CN114024872A

CN114024872A - Voice semantic platform abnormity alarm method, equipment, storage medium and device

Info

Publication number: CN114024872A
Application number: CN202010695882.3A
Authority: CN
Inventors: 吴遐; 张晓芬
Original assignee: Beijing Qihoo Technology Co Ltd
Current assignee: Beijing Qihoo Technology Co Ltd
Priority date: 2020-07-17
Filing date: 2020-07-17
Publication date: 2022-02-08

Abstract

The invention discloses a method, a device, a storage medium and a device for alarming abnormality of a voice semantic platform, wherein the method comprises the following steps: sending an abnormality detection instruction to a voice semantic platform to be detected, so that the voice semantic platform to be detected feeds back target field information and current operation scene information according to a platform debugging log when receiving the abnormality detection instruction; performing anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result; performing abnormity alarm based on the current operation scene information and the abnormity detection result; according to the method and the device, the abnormity detection is carried out on the voice semantic platform to be detected according to the target field information, the abnormity detection result is obtained, and abnormity alarm is carried out according to the current operation scene information and the abnormity detection result, so that various abnormity detections can be carried out on the voice semantic platform, and abnormity alarm is carried out according to the operation scene.

Description

Voice semantic platform abnormity alarm method, equipment, storage medium and device

Technical Field

The invention relates to the technical field of abnormity alarm, in particular to a method, equipment, a storage medium and a device for alarming abnormity of a voice semantic platform.

Background

The voice semantic platform is a multi-scene and multi-functional platform which supports user request input and makes corresponding request reply. At present, the anomaly detection of a voice semantic platform is very simple, and cannot cover multi-scene, multi-function and multi-error types.

The above is only for the purpose of assisting understanding of the technical aspects of the present invention, and does not represent an admission that the above is prior art.

Disclosure of Invention

The invention mainly aims to provide a method, equipment, a storage medium and a device for alarming abnormality of a voice semantic platform, and aims to solve the technical problem of how to detect various abnormalities of the voice semantic platform and alarm the abnormalities in the prior art.

In order to achieve the above object, the present invention provides an abnormality alarm method for a voice semantic platform, which comprises the following steps:

sending an abnormality detection instruction to a voice semantic platform to be detected, so that the voice semantic platform to be detected feeds back target field information and current operation scene information according to a platform debugging log when receiving the abnormality detection instruction;

performing anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result;

and performing abnormity alarm based on the current operation scene information and the abnormity detection result.

Preferably, the step of performing anomaly detection on the to-be-detected speech semantic platform according to the target field information to obtain an anomaly detection result specifically includes:

performing text reply detection on the to-be-detected voice semantic platform according to the target field information to obtain a text reply detection result;

acquiring current slot position information, and performing full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result;

and generating an abnormal detection result according to the text reply detection result and the full-value matching result.

Preferably, before the step of generating an abnormal detection result according to the text reply detection result and the full-value matching result, the method for alarming an abnormality of a speech semantic platform further includes:

acquiring an error log of the voice semantic platform to be detected;

performing return value matching detection on the error log to obtain a return value matching result;

correspondingly, the step of generating an abnormal detection result according to the text reply detection result and the full-value matching result specifically includes:

and generating an abnormal detection result according to the text reply detection result, the full value matching result and the return value matching result.

Preferably, the step of performing return value matching detection on the error log to obtain a return value matching result specifically includes:

generating a list type result according to the error log, and determining a return value according to the list type result;

and matching the return value with a preset expected value to obtain a return value matching result.

Preferably, the step of performing text reply detection on the to-be-detected speech semantic platform according to the target field information to obtain a text reply detection result specifically includes:

determining a text marking value according to the target field information;

and detecting the target field information according to the text label value to obtain a text reply detection result.

Preferably, the step of detecting the target field information according to the text label value to obtain a text reply detection result specifically includes:

determining a text reply detection type according to the text label value, and searching a text detection strategy corresponding to the text reply detection type;

and detecting the target field information through the text detection strategy to obtain a text reply detection result.

Preferably, the step of obtaining the current slot information and performing full-value matching detection on the target field information according to the current slot information to obtain a full-value matching result specifically includes:

acquiring current slot position information, and determining a current slot position identifier according to the current slot position information;

and determining a target verification strategy according to the current slot position identification, and carrying out full-value matching detection on the target field information through the target verification strategy to obtain a full-value matching result.

Preferably, the step of determining a target verification policy according to the current slot identifier, and performing full-value matching detection on the target field information through the target verification policy to obtain a full-value matching result specifically includes:

judging whether the current slot position mark is a preset mark or not;

when the current slot position identification is the preset identification, determining a target verification relation according to the current slot position identification;

and determining a target checking strategy according to the target checking relation, and carrying out full-value matching detection on the target field information through the target checking strategy to obtain a full-value matching result.

Preferably, after the step of determining whether the current slot identifier is a preset identifier, the method for alarming abnormality of a voice semantic platform further includes:

when the current slot position identification is not the preset identification, taking a preset inspection strategy as a target detection strategy;

and carrying out full-value matching detection on the target field information through the target verification strategy to obtain a full-value matching result.

Preferably, the step of performing an anomaly alarm based on the current operating scenario information and the anomaly detection result specifically includes:

determining the current failure rate, the accumulated failure times and the content of an error log according to the abnormal detection result;

and performing abnormity alarm according to the current operation scene information, the current failure rate, the accumulated failure times and the current error log content.

Preferably, the step of performing an exception alarm according to the current operation scenario information, the current failure rate, the accumulated failure times, and the current error log content specifically includes:

respectively searching a failure rate threshold value, an accumulated failure frequency threshold value and important error log contents corresponding to the current operation scene;

judging whether the current failure rate is greater than the failure rate threshold value or not, and obtaining a first judgment result;

judging whether the accumulated failure times is greater than the accumulated failure time threshold value or not, and obtaining a second judgment result;

judging whether the current error log content is the important error log content or not, and obtaining a third judgment result;

and performing abnormity alarm according to the current operation scene, the first judgment result, the second judgment result and the third judgment result.

Preferably, the step of performing an abnormality alarm according to the current operating scene, the first determination result, the second determination result, and the third determination result specifically includes:

searching for a target early warning strategy corresponding to the current operation scene, the first judgment result, the second judgment result and the third judgment result;

and carrying out abnormity early warning according to the target early warning strategy.

In addition, in order to achieve the above object, the present invention further provides a voice semantic platform exception alarm device, where the voice semantic platform exception alarm device includes a memory, a processor, and a voice semantic platform exception alarm program stored in the memory and operable on the processor, and the voice semantic platform exception alarm program is configured to implement the steps of the above voice semantic platform exception alarm method.

In addition, in order to achieve the above object, the present invention further provides a storage medium, where a voice semantic platform exception alarm program is stored on the storage medium, and when executed by a processor, the voice semantic platform exception alarm program implements the steps of the voice semantic platform exception alarm method described above.

In addition, in order to achieve the above object, the present invention further provides an abnormality alarm device for a speech semantic platform, where the abnormality alarm device for a speech semantic platform includes: the device comprises a sending module, a detection module and an alarm module;

the sending module is used for sending an abnormality detection instruction to the voice semantic platform to be detected, so that the voice semantic platform to be detected feeds back target field information and current operation scene information according to a platform debugging log when receiving the abnormality detection instruction;

the detection module is used for carrying out abnormity detection on the to-be-detected voice semantic platform according to the target field information to obtain an abnormity detection result;

and the alarm module is used for carrying out abnormity alarm based on the current operation scene information and the abnormity detection result.

Preferably, the detection module is further configured to perform text reply detection on the to-be-detected speech semantic platform according to the target field information, so as to obtain a text reply detection result;

the detection module is further used for obtaining current slot position information and carrying out full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result;

the detection module is further used for generating an abnormal detection result according to the text reply detection result and the full-value matching result.

Preferably, the voice semantic platform anomaly alarm device further includes: a matching module;

the matching module is used for acquiring an error log of the to-be-detected voice semantic platform;

the matching module is also used for carrying out return value matching detection on the error log to obtain a return value matching result;

correspondingly, the detection module is further configured to generate an abnormal detection result according to the text reply detection result, the full value matching result, and the return value matching result.

Preferably, the matching module is further configured to generate a list type result according to the error log, and determine a return value according to the list type result;

and the matching module is also used for matching the return value with a preset expected value to obtain a return value matching result.

Preferably, the detection module is further configured to determine a text label value according to the target field information;

the detection module is further used for detecting the target field information according to the text marking value to obtain a text reply detection result.

Preferably, the detection module is further configured to determine a text reply detection type according to the text label value, and search for a text detection policy corresponding to the text reply detection type;

the detection module is further configured to detect the target field information through the text detection policy to obtain a text reply detection result.

In the invention, an anomaly detection instruction is sent to a voice semantic platform to be detected, so that when the voice semantic platform to be detected receives the anomaly detection instruction, target field information and current operation scene information are fed back according to a platform debugging log; performing anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result; performing abnormity alarm based on the current operation scene information and the abnormity detection result; according to the method and the device, the abnormity detection is carried out on the voice semantic platform to be detected according to the target field information, the abnormity detection result is obtained, and abnormity alarm is carried out according to the current operation scene information and the abnormity detection result, so that various abnormity detections can be carried out on the voice semantic platform, and abnormity alarm is carried out according to the operation scene.

Drawings

FIG. 1 is a schematic structural diagram of a voice semantic platform exception warning device of a hardware operating environment according to an embodiment of the present invention;

FIG. 2 is a flowchart illustrating a first embodiment of an abnormal alarm method for a speech semantic platform according to the present invention;

FIG. 3 is a flowchart illustrating a second embodiment of the abnormal alarm method for a speech semantic platform according to the present invention;

FIG. 4 is a flowchart illustrating a third embodiment of an abnormal alarm method for a speech semantic platform according to the present invention;

FIG. 5 is a flowchart illustrating a fourth embodiment of an abnormal alarm method for a speech semantic platform according to the present invention;

fig. 6 is a block diagram of a first embodiment of an abnormality warning device for a speech semantic platform according to the present invention.

The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.

Detailed Description

It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

Referring to fig. 1, fig. 1 is a schematic structural diagram of a voice semantic platform exception warning device in a hardware operating environment according to an embodiment of the present invention.

As shown in fig. 1, the voice semantic platform exception alarm device may include: a processor 1001, such as a Central Processing Unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), and the optional user interface 1003 may further include a standard wired interface and a wireless interface, and the wired interface for the user interface 1003 may be a USB interface in the present invention. The network interface 1004 may optionally include a standard wired interface, a WIreless interface (e.g., a WIreless-FIdelity (WI-FI) interface). The Memory 1005 may be a Random Access Memory (RAM) Memory or a Non-volatile Memory (NVM), such as a disk Memory. The memory 1005 may alternatively be a storage device separate from the processor 1001.

Those skilled in the art will appreciate that the configuration shown in FIG. 1 does not constitute a definition of a speech semantic platform anomaly alerting device, and may include more or fewer components than shown, or some components in combination, or a different arrangement of components.

As shown in FIG. 1, memory 1005, identified as one type of computer storage medium, may include an operating system, a network communication module, a user interface module, and a voice semantic platform exception alert program.

In the speech semantic platform anomaly alarm device shown in fig. 1, the network interface 1004 is mainly used for connecting with a background server and communicating with the background server in data; the user interface 1003 is mainly used for connecting user equipment; the voice semantic platform abnormity alarm equipment calls a voice semantic platform abnormity alarm program stored in the memory 1005 through the processor 1001 and executes the voice semantic platform abnormity alarm method provided by the embodiment of the invention.

Based on the hardware structure, the embodiment of the abnormal alarm method of the voice semantic platform is provided.

Referring to fig. 2, fig. 2 is a schematic flow chart of a first embodiment of the abnormal alarm method for the voice semantic platform according to the present invention, and the first embodiment of the abnormal alarm method for the voice semantic platform according to the present invention is provided.

In a first embodiment, the method for alarming abnormality of a speech semantic platform comprises the following steps:

step S10: and sending an abnormality detection instruction to the voice semantic platform to be detected, so that the voice semantic platform to be detected feeds back target field information and current operation scene information according to a platform debugging log when receiving the abnormality detection instruction.

It should be understood that an execution main body of this embodiment is the voice semantic platform exception alarm device, where the voice semantic platform exception alarm device may be an electronic device such as a personal computer or a server, and may also be another device that can implement the same or similar functions.

It can be understood that the sending of the abnormality detection instruction to the to-be-detected voice semantic platform may be sending of the abnormality detection instruction to the to-be-detected voice semantic platform through a wireless network, for example, in a 5G, 4G, WIFI, or other manners; the abnormality detection instruction may also be sent to the to-be-detected speech semantic platform through a wired network, for example, an optical fiber, and the like, which is not limited in this embodiment.

It should be understood that the voice semantic platform to be detected may obtain the target field information from the platform debug log by two ways, namely, by accurately obtaining and path naming analysis, according to the platform debug log feedback target field information and the current operating scene information. Accurately acquiring, namely acquiring field values of fields which must be returned in the Debug log by adopting a fixed JSON nested content analysis mode; path naming resolution, namely for fields which are not necessarily returned in the Debug log, naming the fields in an underline cascading mode by adopting a JSON (Java Server pages) level nesting rule, for example, Errors _ NLU represents NLU fields under the Errors, and obtaining expected field values from the platform Debug log by resolving field naming. And storing the obtained field value in a structure form of Pandas Series, so as to facilitate subsequent result detection.

Step S20: and carrying out anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result.

It can be understood that, the abnormality detection is performed on the to-be-detected voice semantic platform according to the target field information, and the obtained abnormality detection result may be a text reply detection performed on the to-be-detected voice semantic platform according to the target field information to obtain a text reply detection result; acquiring current slot position information, and performing full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result; generating an abnormal detection result according to the text reply detection result and the full-value matching result;

or performing text reply detection on the to-be-detected voice semantic platform according to the target field information to obtain a text reply detection result; acquiring current slot position information, and performing full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result; acquiring an error log of the voice semantic platform to be detected; performing return value matching detection on the error log to obtain a return value matching result; and generating an abnormal detection result according to the text reply detection result, the full value matching result and the return value matching result.

It should be noted that the text label value may be a numerical value for labeling a text reply detection type; the current slot information may be key information of the user request at the current time, such as a user request type (type), a user request resource (resource), a user request set (album), a current slot identifier, and the like; the current slot position identification can be identification information used for marking a verification strategy; the error log is a text file used for recording error information during operation, and programmers, maintainers and the like can debug and maintain the system by using the error log; the List type result may be a List type result, and the return value may be a numerical value returned by the List type result; the preset expected value may be an expected value corresponding to each content in the List-type result.

Step S30: and performing abnormity alarm based on the current operation scene information and the abnormity detection result.

It can be understood that performing an exception alarm based on the current operation scenario information and the exception detection result may be determining a current failure rate, an accumulated failure number, and an error log content according to the exception detection result, and performing an exception alarm according to the current operation scenario information, the current failure rate, the accumulated failure number, and the current error log content.

It should be noted that the current failure rate is a ratio between the failure times at the current time and the total times; the accumulated failure times can be the total failure times from the beginning of the operation of the abnormal alarm device of the voice semantic platform to the current time.

It can be understood that, performing an exception alarm according to the current operation scene information, the current failure rate, the accumulated failure times, and the current error log content may be to search a failure rate threshold, an accumulated failure time threshold, and an important error log content corresponding to the current operation scene, determine whether the current failure rate is greater than the failure rate threshold, obtain a first determination result, determine whether the accumulated failure times is greater than the accumulated failure time threshold, obtain a second determination result, determine whether the current error log content is the important error log content, obtain a third determination result, and perform an exception alarm according to the current operation scene, the first determination result, the second determination result, and the third determination result.

In the first embodiment, an anomaly detection instruction is sent to a voice semantic platform to be detected, so that when the voice semantic platform to be detected receives the anomaly detection instruction, target field information and current operation scene information are fed back according to a platform debugging log; performing anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result; performing abnormity alarm based on the current operation scene information and the abnormity detection result; according to the method and the device, the voice semantic platform to be detected is subjected to abnormity detection according to the target field information, an abnormity detection result is obtained, and abnormity alarm is performed according to the current operation scene information and the abnormity detection result, so that various abnormity detections can be performed on the voice semantic platform, and abnormity alarm is performed according to the operation scene.

Referring to fig. 3, fig. 3 is a flowchart illustrating a second embodiment of the abnormal alarm method for a speech semantic platform according to the present invention, and the second embodiment of the abnormal alarm method for a speech semantic platform according to the present invention is proposed based on the first embodiment illustrated in fig. 2.

In the second embodiment, the step S20 includes:

step S201: and performing text reply detection on the to-be-detected voice semantic platform according to the target field information to obtain a text reply detection result.

It should be understood that, the text reply detection is performed on the to-be-detected speech semantic platform according to the target field information, and the text reply detection result may be obtained by determining a text label value according to the target field information, and detecting the target field information according to the text label value to obtain a text reply detection result.

It should be noted that the text annotation value may be a numerical value for marking the text reply detection type.

Step S202: and acquiring current slot position information, and performing full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result.

It can be understood that obtaining the current slot information and performing full-value matching detection on the target field information according to the current slot information may be obtaining the current slot information and determining a current slot identifier according to the current slot information, determining a target verification policy according to the current slot identifier, and performing full-value matching detection on the target field information according to the target verification policy to obtain a full-value matching result.

It should be noted that the current slot information may be key information of the user request at the current time, for example, a user request type (type), a user request resource (resource), a user request set (album), a current slot identifier, and the like; the current slot identification may be identification information for marking a check policy.

Step S203: and acquiring an error log of the voice semantic platform to be detected.

It should be noted that the error log is a text file used for recording error information during operation, and programmers and maintenance personnel and the like can use the error log to debug and maintain the system.

It should be understood that obtaining the error log of the voice semantic platform to be detected may be directly extracting the error log from the voice semantic platform to be detected.

Step S204: and performing return value matching detection on the error log to obtain a return value matching result.

It can be understood that, the return value matching detection is performed on the error log, and the obtaining of the return value matching result may be that a list type result is generated according to the error log, and a return value is determined according to the list type result to match the return value with a preset expected value, so as to obtain a return value matching result.

It should be noted that the List type result may be a List type result, and the return value may be a numerical value returned by the List type result; the preset expected value may be an expected value corresponding to each content in the List-type result.

Step S205: and generating an abnormal detection result according to the text reply detection result, the full value matching result and the return value matching result.

It should be understood that, the generating of the anomaly detection result according to the text reply detection result, the full-value matching result and the return-value matching result may be directly generating the anomaly detection result according to the text reply detection result, the full-value matching result and the return-value matching result.

In the second embodiment, the step S30 includes:

step S301: and determining the current failure rate, the accumulated failure times and the error log content according to the abnormal detection result.

It should be understood that the determining of the current failure rate, the accumulated failure times and the content of the error log according to the abnormal detection result may be counting the abnormal detection result, obtaining a statistical result, and determining the current failure rate, the accumulated failure times and the content of the error log according to the statistical result.

Step S302: and performing abnormity alarm according to the current operation scene information, the current failure rate, the accumulated failure times and the current error log content.

In a second embodiment, performing text reply detection on the to-be-detected voice semantic platform according to the target field information to obtain a text reply detection result; acquiring current slot position information, and performing full-value matching detection on the target field information according to the current slot position information to obtain a full-value matching result; acquiring an error log of the voice semantic platform to be detected; performing return value matching detection on the error log to obtain a return value matching result; generating an abnormal detection result according to the text reply detection result, the full value matching result and the return value matching result; the embodiment generates the abnormal detection result through text reply detection, full value matching detection and return value matching detection, thereby realizing abnormal detection with multiple scenes, multiple functions and multiple error types;

in a second embodiment, the current failure rate, the accumulated failure times and the error log content are determined according to the abnormal detection result, and abnormal alarm is performed according to the current operation scene information, the current failure rate, the accumulated failure times and the current error log content, so that the accuracy of abnormal alarm can be improved.

Referring to fig. 4, fig. 4 is a flowchart illustrating a voice semantic platform anomaly alarm method according to a third embodiment of the present invention, and the third embodiment of the voice semantic platform anomaly alarm method according to the present invention is proposed based on the second embodiment illustrated in fig. 3.

In the third embodiment, the step S201 includes:

step S2011: and determining a text marking value according to the target field information.

It should be understood that, determining the text label value according to the target field information may be analyzing the target field information to obtain an analysis result, and determining the text label value according to the analysis result.

Step S2012: and detecting the target field information according to the text label value to obtain a text reply detection result.

It can be understood that, the detecting the target field information according to the text label value, and obtaining the text reply detection result may be determining a text reply detection type according to the text label value, searching for a text detection policy corresponding to the text reply detection type, and detecting the target field information according to the text detection policy to obtain the text reply detection result.

Further, the step S2012 includes:

It should be noted that the text reply detection type may be at least one of a non-null text reply detection type, an expected value text reply detection type, and an unexpected value text reply detection type.

It should be understood that determining the text reply detection type according to the text annotation value may be finding a text reply detection type corresponding to the text annotation value.

In a specific implementation, for example, the text label value 0 corresponds to a non-empty text reply detection type; the text label value 1 corresponds to the expected value text reply detection type; the text label value 2 corresponds to an unexpected value text reply detection type.

It can be understood that the finding of the text detection policy corresponding to the text reply detection type may be finding the text detection policy corresponding to the text reply type in a preset mapping relation table, where the preset mapping relation table includes a correspondence relation between the text reply type and the text detection policy.

It should be understood that the text detection strategy corresponding to the non-empty text reply detection type is a non-empty text detection strategy; the text detection strategy corresponding to the expected value text reply detection type is an expected value text detection strategy; and the text detection strategy corresponding to the unexpected value text reply detection type is an unexpected value text detection strategy.

In a specific implementation, for example, the non-empty text detection policy may be to perform non-empty determination on the text reply, that is, the text reply must have content, otherwise, the text reply is regarded as an erroneous text reply; the expected value text reply detection type can be that text content corresponding to the text label value should not appear in the text reply, otherwise, the text reply is regarded as an error text reply; the unexpected value text reply detection type can be that text content corresponding to the text label value should appear in the text reply, otherwise, the text reply is regarded as an error text reply.

It can be understood that, the detecting the target field information by the text detection policy to obtain the text reply detection result may be determining a text reply according to the target field information and detecting the text reply by the text detection policy, where a specific detection manner is as described above to obtain the text reply detection result.

In a third embodiment, the step S202 includes:

step S2021: and acquiring current slot position information, and determining a current slot position identifier according to the current slot position information.

It should be understood that determining the current slot id according to the current slot information may be extracting the current slot information to obtain the current slot id.

Step S2022: and determining a target verification strategy according to the current slot position identification, and carrying out full-value matching detection on the target field information through the target verification strategy to obtain a full-value matching result.

It should be understood that, determining a target verification policy according to the current slot identifier, and performing full-value matching detection on the target field information through the target verification policy, where obtaining a full-value matching result may be determining whether the current slot identifier is a preset identifier, when the current slot identifier is the preset identifier, determining a target verification relationship according to the current slot identifier, determining a target verification policy according to the target verification relationship, and performing full-value matching detection on the target field information through the target verification policy, so as to obtain a full-value matching result;

or judging whether the current slot position identifier is a preset identifier, when the current slot position identifier is not the preset identifier, using a preset inspection strategy as a target detection strategy, and performing full-value matching detection on the target field information through the target verification strategy to obtain a full-value matching result.

Further, the step S2022 includes:

judging whether the current slot position mark is a preset mark or not;

It should be noted that the preset identifier may be a joint verification identifier, and the present embodiment is not limited thereto.

It should be understood that when the current slot identifier is the preset identifier, it indicates that joint verification needs to be performed at this time, and at this time, a target verification relationship needs to be determined according to the current slot identifier, for example, if the current slot identifier is type | album/type | resource, the target verification relationship is that "type" and "album" both meet the condition or "type" and "resource" both meet the condition.

It can be understood that, the determining of the target verification policy according to the target verification relationship may be to use a slot in the target verification relationship as a slot to be detected, and detect the slot to be detected as the target verification policy.

It should be understood that, the full-value matching detection is performed on the target field information through the target verification policy, and the full-value matching result may be obtained by determining user request information according to the target field information and matching the current slot position information with the user request information through the target verification policy.

Further, after determining whether the current slot identifier is a preset identifier, the method further includes:

It should be noted that the preset verification policy may be an independent verification policy, and the independent verification policy may be to perform independent verification on the current slot information.

It should be understood that, when the current slot identifier is not the preset identifier, it indicates that independent verification needs to be performed at this time, and at this time, an independent verification policy needs to be used as a target verification policy.

It can be understood that, the target field information is subjected to full-value matching detection through the target verification policy, and the full-value matching result is obtained by determining user request information according to the target field information and matching the current slot position information with the user request information through the target verification policy.

Also, for ease of understanding, reference is made to table 1 for illustration:

as shown in table 1, type, resource, and album are current slot information, and when no information is marked in the current slot identifier, all slots are independently checked, for example, Case1 needs to check whether the returned value of the slot of "resource" is "healthy song" for the request of "playing a healthy song" of the user, which means that the check is successful; when the current slot position identification is a preset identification, all marked slot positions need to be subjected to joint verification, for example, according to a request of 'playing the radish songgar' of a user by Case2, a target verification relation is analyzed according to the content of the current slot position identification, namely, a 'type' and an 'album' simultaneously meet a condition or a 'type' and a 'resource' simultaneously meet a condition, slot position joint verification is carried out according to the analyzed target verification relation, when a verification result meets the target verification relation, the verification is successful, otherwise, the verification fails; in addition, the check for the slot supports multiple result matching, that is, one slot can have multiple expected results, the test result can be matched with one of the values completely to indicate that the result meets the requirement, for example, Case3, and the detection for "album" can indicate that the check is successful as long as the value of "album" is "baby bus" or "theme song".

TABLE 1 full value matching Table

In a third embodiment, the step S204 includes:

step S2041: and generating a list type result according to the error log, and determining a return value according to the list type result.

It should be noted that the List type result may be a List type result, and the returned value may be a numerical value returned by the List type result.

It should be understood that generating the List type result according to the error log and determining the return value according to the List type result may be extracting the error log, obtaining a List type result, and receiving the return value returned by the List type result.

Step S2042: and matching the return value with a preset expected value to obtain a return value matching result.

It should be noted that the preset expected value may be an expected value corresponding to each content in the List-type result.

It should be understood that when the matching between the returned value and the preset expected value is successful, it indicates that no error occurs in the content corresponding to the preset expected value, and the matching result of the returned value is normal; and when the matching of the return value and the preset expected value fails, indicating that the content corresponding to the preset expected value is wrong, and judging that the matching result of the return value is abnormal.

In a third embodiment, a text label value is determined according to the target field information, and the target field information is detected according to the text label value to obtain a text reply detection result; in the embodiment, the text reply detection is carried out by extracting the text label value in the target field information, so that the text reply detection is more accurate;

in a third embodiment, a text reply detection type is determined according to the text label value, and a text detection strategy corresponding to the text reply detection type is searched; detecting the target field information through the text detection strategy to obtain a text reply detection result; in the embodiment, the text detection strategy is determined through the text label value, and the target field information is detected through the text detection strategy, so that the text reply detection is more reasonable and meets the user requirements;

in a third embodiment, current slot position information is obtained, a current slot position identifier is determined according to the current slot position information, a target verification strategy is determined according to the current slot position identifier, and full-value matching detection is performed on the target field information through the target verification strategy to obtain a full-value matching result; in the embodiment, the target verification strategy is determined through the front slot position identification, and the target field information is subjected to full-value matching detection through the target verification strategy, so that the full-value matching detection is more accurate;

in a third embodiment, it is determined whether the current slot identifier is a preset identifier; when the current slot position identification is the preset identification, determining a target verification relation according to the current slot position identification; determining a target checking strategy according to the target checking relation, and carrying out full-value matching detection on the target field information through the target checking strategy to obtain a full-value matching result; in the embodiment, the target inspection strategy is determined by comparing the current slot position identification with the preset identification, and full-value matching detection is performed through the target inspection strategy, so that the speed of full-value matching detection can be increased;

in a third embodiment, a list type result is generated according to the error log, a return value is determined according to the list type result, the return value is matched with a preset expected value, and a return value matching result is obtained, so that return value matching can be more accurate.

Referring to fig. 5, fig. 5 is a flowchart illustrating a fourth embodiment of the abnormal alarm method for a speech semantic platform according to the present invention, and the fourth embodiment of the abnormal alarm method for a speech semantic platform according to the present invention is provided based on the second embodiment illustrated in fig. 3.

In a fourth embodiment, the step S302 includes:

step S3021: and respectively searching a failure rate threshold value, an accumulated failure frequency threshold value and important error log contents corresponding to the current operation scene.

It should be understood that, the respectively searching for the failure rate threshold, the cumulative failure number threshold and the important error log content corresponding to the current operating scenario may be to search for the failure rate threshold corresponding to the current operating scenario, search for the cumulative failure number threshold corresponding to the current operating scenario, and search for the important error log content corresponding to the current operating scenario.

It should be noted that the failure rate threshold corresponding to the current operating scenario may be a numerical value set by the user according to actual requirements; the threshold value of the accumulated failure times corresponding to the current operation scene can be a numerical value set by a user according to actual requirements; the important error log content corresponding to the current operating scenario may be a log content set by a user according to an actual requirement, which is not limited in this embodiment.

Step S3022: and judging whether the current failure rate is greater than the failure rate threshold value or not, and obtaining a first judgment result.

It should be understood that when the current failure rate is greater than the failure rate threshold, the current failure rate is declared too high; and when the current failure rate is less than or equal to the failure rate threshold value, the current failure rate meets the condition.

Step S3023: and judging whether the accumulated failure times is greater than the threshold of the accumulated failure times or not, and obtaining a second judgment result.

It can be understood that when the accumulated failure number is greater than the threshold value of the accumulated failure number, the accumulated failure number is over-high; and when the accumulated failure times are less than or equal to the accumulated failure times threshold value, the accumulated failure times meet the condition.

Step S3024: and judging whether the current error log content is the important error log content or not, and obtaining a third judgment result.

It should be understood that the determining whether the current error log content is the important error log content may be matching the current log content with the important error log content to obtain a matching result, and determining whether the current error log content is the important error log content according to the matching result.

Step S3025: and performing abnormity alarm according to the current operation scene, the first judgment result, the second judgment result and the third judgment result.

It can be understood that, the performing of the abnormal warning according to the current operation scene, the first determination result, the second determination result, and the third determination result may be searching for a target early warning policy corresponding to the current operation scene, the first determination result, the second determination result, and the third determination result, and performing the abnormal early warning according to the target early warning policy.

Further, the step S3023 includes:

It should be noted that the target alarm policy may be to perform alarm notification according to a corresponding mail or bluetooth group configured by the user in advance.

It should be understood that the finding of the target early warning policy corresponding to the current operating scenario, the first determination result, the second determination result, and the third determination result may be determining an early warning policy to be selected according to the current operating scenario, and determining a target early warning policy from the early warning policy to be selected according to the first determination result, the second determination result, and the third determination result.

In a fourth embodiment, a failure rate threshold, an accumulated failure frequency threshold and important error log contents corresponding to the current operation scene are respectively searched; judging whether the current failure rate is greater than the failure rate threshold value or not, and obtaining a first judgment result; judging whether the accumulated failure times is greater than the accumulated failure time threshold value or not, and obtaining a second judgment result; judging whether the current error log content is the important error log content or not, and obtaining a third judgment result; performing abnormity alarm according to the current operation scene, the first judgment result, the second judgment result and the third judgment result; according to the embodiment, the current failure rate, the accumulated failure times and the current error log content are matched with the failure rate threshold, the failure times threshold and the important error log content, and abnormal alarm is performed according to the matching result, so that personalized alarm information can be generated, and irrelevant personnel are prevented from being interfered.

In addition, an embodiment of the present invention further provides a storage medium, where a voice semantic platform exception alarm program is stored on the storage medium, and when executed by a processor, the voice semantic platform exception alarm program implements the steps of the voice semantic platform exception alarm method described above.

In addition, referring to fig. 6, an embodiment of the present invention further provides a voice semantic platform anomaly alarm device, where the voice semantic platform anomaly alarm device includes: the device comprises a sending module 10, a detection module 20 and an alarm module 30;

the sending module 10 is configured to send an anomaly detection instruction to the to-be-detected voice semantic platform, so that when the to-be-detected voice semantic platform receives the anomaly detection instruction, target field information and current operation scene information are fed back according to a platform debugging log.

The detection module 20 is configured to perform anomaly detection on the to-be-detected speech semantic platform according to the target field information, and obtain an anomaly detection result.

The alarm module 30 is configured to perform an abnormal alarm based on the current operating scenario information and the abnormal detection result.

In this embodiment, an anomaly detection instruction is sent to a to-be-detected voice semantic platform, so that when the to-be-detected voice semantic platform receives the anomaly detection instruction, target field information and current operation scene information are fed back according to a platform debugging log; performing anomaly detection on the voice semantic platform to be detected according to the target field information to obtain an anomaly detection result; performing abnormity alarm based on the current operation scene information and the abnormity detection result; according to the method and the device, the voice semantic platform to be detected is subjected to abnormity detection according to the target field information, an abnormity detection result is obtained, and abnormity alarm is performed according to the current operation scene information and the abnormity detection result, so that various abnormity detections can be performed on the voice semantic platform, and abnormity alarm is performed according to the operation scene.

Other embodiments or specific implementation manners of the abnormality warning device for the voice semantic platform may refer to the above method embodiments, and are not described herein again.

It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, third, etc. do not denote any order, but rather the words first, second, third, etc. are to be interpreted as names.

Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention or portions thereof that contribute to the prior art may be embodied in the form of a software product, where the computer software product is stored in a storage medium (e.g., a Read Only Memory (ROM)/Random Access Memory (RAM), a magnetic disk, an optical disk), and includes several instructions for enabling a terminal device (e.g., a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.

The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes, which are made by using the contents of the present specification and the accompanying drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.

The invention discloses A1 and a voice semantic platform abnormity alarm method, which comprises the following steps:

A2, the method for alarming abnormality of a speech semantic platform according to A1, wherein the step of performing abnormality detection on the speech semantic platform to be detected according to the target field information to obtain an abnormality detection result, specifically includes:

A3, before the step of generating an abnormal detection result according to the text reply detection result and the full-value matching result as described in A2, the method further comprises:

acquiring an error log of the voice semantic platform to be detected;

A4, the method for alarming abnormality of speech semantic platform as A3, the step of performing return value matching detection on the error log to obtain a return value matching result, specifically comprising:

The method for alarming abnormality of a speech semantic platform according to a5, as described in A3, includes the steps of performing text reply detection on the speech semantic platform to be detected according to the target field information to obtain a text reply detection result, and specifically includes:

determining a text marking value according to the target field information;

A6, the method for alarming abnormality of a speech semantic platform as in a5, wherein the step of detecting the target field information according to the text label value to obtain a text reply detection result specifically includes:

A7, the method for alarming abnormality of a voice semantic platform as in A3, wherein the step of obtaining current slot information, performing full-value matching detection on the target field information according to the current slot information, and obtaining a full-value matching result specifically comprises:

A8, the method for alarming abnormality of a voice semantic platform as in a7, wherein the step of determining a target verification policy according to the current slot identifier, and performing full-value matching detection on the target field information through the target verification policy to obtain a full-value matching result specifically includes:

judging whether the current slot position mark is a preset mark or not;

A9, the method for alarming abnormality of voice semantic platform as in A8, wherein after the step of determining whether the current slot position identifier is a preset identifier, the method for alarming abnormality of voice semantic platform further comprises:

A10, in which the method for alarming abnormality of a speech semantic platform according to any one of a1-a9, the step of alarming abnormality based on the current operation scene information and the abnormality detection result includes:

A11, the method for alarming abnormality of voice semantic platform as in a10, wherein the step of alarming abnormality according to the current operation scene information, the current failure rate, the accumulated failure times and the current error log content specifically includes:

A12, the method for alarming abnormality of a speech semantic platform as in a11, wherein the step of alarming abnormality according to the current operation scene, the first determination result, the second determination result, and the third determination result specifically includes:

The invention discloses B13 and a voice semantic platform abnormity alarm device, which comprises: the system comprises a memory, a processor and a voice semantic platform exception alarm program stored on the memory and capable of running on the processor, wherein the voice semantic platform exception alarm program is configured to realize the steps of the voice semantic platform exception alarm method.

The invention discloses C14 and a storage medium, wherein a voice semantic platform abnormity alarm program is stored on the storage medium, and when being executed by a processor, the voice semantic platform abnormity alarm program realizes the steps of the voice semantic platform abnormity alarm method.

The invention discloses D15 and a voice semantic platform abnormity alarm device, which comprises: the device comprises a sending module, a detection module and an alarm module;

D16, the abnormal voice semantic platform alarm device according to D15, and the detection module, according to the target field information, performing text reply detection on the voice semantic platform to be detected to obtain a text reply detection result;

D17, the voice semantic platform anomaly alerting device of D16, further comprising: a matching module;

D18, the speech semantic platform abnormity warning device as D17, the matching module is further used for generating a list type result according to the error log, and determining a return value according to the list type result;

D19, the abnormality alarm device for speech semantic platform as D17, the detecting module is further used for determining a text label value according to the target field information;

D20, the abnormal alarm device for speech semantic platform as D19, the detecting module further configured to determine a text reply detection type according to the text label value, and search for a text detection policy corresponding to the text reply detection type;

Claims

1. An abnormity alarm method of a voice semantic platform is characterized by comprising the following steps:

2. The method for alarming abnormality of a speech semantic platform according to claim 1, wherein the step of performing abnormality detection on the speech semantic platform to be detected according to the target field information to obtain an abnormality detection result specifically comprises:

3. The method for alarming abnormality of speech semantic platform according to claim 2, wherein before the step of generating an abnormality detection result based on the text reply detection result and the full-value matching result, the method for alarming abnormality of speech semantic platform further comprises:

acquiring an error log of the voice semantic platform to be detected;

4. The method for alarming abnormality of a speech semantic platform according to claim 3, wherein the step of performing return value matching detection on the error log to obtain a return value matching result specifically comprises:

5. The method for alarming abnormality of a speech semantic platform according to claim 3, wherein the step of performing text reply detection on the speech semantic platform to be detected according to the target field information to obtain a text reply detection result specifically comprises:

determining a text marking value according to the target field information;

6. The method for alarming abnormality of a speech semantic platform according to claim 5, wherein the step of detecting the target field information according to the text label value to obtain a text reply detection result specifically comprises:

7. The method for alarming abnormality of a voice semantic platform according to claim 3, wherein the step of obtaining current slot information, performing full-value matching detection on the target field information according to the current slot information, and obtaining a full-value matching result specifically comprises:

8. An abnormality alarm device for a voice semantic platform, characterized by comprising: a memory, a processor and a voice semantic platform exception alert program stored on the memory and executable on the processor, the voice semantic platform exception alert program when executed by the processor implementing the steps of the voice semantic platform exception alert method of any of claims 1 to 7.

9. A storage medium having stored thereon a voice semantic platform exception alert program which, when executed by a processor, implements the steps of the voice semantic platform exception alert method of any one of claims 1 to 7.

10. The abnormal alarm device for the voice semantic platform is characterized by comprising the following steps: the device comprises a sending module, a detection module and an alarm module;