CN111831552A - Automatic testing method for real-time user behavior system - Google Patents

Abstract

The invention provides an automatic testing method for a real-time user behavior system, which comprises the following steps: step 1: recording production flow and adding the production flow into a test data set; step 2: adding mock abnormal data into a test data set aiming at the real-time data processing node; and step 3: importing the test data set into a real-time user behavior system to serve as an input parameter of a real-time data processing node, and exporting an output parameter of the real-time data processing node; and 4, step 4: checking the data format of the real-time data processing node, and carrying out accuracy check on the data format and the imported data; and 5: adding the abnormal data of the next-stage real-time data processing node into the output parameter of the real-time data processing node as the input parameter of the next-stage real-time data processing node; step 6: and repeating the steps 4-5 until the test data set is processed by all the real-time data processing nodes. After the system is divided into a plurality of subsystems, the intermediate flow can be covered, and the overall test coverage rate is improved greatly.

Description

Automatic testing method for real-time user behavior system

Technical Field

The invention relates to the field of automatic testing, in particular to an automatic testing method for a real-time user behavior system.

Background

For an internet application, services are various and relatively independent, each service has a behavior dotting reporting and analyzing system, a real-time portrait of a user needs to be constructed for a fine operation scene and sensing the behavior of the user in real time, a uniform real-time user behavior system is derived, layer-by-layer processing is carried out in a data flow mode, and finally, uniform standard data is formed and a real-time user behavior portrait is formed. As shown in fig. 2.

For such internally closed systems, the usual funnel-shaped (see-and-go) test protocols are not applicable:

the first reason is as follows: the process is long, and once a certain node has a problem, abnormal positioning is not easy to occur;

the second reason is that: each node may have abnormal data, and the processing of the abnormal data by the next node cannot be covered.

Disclosure of Invention

Aiming at the defects of the prior art, the invention provides an automatic testing method for a real-time user behavior system, which can improve the overall testing coverage rate.

In order to achieve the purpose, the invention adopts the following technical scheme:

an automated testing method for a real-time user behavior system, comprising the steps of:

step 1: recording production flow and adding the production flow into a test data set;

step 2: adding mock abnormal data into a test data set aiming at the real-time data processing node;

and step 3: importing the test data set into a real-time user behavior system to serve as an input parameter of a real-time data processing node, and exporting an output parameter of the real-time data processing node;

and 4, step 4: checking the data format of the real-time data processing node, and carrying out accuracy check on the data format and the imported data;

and 5: adding the abnormal data of the next-stage real-time data processing node into the output parameter of the real-time data processing node as the input parameter of the next-stage real-time data processing node;

step 6: and repeating the steps 4-5 until the test data set is processed by all the real-time data processing nodes.

Has the advantages that:

1. compared with the conventional automatic testing method, only the inlet and the outlet of the system are concerned, the intermediate flow is difficult to cover, and the intermediate flow can be covered after the system is divided into a plurality of subsystems, so that the overall testing coverage rate is improved greatly.

2. The problem positioning is facilitated, once a problem occurs in a system with a long flow, the problem cannot be quickly positioned, and through the segmentation system, each node is independently checked, so that the problem is positioned more easily.

3. The robustness of each node is guaranteed, the system is a continuous iteration process, each node can correctly process abnormal data of the node under normal conditions, but if the abnormal data flows into the next node in the continuous iteration process or under other abnormal conditions, the data processing flow of the whole system is likely to be broken if the next node is not robust enough. Therefore, by slicing, adding anomaly data to each child node, the robustness of each child node can be verified.

Drawings

FIG. 1 is a flow chart of a method for automated testing of a real-time user behavior system of the present invention;

FIG. 2 is a schematic diagram of a conventional real-time user behavior system;

fig. 3 is a schematic structural diagram of a real-time user behavior system according to the present invention.

Detailed Description

The following further describes the testing method and data processing with reference to examples, but the embodiments of the present invention are not limited thereto.

As shown in fig. 1, the present invention employs the following test protocol:

firstly, by using the concept of software slicing, slicing a real-time user behavior system through a middleware, and slicing the system to form a plurality of sub-nodes, wherein each sub-node has independent input and output parameters;

recording the inlet flow (original data) of the production environment through a tool, mainly reporting data of the behavior of each service, and adding the data into a test data set;

analyzing each node, and according to the entry format of each node and in combination with the test boundary, mock outputs an abnormal test data set of each node in a common mode as follows: the method comprises the steps of setting a field to be null, setting a field format to be abnormal, lacking the field, increasing the field, exceeding a threshold value and the like, and meanwhile, adding some specific abnormal data according to the implementation logic inside the node and adding the abnormal data into an abnormal test data set;

fourthly, the flow recorded in the production environment and the abnormal data set of the node 1 are led into a system inlet (middleware A), and in the process of processing the flow data, the interruption is not carried out, but the output parameter (middleware B) of the node 1 is led out and processed subsequently by adding consumers, so that the flow direction of the real-time system data is ensured;

and fifthly, checking the data format of the node 1 (middleware B), carrying out 1-to-1 accuracy check on the data and the imported data, checking whether the abnormal data is correctly processed or not, avoiding the situation that the abnormal data is continuously lost to the next node, checking a system log, and judging whether the abnormality is thrown or not.

Adding the abnormal data of the node 2 into the output parameter of the node 1 (middleware B) as the input parameter of the node 2 (middleware C), wherein in the process, the output parameter of the node 1 is consumed by the node 2 naturally as a real-time data stream of the whole system, and in the actual operation process, only the abnormal data of the node 2 needs to be led into the middleware B, and in the stream data processing process, the interruption is not carried out, but the output parameter of the node 2 (middleware C) is led out and subsequently processed by adding consumers, so that the data flow direction of the real-time system is ensured;

and seventhly, checking the data format of the node 2, performing 1-to-1 accuracy check on the data and the imported data, checking whether abnormal data is correctly processed or not, avoiding the situation that the abnormal data is continuously lost to the next node, checking a system log, and judging whether the abnormal data is thrown or not.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (1)

1. An automated testing method for a real-time user behavior system is characterized by comprising the following steps:

step 1: recording production flow and adding the production flow into a test data set;

step 2: adding mock abnormal data into a test data set aiming at the real-time data processing node;

and step 3: importing the test data set into a real-time user behavior system to serve as an input parameter of a real-time data processing node, and exporting an output parameter of the real-time data processing node;

and 4, step 4: checking the data format of the real-time data processing node, and carrying out accuracy check on the data format and the imported data;

and 5: adding the abnormal data of the next-stage real-time data processing node into the output parameter of the real-time data processing node as the input parameter of the next-stage real-time data processing node;

step 6: and repeating the steps 4-5 until the test data set is processed by all the real-time data processing nodes.

Images