

If the thesaurus addition process can be completed in a short period of time (for example, no more than 1 second), it can be inserted directly into the user's input process due to its small impact. However, if it cannot be completed in a short period of time and may affect the user's experience, the thesaurus addition process should be performed in a separate cached lexicon. In the process, the original vocabulary of the input method is not affected, and the user can use it normally. When the cached thesaurus is created, you can directly replace the original thesaurus of the input method. Since this replacement process can be fast, it can be avoided to interfere with the normal use of the user.

The above provides an input method system, a character input method, a word library update method and a thesaurus publishing system provided by the present invention, and a specific example is applied to the principle of the present invention. The embodiments have been described, and the description of the above embodiments is only for helping to understand the method of the present invention and its core ideas. Meanwhile, for those skilled in the art, according to the idea of the present invention, in the specific embodiments and application scope There are variations, and the description should not be construed as limiting the invention.

Claims

Rights request

An input method system, comprising an input interface unit, an information conversion unit, and a display output unit, further comprising:

System vocabulary for recording basic words and related information;

2. The input method system of claim 1, further comprising:

The automatic update module is configured to obtain the required update data from the server by comparing and judging according to the list of existing cell vocabularies.

The input method system according to claim 1, wherein the type of related information stored in the cell word set is less than or equal to a type of related information stored in the system vocabulary; and the server side stores At least one cell vocabulary in multiple cell lexicons is manually generated manually.

4. The input method system of claim 1 further comprising a user vocabulary.

5. The input method system of claim 1, further comprising:

And a module for adding cell lexicon information acquired from the server to the cell word set; the cell word set is an independent lexicon or a lexicon collection in which a plurality of lexicons are juxtaposed.

6. The input method system of claim 5, wherein

The adding manner is: after completing downloading of a cell vocabulary to be updated, adding the cell lexicon information to the cell word set;

Alternatively, the adding manner is: after all the cell vocabularies to be updated are downloaded, the cells are uniformly added to the cell word set.

7. The input method system of claim 6 wherein:

The addition process takes place in a separate cached lexicon.

8. The input method system of claim 1, further comprising:

Cell lexicon deactivation module, for receiving user instructions, removing from the cell word set belongs to the user selected The entry of the cell lexicon.

9. The input method system of claim 8 wherein:

The removing process is: receiving a user instruction, deleting the cell vocabulary selected by the user from the list, and re-adding the cell vocabulary in the list to obtain a new cell word set;

Or the removing process is: receiving a user instruction, deleting, from the cell word set, a term record belonging to a cell vocabulary selected by the user, where the cell word set records a cell vocabulary to which each term belongs; or The removing process is: receiving a user instruction, in the cell word set, adding a deletion mark to the entry record belonging to the cell vocabulary selected by the user, wherein the cell word set records the fine monthly package vocabulary to which each entry belongs.

10. A method of character input, comprising:

Loading a system vocabulary and a cell word set; the cell word set is for recording an extended word and related information; the cell word set is obtained by at least one cell vocabulary obtained from a plurality of cell lexicons stored on the server side Get; the words in each cell lexicon have at least one common attribute;

Receiving input information of the user;

11. The method of claim 10, wherein

The loading is: merging the cell word set and the system vocabulary into a vocabulary and placing it in a cache; or, the loading is: placing the cell word set and the system vocabulary as two or more independent lexicons In the cache, the lexicon priority is set according to a preset rule; the priority is used for display ordering of candidates.

12. The method according to claim 10, wherein the cell word set records a cell vocabulary to which each term belongs and a corresponding cell lexicon priority; the priority is used for display ordering of candidates .

13. The method of claim 12, further comprising:

During the loading process, the cell lexicon priority is dynamically adjusted according to the usage environment of the input method.

14. A method of updating a thesaurus, characterized in that the updated thesaurus relates to a set of cell words for recording extended words and related information, the plurality of cells stored by the server side. At least one cell lexicon selected in the lexicon is obtained; the words in each cell lexicon have at least one total Same attribute

The method includes:

The method according to claim 14, wherein before the triggering, the method further comprises: manually or automatically upgrading the cell vocabulary stored on the server side, and changing the corresponding version information.

16. The method of claim 14 wherein:

17. The method of claim 16 wherein said adding process is performed in a separate cache vocabulary.

18. A thesaurus publishing system, characterized in that:

The vocabulary distribution system according to claim 18, wherein the cell vocabulary generating unit further comprises:

The update module is configured to modify the updated cell lexicon stored information, and notify the identifier module to generate new version information for the cell vocabulary.

The vocabulary publishing system according to claim 18, further comprising: an identification module, configured to compare the cell vocabulary list of the server side and the cell vocabulary list of the client, and communicate according to the obtained comparison result The unit transmits the required update data to the client.

21. The thesaurus publishing system of claim 18, wherein:

A plurality of entry information is stored in the cell vocabulary obtained according to the received information;

Or the index information is stored in the cell vocabulary obtained according to the received information, and the index information The information corresponds to other cell lexicons.

The vocabulary publishing system according to claim 18, further comprising: a merging module, configured to merge the plurality of cell lexicon entry information into one download vocabulary, and notify the communication unit to download the word The library is transferred to the client.