WO2022180815A1

WO2022180815A1 - Information processing program, information processing method, and information processing device

Info

Publication number: WO2022180815A1
Application number: PCT/JP2021/007476
Authority: WO
Inventors: 壮吉藤田; 仙知大三島; チャーミンハンシラナウィーラ，ジャヤラタゲー; 源高
Original assignee: 富士通株式会社
Priority date: 2021-02-26
Filing date: 2021-02-26
Publication date: 2022-09-01
Also published as: JPWO2022180815A1

Abstract

An information processing program that makes a computer execute the following processing. Processing that receives a combination request for a first API and a second API. Processing that, on the basis of information for the first API that includes definition information, determines whether the first API has the completeness to make it possible for a request to the second API to be created from information included in a response from the first API. Processing that, when it has been determined that the first API does not have the completeness, searches for and acquires a linking API to generate information for generating the request to the second API on the basis of the information included in the response from the first API.

Description

Information processing program, information processing method, and information processing apparatus

The present invention relates to an information processing program, an information processing method, and an information processing apparatus.

In recent years, cloud services have been proposed that have functions that allow users to automate their work by combining multiple APIs (Application Programming Interfaces). This service targets not only IT (Information Technology) engineers but also business users.

When using such a cloud service, the correct combination of APIs will properly automate operations. In order to combine APIs correctly, it is necessary to understand API specifications and write code to connect APIs, but it is difficult for users who do not have a good understanding of technical details to do this. .

For example, when combining two APIs, generally after the response of the first API is properly processed, the processing result is used to create a request for the second API. This is the same between each API even when combining three or more APIs.

However, if the response of the first API is incomplete, it is difficult to combine the second API as it is. In such a case, another API that connects the first API and the second API is searched, and the processing result of the first API and the request for the second API are sent by the searched API. Work to describe the connecting process occurs. For example, APIs that handle image recognition and natural language often do not include execution results in responses. Such a response often indicates the storage location of the execution result. In such a case, in order to use the execution result in the second API, another API that acquires the execution result is called using the processing result included in the response of the first API, and the second API It is required to obtain information used in requests to

Conventionally, as a technology for connecting APIs, multiple APIs are collected, the correlation of APIs is determined from the information of each API, an API mashup database is created, and appropriate There is a technology that responds to various APIs. There is also a technique for efficiently executing a program in consideration of parallelism when performing API call processing.

JP 2020-126641 A JP 2019-179383 A

However, when business users who are not IT engineers perform the task of connecting the two APIs, there is concern that various mistakes may occur. For example, there is a risk of an error in the combination of APIs connecting two APIs to be used, an error in input information, and the like. When an abnormality occurs due to such a mistake, it is required to identify the cause of the abnormality and correct it. If there are multiple mistakes, these steps will be repeated many times. Therefore, it becomes difficult to efficiently combine a plurality of APIs. In addition, it takes a long time to normally operate the service to be realized by combining the required APIs, making it difficult to easily obtain the desired service.

In addition, in the technique of creating an API mashup database and proposing an appropriate API, it is possible to propose an appropriate combination of APIs, but it is difficult to automatically detect an API that connects two APIs and combine the APIs. It is difficult. Moreover, even with a technique for efficiently executing a program in consideration of parallelism when performing API call processing, it is difficult to automatically detect an API that connects two APIs and combine the APIs. Therefore, no matter which technology is used, it is difficult to efficiently combine multiple APIs.

The disclosed technology has been made in view of the above, and aims to provide an information processing program, an information processing method, and an information processing apparatus that efficiently combine a plurality of APIs.

The information processing program disclosed in this application causes a computer to execute the following processes. A process of accepting a combination request for a first API and a second API. The first API has integrity that enables the creation of a request to the second API based on the information of the first API including definition information and the information contained in the response of the first API. Processing to determine whether or not. generating information for generating the request to the second API based on information included in the response of the first API if it is determined that the first API does not have the integrity; A process of searching for and acquiring a connection API.

In one aspect, the present invention can efficiently combine multiple APIs.

FIG. 1 is a system configuration diagram of an API combination support system. FIG. 2 is a block diagram of an API combination support device. FIG. 3 is a diagram showing an example of connecting two APIs. FIG. 4 is a diagram showing an example of connecting two APIs by a plurality of APIs. FIG. 5 is a flowchart of API combination processing by the API combination support device according to the embodiment. FIG. 6 is a flow chart of search processing for the link API. FIG. 7 is a hardware configuration diagram of an API combination support device.

Hereinafter, embodiments of the information processing program, the information processing method, and the information processing apparatus disclosed in the present application will be described in detail based on the drawings. The information processing program, the information processing method, and the information processing apparatus disclosed in the present application are not limited to the following embodiments.

Fig. 1 is a system configuration diagram of the API combination support system. The API combination support system 1 has an API combination support device 10 , a repository 20 and a user terminal 30 .

The API combination support device 10 is an information processing device such as a server, and performs combination processing for combining designated APIs. API combination support device 10 is connected to repository 20 and user terminal 30 .

The repository 20 includes a database of definition information for multiple APIs. The API definition information includes schema definition information indicating data information used in a request for each API and data information returned in a response. In addition, the definition information includes various kinds of information that serve as criteria for judging the completeness of the API. An incomplete API is an API that cannot pass processing to the next API without the help of another API. That is, an API having integrity with respect to the second API can be said to be an API that has the property of being able to create a request for the second API based on the information included in the response.

The definition information also includes, for example, the presence or absence of external references in each API, the size of the request value, the size of the response value, the number of request keys, the number of response keys, and the like. The definition information also includes the name of the key of the response, the type and content of the value, and the outline and details of each API. Furthermore, in the repository 20, connectable APIs are arranged so that they have close addresses.

Also, the repository 20 may include integrity information indicating whether each API is complete or incomplete. For example, the completeness information may be registered by the administrator of the API combination support system 1 based on past combination results, or information defined by the administrator may be registered. Also, the repository 20 may store past combination information. Here, a collection of definition information and completeness information for each API included in the repository 20 is simply referred to as "API information".

The user terminal 30 is a terminal that requests the API combination support device 10 to combine APIs. The user terminal 30 receives an input from the user P specifying APIs to be combined. Then, the user terminal 30 transmits identification information of the specified API to the API combination support device 10 and requests execution of API combination.

FIG. 2 is a block diagram of the API combination support device. The API combination support device 10 has an API acquisition unit 101, an API information acquisition unit 102, a linked API search unit 103, an integrity determination unit 104, and a combination generation unit 105, as shown in FIG. Here, the number of APIs to be combined is not limited, but the procedure for combining individual APIs is the same between any APIs. That is, when combining N APIs in order, the API combination support device 10 can: Combinations can be performed in a similar manner. Therefore, the case where two APIs are combined will be described below. Here, of the two APIs to be combined, the API that is read and executed earlier is called the "first API", and is called after the first API, and based on the execution result of the first API The API that is executed is called the "second API". Also, an API for connecting the first API and the second API is called a "connection API". The first AIP corresponds to an example of the first API, and the second API corresponds to an example of the second API.

The API acquisition unit 101 receives the identification information of the first API and the second API from the user terminal 30, and receives a request to execute the combination of the first API and the second API. Then, API acquisition section 101 outputs the identification information of the first API to API information acquisition section 102 . API acquisition section 101 also outputs the identification information of the first API and the second API to linked API search section 103 and completeness determination section 104 .

The API information acquisition unit 102 receives the input of the identification information of the first API from the API acquisition unit 101 . Then, the API information acquisition unit 102 determines whether or not the integrity information for the first API exists in the repository 20 . When the integrity information exists, the API information acquisition unit 102 acquires the integrity information of the first API from the repository 20 and outputs it to the integrity determination unit 104 .

On the other hand, if the integrity information of the first API does not exist, the API information acquisition unit 102 acquires the definition information of the first API from the repository 20. Then, API information acquisition section 102 outputs the acquired definition information of the first API to integrity determination section 104 .

Also, the API information acquisition unit 102 receives from the link API search unit 103 a request to acquire past combination information of the first API or the link API and the second API. When the specified past combination information exists in the repository 20 , the API information acquisition unit 102 acquires the specified past combination information from the repository 20 and outputs the acquired past combination information to the linked API search unit 103 . When there is no designated past combination information, API information acquisition section 102 notifies link API search section 103 of the absence of past combination information sound.

The API information acquisition unit 102 also receives an API information acquisition request from the linked API search unit 103 . Then, the API information acquisition unit 102 determines whether or not the integrity information for the specified API exists in the repository 20 . When the integrity information exists, the API information acquisition unit 102 acquires the integrity information of the designated API from the repository 20 and outputs it to the integrity determination unit 104 .

On the other hand, if there is no completeness information, the API information acquisition unit 102 acquires the definition information of the specified API from the repository 20. Then, API information acquisition section 102 outputs the definition information of the designated API to integrity determination section 104 .

The integrity determination unit 104 receives input of the first API information and the second API identification information from the API acquisition unit 101 . Next, when the integrity information of the first API exists in the repository 20 , the integrity determination unit 104 acquires the integrity information of the first API from the API information acquisition unit 102 . Then, the integrity determination unit 104 determines whether or not the first API has integrity based on the integrity information.

On the other hand, if the integrity information of the first API does not exist in the repository 20, the integrity determination unit 104 receives the input of the definition information of the first API from the API information acquisition unit 102. Then, the integrity determination unit 104 determines the integrity of the first API based on the scoring and classification results of a predetermined rule-based algorithm using the information included in the acquired definition information. An example of scoring and classification results by a rule-based algorithm will be described below.

For example, the integrity determination unit 104 acquires API schema definition information from the definition information and checks the presence or absence of external references based on whether the URL (Uniform Resource Locator) or ID (Identifier) of the external resource exists. do. Then, the completeness determination unit 104 increases the incompleteness score when there is an external reference. Further, the completeness determination unit 104 acquires the request value and the response value from the definition information, and calculates the difference between the magnitude of the request value and the magnitude of the response value. Then, the completeness determination unit 104 increases the score of incompleteness as the difference between the magnitude of the request value and the magnitude of the response value increases. Further, the integrity determination unit 104 acquires the number of request keys and the number of response keys from the definition information, and calculates the difference between the number of request keys and the number of response keys. Then, the completeness determination unit 104 increases the incompleteness score as the difference between the number of request keys and the number of response keys increases. Further, the completeness determination unit 104 acquires one or a combination of the name of the response key, the type of the value, and the contents of the value from the definition information. Then, the integrity determination unit 104 determines whether or not the data acquired from the response can be processed by the second API from the acquired information. Then, the completeness determination unit 104 increases the incompleteness score as the probability of the data that can be processed by the second API is low. Further, the integrity determination unit 104 acquires the outline and details of the first API from the definition information, checks the content of the first API, and determines the possibility of having integrity. Then, the completeness determination unit 104 increases the incompleteness score as the possibility of completeness is low. After that, the completeness determination unit 104 calculates the sum of the incompleteness scores, and determines that the first API is incomplete if the calculated result is higher than a predetermined threshold. Conversely, if the calculation result is equal to or less than the predetermined threshold, the integrity determination unit 104 determines that the first API has integrity.

In addition, the completeness judging unit 104 performs machine learning from learning data using each item obtained from the definition information described above as a feature amount, and classifies whether or not there is completeness using the generated learning model. may be performed to determine whether or not the first API is complete.

When the first API has integrity, the integrity determination unit 104 outputs information on the first API and the second API to the combination generation unit 105 along with a notification that the first API has integrity. do. At this time, the completeness determination unit 104 also outputs the definition information of the first API and the second API to the combination generation unit 105 . On the other hand, if the first API is incomplete, the completeness determination unit 104 notifies the link API search unit 103 of an instruction to search for a link API for linking the first API and the second API. do.

Also, the integrity determination unit 104 receives a request to confirm the integrity of the linked API from the linked API search unit 103 when the first API is incomplete. If the integrity information exists in the repository 20, the integrity determination unit 104 acquires the integrity information and determines the integrity of the designated API. Further, when the integrity information of the specified linked API does not exist in the repository 20 or is incomplete due to the integrity information, the integrity determination unit 104 obtains the definition information of the specified linked API from the API information acquisition unit. 102. The completeness determination unit 104 then performs scoring and classification using a predetermined rule-based algorithm, similar to the first API.

Then, the completeness determination unit 104 determines whether or not the specified API has a higher completeness index value obtained by scoring or classification than the preceding linked API. If the completeness index value of the designated API is less than or equal to the previous linked API, the integrity determination unit 104 instructs to re-search the linked API to connect the previous linked API and the second API. is notified to the link API search unit 103 .

On the other hand, if the index value of the specified API is higher than the preceding linked API, the completeness determination unit 104 determines the completeness from the results of scoring and classification of the specified API.

When the specified concatenated API has integrity, the integrity determination unit 104 notifies the combination generation unit 105 that the specified concatenated API has integrity. On the other hand, if the specified linking API is incomplete, the completeness determination unit 104 determines whether the specified API is selected together with the following command for linking the specified linking API and the second API. It notifies the linked API search unit 103 of a link API re-search instruction.

The linked API search unit 103 receives input of the identification information of the first API and the second API from the API acquisition unit 101 . After that, when the first API is incomplete, the link API search unit 103 receives a link API search instruction from the completeness determination unit 104 . Then, the linked API search unit 103 searches for linked APIs connected to the first API. Further, when there is a selected linked API, the linked API search unit 103 searches for the next linked API if the previous linked API does not have completeness. Finding the Concatenation API is described in detail below.

The linked API search unit 103 requests the API information acquisition unit 102 to acquire the past combination information of the previous API and the second API, and checks the presence or absence of the past combination information of the previous linked API. judge. When past combination information of the linked API immediately before and the API in the second place exists in the repository 20, the completeness determination unit 104 acquires the past combination information of the linked API immediately before the API information acquisition unit 102. Get from Then, the link API search unit 103 requests the API information acquisition unit 102 to acquire the definition information of the link API specified by the acquired past combination information. Get information. After that, the linked API searching unit 103 outputs the information of the first API, the second API, and all the APIs arranged between the first API and the second API to the combination generating unit 105 . At this time, the linked API searching unit 103 also outputs definition information of all APIs arranged between the first API and the second API to the combination generating unit 105 .

On the other hand, when the past combination information of the immediately previous linked API does not exist in the repository 20, the linked API search unit 103 requests the API information acquisition unit 102 to acquire the definition information of the immediately previous API. Get the definition information of the previous API. Then, the linked API search unit 103 acquires the key or value of the response of the previous API from the definition information. Here, the linked API search unit 103 stores a predetermined priority for each key and value of the response. For example, the priority is determined in the order in which the following APIs can obtain a response with as much completeness as possible. Therefore, the linked API search unit 103 sorts the keys or values of the previous API response in order of priority.

Next, the linked API search unit 103 selects the key or value with the highest priority from among the unselected keys and values arranged in order of priority. Then, the linked API search unit 103 determines whether or not there is an API having the selected key or value in the request item among the APIs having addresses close to the previous linked API in the repository 20 . If there is an API having the selected key or value in the request item among APIs with close addresses, the linked API search unit 103 determines that API as a candidate for the next linked API.

For example, when information about a data storage location is returned as a response to the first API, the link API search unit 103 determines an API that acquires data from that storage location as a candidate for the next link API. In addition, when extracting a mail address from text information returned as a response from the first API, the linking API search unit 103 determines an API that performs a process of extracting the mail address as a candidate for the next linking API.

Next, the linked API search unit 103 requests the integrity determination unit 104 to determine the integrity of the APIs determined as candidates for the linked API. After that, the link API search unit 103 receives a response to the integrity determination request from the integrity determination unit 104 .

When receiving from the completeness determination unit 104 an instruction to re-search the concatenated API for connecting the immediately preceding concatenated API and the second API, the concatenated API search unit 103 selects a key or value according to the priority. Do it again. Then, using the selected key or value, the linked API search unit 103 searches again for an API that is a candidate for the next linked API, and requests the completeness determination unit 104 to perform completeness determination.

Further, when a response indicating that the specified API is incomplete is received from the completeness determination unit 104 together with the selection instruction of the selected API, the link API search unit 103 selects the determined API as the next link API. do. Then, the linked API searching unit 103 similarly searches for the linked API next to the linked API selected at this time. The linked API search unit 103 repeats determination of the next linked API until the linked API has completeness.

When the completeness determination unit 104 determines that the designated link API has integrity, the link API search unit 103 terminates the search for the link API. Linked API searching section 103 then outputs to combination generating section 105 the information of the first API, the second API, and all the APIs arranged between the first API and the second API. At this time, the linked API searching unit 103 also outputs definition information of all APIs arranged between the first API and the second API to the combination generating unit 105 .

Here, in this embodiment, the key and value of the response are used to search for the next linked API. good. For example, among the values included in the definition information transfer, other information may be used as long as it is information that can be linked to the second API. Extract to Then, the linked API search unit 103 may determine the most effective API among the APIs corresponding to the extracted information as a candidate for the next linked API.

If the first API has integrity, the combination generation unit 105 notifies the integrity determination unit 105 that the first API has integrity and the information and definition information of the first and second APIs. Get from After that, the combination generation unit 105 combines the code for calling the first API and the code for calling the second API when the response of the first API is returned. Accordingly, the combination generator 105 combines the first API and the second API. After that, the combination generator 105 transmits the result of direct combination of the first API and the second API to the user terminal 30 for display, and also transmits the generated code to the user terminal 30 .

On the other hand, if the first API is incomplete, the combination generation unit 105 generates the first API, the second API, and all the APIs lined up between the first API and the second API. and definition information from the combination generation unit 105 . Then, the combination generating unit 105 generates a code for sequentially calling all linked APIs arranged between the first API and the second API, and generates a code for calling the first API and a code for calling the second API. Insert between Accordingly, the combination generator 105 combines the first API and the second API. After that, the combination generation unit 105 transmits the combination result of arranging the linked APIs between the first API and the second API to the user terminal 30 for display, and transmits the generated code to the user terminal 30. Send.

FIG. 3 is a diagram showing a case where one linked API is inserted between the first API and the second API. Next, referring to FIG. 3, a procedure for inserting one linking API between the first API and the second API to combine the first API and the second API will be described. For example, the first API 201 is a natural language processing API, and the second API 202 is a chat API. The first API is an API that executes POST for an image under domainA. The second API is an API for executing POST for photos under domainB.

At this time, the integrity determination unit 104 acquires the value of the response 211 specified from the definition information of the first API 201 . In this case, since the response 211 includes status information and id indicating natural language identification information, the completeness determination unit 104 determines that the first API 201 is incomplete. Then, the linking API search unit 103 determines, as the linking API 203 , an API for acquiring the natural language corresponding to the id indicating the identification information as the response 212 specified from the definition information of the first API 201 . The linking API 203 is an API that executes GET on the following images for domainA.

The integrity determination unit 104 determines that the linkage API 203 has integrity because the response of the linkage API 203 includes text information representing a natural language. Linked API searching section 103 then outputs information on first API 201 , linked API 203 and second API 202 to combination generating section 105 . The combination generation unit 105 arranges codes for reading the first API 201, the concatenated API 203, and the second API 202 using the respective results, and combines the first API 201 and the second API 202. FIG. This completes a series of processes of acquiring the natural language id as a response to the first API 201, which is a natural language API, acquiring the natural language corresponding to the acquired id, and writing the acquired natural language to the chat.

Also, FIG. 4 is a diagram showing a case where a plurality of linked APIs are inserted between the first API and the second API. Next, referring to FIG. 4, a procedure for inserting a plurality of linking APIs between the first API and the second API to combine the first API and the second API will be described. Here, the case where the first API is API#A, the second API is API#B, and the repository 20 contains information on API#C to #E will be described.

The completeness determination unit 104 determines that API#A is incomplete, as in state 301, based on the value of the API#A response.

Therefore, the link API search unit 103 requests the API information acquisition unit 102 to acquire an API that is a candidate for the link API from the key or value of the response of API #A. Then, the link API search unit 103 acquires the information of API #C as an API that is a candidate for the link API. After that, the completeness determination unit 104 determines that the API#C is incomplete as shown in the state 302 from the value of the API#C response.

Therefore, the link API search unit 103 requests the API information acquisition unit 102 to acquire the next link API candidate from the key or value of the API #C response. Then, the link API search unit 103 acquires the information of the API #D as a link API candidate. After that, the completeness determination unit 104 determines that API#D is incomplete, as in state 303, based on the value of the response to API#D. As a result, by arranging API#A, API#C, API#D, and API#B, API#A and API#B are combined.

FIG. 5 is a flowchart of API combination processing by the API combination support device according to the embodiment. Next, the flow of API combination processing by the API combination support device 10 according to the embodiment will be described with reference to FIG.

User P uses the user terminal 30 to specify APIs to be combined. The API acquisition unit 101 of the API combination support device 10 receives the identification information of each API to be combined from the user terminal 30 (step S1).

The API information acquisition unit 102 acquires the identification information of the first API from the API acquisition unit 101 . Then, the API information acquisition unit 102 acquires the information of the first API from the repository 20 (step S2).

The integrity determination unit 104 acquires information on the first API from the API information acquisition unit 102 . Then, the completeness determination unit 104 uses the completeness information or the definition information to determine whether or not the first API is complete (step S3). If the first API has integrity (step S3: affirmative), the integrity determination unit 104 obtains the information indicating that the first API has integrity and the definition information of the first API and the second API. to the combination generation unit 105 . Thereafter, the API combination processing proceeds to step S6. On the other hand, if the first API is incomplete (step S3: No), the completeness determination unit 104 sends the definition information of the first API to the linked API search unit 103 and searches for the linked API. to the link API search unit 103 .

If the first API or the previous linked API is incomplete, the linked API search unit 103 searches for the next linked API (step S4).

After that, the linked API search unit 103 requests the integrity determination unit 104 to determine whether or not the next linked API is complete. The integrity determination unit 104 determines whether or not the next link API has integrity (step S5). When the next linked API is incomplete (step S5: No), the completeness determination unit 104 notifies the linked API search unit 103 that the next linked API is incomplete. Thereafter, the API combination processing returns to step S4.

If the linked API specified by the linked API search unit 103 has integrity (step S5: Yes), the integrity determination unit 104 notifies the linked API search unit 103 that the next linked API has integrity. Linked API search unit 103 arranges extracted APIs for combining the first API and second API between the first API and second API, and notifies combination generation unit 105 of the respective definition information. . The combination generation unit 105 generates codes for calling each API used to combine the first API and the second API, and arranges them in order (step S6). Here, when the concatenated API is used, the combination generation unit 105 inserts a code for sequentially reading each concatenated API between the code for calling the first API and the code for calling the second API.

After that, the combination generation unit 105 transmits the combination result of combining the first API and the second API to the user terminal 30 for display, and also transmits the generated code to the user terminal 30 to perform the combination processing. The result is notified to the user P (step S7).

FIG. 6 is a flow chart of the search processing for the link API. Each process shown in the flowchart of FIG. 6 corresponds to an example of the process executed in step S4 in FIG.

The linked API search unit 103 requests the API information acquisition unit 102 to acquire past combination information of the previous API and the second API, and determines whether the past combination information exists in the repository 20. is determined (step S11).

If the past combination information of the previous linked API and the second API exists in the repository 20 (step S11: Yes), the linked API searching unit 103 proceeds to step S16.

On the other hand, if the past combination information of the previous API and the second API does not exist in the repository 20 (step S11: No), the linked API searching unit 103 retrieves the definition information of the previous API. to the API information acquisition unit 102 . Then, the linked API search unit 103 acquires the definition information of the previous API stored in the repository 20 from the API information acquisition unit 102 . Next, the link API search unit 103 extracts the key and value of the response from the acquired definition information. Then, the link API searching unit 103 sorts the keys and values of the extracted response in order of priority (step S12).

Next, the linked API search unit 103 selects one unselected item in order of priority from among the keys and values arranged in order of priority (step S13).

Next, the linked API search unit 103 determines whether or not there is an API having the selected key or value in the request item among the APIs with addresses close to the previous linked API in the repository 20. (Step S14).

If there is no API having the selected key or value in the request item among APIs with close addresses (step S14: No), the linked API search unit 103 returns to step S13.

On the other hand, if there is an API having the selected key or value in the request item among the APIs with close addresses (step S14: Yes), the link API search unit 103 searches the API for the next link API. Decide as a candidate. Next, the linked API search unit 103 requests the completeness determination unit 104 to determine the integrity of the APIs determined as candidates for the linked API. The integrity determination unit 104 determines whether or not the index value of the integrity of the designated API is higher than that of the previous API (step S15). If the completeness index value does not increase (step S15: No), the link API search process returns to step S13.

On the other hand, if the index value of completeness increases (step S15: affirmative), the completeness determination unit 104 instructs the linked API search unit 103 to select the designated API.

The linked API search unit 103 receives the instruction from the completeness determination unit 104 and selects the corresponding API as the next linked API (step S16).

As described above, the API combination support device according to the present embodiment determines the completeness of the first API when combining APIs, and selects a linked API based on the value of the response if the first API is incomplete. and list until you have a complete concatenated API. As a result, the API combination support device can easily detect the linking APIs used to combine the specified APIs, and provide information on what kind of APIs are interposed to combine the specified APIs. becomes possible. Therefore, it is possible to efficiently combine a plurality of APIs. In addition, it can contribute to shortening the time required for combination and reducing costs.

(Hardware configuration)
FIG. 6 is a hardware configuration diagram of the server. The API combination support device 10 according to this embodiment has, for example, a hardware configuration as shown in FIG. The API combination support device 10 has a CPU 91 , a memory 92 , a storage device 93 and a communication module 94 . The CPU 91, memory 92, storage device 93 and communication module 94 are connected by buses.

The storage device 93 is an auxiliary storage device such as an SSD (Solid State Drive) or hard disk. The storage device 93 stores various programs including programs for operating the API acquisition unit 101, the API information acquisition unit 102, the linked API search unit 103, the completeness determination unit 104, and the combination generation unit 105 illustrated in FIG. be. The storage device 50 may also store the repository 20 .

The communication module 94 is a network interface for the API combination support device 10 to communicate with external devices. For example, the CPU 91 communicates with the repository 20 and the user terminal 30 via the communication module 94 .

The memory 92 is a main storage device such as SDRAM (Synchronous Dynamic Random Access Memory).

The CPU 91 reads out various programs stored in the storage device 93, develops them on the memory 92, and executes them to obtain the API acquisition unit 101, the API information acquisition unit 102, and the linked API shown in FIG. The functions of the search unit 103, the completeness determination unit 104, and the combination generation unit 105 are realized.

1 API combination support system 10 API combination support device 20 repository 30 user terminal 101 API acquisition unit 102 API information acquisition unit 103 linked API search unit 104 completeness determination unit 105 combination generation unit

Claims

Receiving a combination request of a first API (Application Programming Interface) and a second API,
The first API has integrity that enables the creation of a request to the second API based on the information of the first API including definition information and the information contained in the response of the first API. determine whether or not
generating information for generating the request to the second API based on information included in the response of the first API if it is determined that the first API does not have the integrity; An information processing program that causes a computer to execute a process of searching for and acquiring a linking API.
determining the presence or absence of the completeness of the linked API obtained by the search;
If it is determined that the linked API does not have the integrity, the process of searching for and acquiring the next linked API based on the information of the linked API is repeated until the linked API having integrity is acquired. The information processing program according to claim 1, characterized by:
if the information of the first API includes integrity information indicating the presence or absence of integrity, using the integrity information to determine the presence or absence of integrity of the first API;
If the integrity information is not included in the information of the first API, the computer executes a process of determining presence or absence of integrity of the first API based on the definition information of the first API. 2. The information processing program according to claim 1, characterized in that it causes
A computer is caused to execute a process of determining whether or not the first API is complete according to a predetermined algorithm based on response information from the first API included in the definition information. Item 1. The information processing program according to item 1.
performing learning based on the definition information of each of the first APIs to generate a learning model, and causing a computer to execute a process of determining whether or not the first API is complete using the generated learning model; The information processing program according to claim 1, characterized by:
The information processing program according to claim 1, further causing a computer to execute a process of notifying a result of combining the first API, the linking API, and the second API.
receiving a combination request of the first API and the second API;
The first API has integrity that enables the creation of a request to the second API based on the information of the first API including definition information and the information contained in the response of the first API. determine whether or not
generating information for generating the request to the second API based on information included in the response of the first API if it is determined that the first API does not have the integrity; An information processing method characterized by searching for and acquiring a linking API.
an API acquisition unit that receives a combination request for the first API and the second API;
an integrity determination unit that determines whether or not each API has integrity that enables creation of a request to the second API based on information included in the response of each of the APIs;
generating the request for the second API based on information included in the response of the first API when the integrity determination unit determines that the first API does not have integrity; an information processing apparatus comprising: a link API search unit that searches for and acquires a link API that generates information for.