Integrated test version control linkage method and system based on quality access control

Abstract

The application discloses a quality gate-based integrated test version control linkage method and system, relates to the technical field of software quality management, and comprises the following steps: after each time a code module participates in gate detection, the gate detection result is stored in a historical database in a structured manner; the value sequence of each quality index changing with the version is extracted from the historical record in the unit of module, and the time sequence of the quality index of each module is constructed; the gate decision result is automatically synchronized to the version control system through an interface, linkage operations such as label marking, state setting, comment adding, problem creating and notification sending are automatically executed for four decision types, a complete artificial approval exemption process is provided for trend prediction interception, the artificial transmission link between the gate decision and the version control is eliminated, the decision result is ensured to be real-time visible, and complete closed-loop management from quality detection to version control is formed.

Description

Technical Field

[0001] This invention belongs to the field of software quality management technology, specifically an integrated testing version control linkage method and system based on quality access control. Background Technology

[0002] Software integration testing is a crucial step in the software development process to ensure code quality. Common quality control methods in current engineering projects mainly include setting quality gates before code merging, conducting actual tests on indicators such as code coverage, test pass rate, static scan defect count, and cyclomatic complexity, and allowing merging if the indicators meet preset thresholds; otherwise, merging is rejected.

[0003] However, existing technologies for quality access control of code versions typically rely solely on comparing the measured values ​​of current version metrics with fixed thresholds. This static, single-point judgment method is ineffective when a quality metric shows a slow, continuous decline across multiple consecutive versions, yet still barely exceeds the threshold. The access control system cannot detect this gradual deterioration trend and only triggers interception after the metric actually falls below the threshold, by which time a large amount of potentially problematic code has already entered the main codebase. Secondly, existing access control decisions only offer two outcomes: pass or reject. They cannot distinguish between two fundamentally different risk states: current quality meets the standard but the trend is continuously deteriorating, and current quality meets the standard and the trend is healthy. This results in access control decisions lacking the ability to predict future risks. Based on these issues, this paper proposes an integrated test version control linkage method and system based on quality access control. Summary of the Invention

[0004] The purpose of this invention is to provide an integrated test version control linkage method and system based on quality access control, so as to solve the problems mentioned in the background art.

[0005] Integration test version control linkage methods based on quality access control include: Step 1: After each access control test is completed by the code module, the access control test results are stored in the historical database in a structured manner. The numerical sequence of each quality indicator as the version changes is extracted from the historical records by module, and the time series of quality indicators of each module is constructed. Step 2: Based on the time series of quality indicators, calculate the trend slope and oscillation coefficient for each quality indicator. Determine the trend type of each quality indicator based on the trend slope and oscillation coefficient. The trend types include continuous deterioration, oscillating deterioration, continuous improvement, continuous stability, and recent sudden change. Step 3: Based on the trend slope of each quality indicator, first determine whether each quality indicator belongs to the type of higher quality is better or lower quality is better. Then, calculate the predicted value and prediction confidence interval of each quality indicator after K iterations. Finally, based on the type label, predicted value, prediction confidence interval and preset access threshold, determine the risk level of each quality indicator as having foreseeable risk, having low risk warning or predicting compliance. Step 4: Compare the measured values ​​of each quality indicator in the current version with the access control threshold. Combine the risk level judgment results and trend type, and determine the differentiated access control decision according to the priority from high to low. The access control decision types include traditional rejection, trend predictive interception, trend warning access, and traditional access. Step 5: Automatically synchronize the generated access control decision results to the version control system via the interface, and execute the corresponding version control linkage operation for each of the four decision types.

[0006] As a further aspect of the present invention, the specific method for constructing the time series of quality indicators for each module is as follows: The system queries the access control history database to retrieve historical access control records for this module, sorting them from oldest to newest by submission time. For each quality indicator, it extracts the measured values ​​from each version and constructs a time series from the most recent N valid records, where N is 15. If there are fewer than 15 records, all historical records are used. Valid records refer to records where the corresponding module has actually undergone code changes or participated in test execution in that version. If a module has fewer than 3 valid historical records, the trend determination is skipped, and only the current value is checked. If a quality indicator has more than 30% missing values ​​in the time series, it is not included in the trend analysis. This process is used to obtain the time series of quality indicators for each module.

[0007] As a further aspect of the present invention, the specific method for calculating the trend slope and oscillation coefficient is as follows: Assign a location number to each data point in the time series, calculate the mean of the location numbers and the mean of the index value, calculate the location deviation and value deviation for each data point, and sum the product of the location deviations and value deviations of all data points to obtain a total A. Sum the squares of the location deviations of all data points to obtain a total B. The trend slope is equal to the total A divided by the total B. Calculate the difference for each pair of adjacent data points in the quality index time series, and generate a sign sequence based on the sign of the difference. A difference greater than 0 is recorded as positive, less than 0 as negative, and equal to 0 uses the sign of the previous difference. Count the total number of sign alternations in the sign sequence. The oscillation coefficient is equal to the total number of sign alternations divided by the number of adjacent differences.

[0008] As a further aspect of the present invention, the specific method for determining the trend type of each quality indicator is as follows: For each quality indicator, a trend judgment threshold T1 and a sudden change judgment threshold T2 are preset. When the trend slope is negative and its absolute value is greater than T1, and the oscillation coefficient is less than 0.3, it is judged as continuous deterioration. When the trend slope is negative and its absolute value is greater than T1, and the oscillation coefficient is greater than or equal to 0.3, it is judged as oscillating deterioration. When the trend slope is positive and its absolute value is greater than T1, it is judged as continuous improvement. When the absolute value of the trend slope is less than or equal to T1, it is judged as continuous stability. When the absolute value of the difference between the last two data points of the time series is greater than T2, and the absolute value of the trend slope is less than T1, it is judged as a recent sudden change, and the judgment of recent sudden changes has the highest priority.

[0009] As a further aspect of the present invention, the specific method for determining the type of quality indicator is as follows: A pre-defined quality indicator type mapping table is used, where code coverage and test pass rate belong to the type where higher values ​​mean better quality, while static scan defect count and average cyclomatic complexity belong to the type where lower values ​​mean better quality. When an indicator to be judged falls into the corresponding list, it is assigned a corresponding mark; otherwise, it is automatically identified according to the access control threshold expression of the indicator. If the access control threshold is greater than or equal to or greater than the form, it is judged as the type where higher values ​​mean better quality; if it is less than or equal to or less than the form, it is judged as the type where lower values ​​mean better quality.

[0010] As a further aspect of this invention, the specific method for calculating the predicted values ​​and prediction confidence intervals of each quality indicator after K iterations is as follows: The predicted value is equal to the current measured value plus the product of the trend slope and the prediction step size K, where the default value of K is 3. For each data point in the quality index time series, the predicted value of the trend line is calculated as the mean of the index value plus the product of the trend slope and the corresponding position deviation. The residual is equal to the actual quality index value of the data point minus the predicted value of the trend line. The standard deviation of the residuals of all data points is calculated. The lower bound of the prediction confidence interval is equal to the predicted value minus twice the standard deviation of the residuals, and the upper bound is equal to the predicted value plus twice the standard deviation of the residuals.

[0011] As a further aspect of the present invention, the specific method for determining the risk level of each quality indicator is as follows: If the quality indicator is categorized as "higher value, better quality," and the predicted value is below the access control threshold, and the lower bound of the prediction confidence interval is also below the access control threshold, then a foreseeable risk is identified. If the predicted value is above the access control threshold but the lower bound of the prediction confidence interval is below the access control threshold, then a low-risk warning is identified. If both the predicted value and the lower bound of the prediction confidence interval are above the access control threshold, then the prediction is considered satisfactory.

[0012] As a further aspect of the present invention, the specific method for determining differentiated access control decisions is as follows: First priority: If the current measured value of any quality indicator does not meet the corresponding access control threshold, then traditional rejection is executed. Second priority: If the measured values ​​of all quality indicators meet the access control threshold, and any quality indicator simultaneously meets the following conditions: the risk level judgment result indicates foreseeable risk, the trend type is continuous deterioration, and the safety margin is less than twice the absolute value of the trend slope of the indicator, then trend predictive blocking is executed. Among these, for indicators of the type where higher values ​​indicate better quality, the safety margin is equal to the current measured value minus the access control threshold; for indicators of the type where lower values ​​indicate better quality, the safety margin is equal to the access control threshold minus the current measured value. Third priority: If the second priority is not triggered and the risk judgment result of any quality indicator indicates a low-risk warning, or the trend type of any quality indicator is oscillating deterioration or recent abrupt change, then trend warning is passed, and the number of warnings is recorded. If the same quality indicator triggers trend warnings three times consecutively, then when the indicator triggers a warning again for the fourth time, it is automatically upgraded to trend predictive blocking. Fourth priority: If none of these conditions are triggered, then traditional pass is executed.

[0013] As a further aspect of the present invention, the specific method for performing corresponding version control linkage operations for the four decision types is as follows: When the decision is "Traditional Approval," the merge request is tagged with "Quality Gated Approval," the status is set to "Approved," and code merging is allowed. When the decision is "Traditional Rejection," the merge request is tagged with "Quality Gated Rejection," the status is set to "Blocked," the names, measured values, and gate thresholds of the non-compliant metrics are listed in the comments section, code merging is blocked, and the submitter is notified. When the decision is "Trend Alert Approval," the merge request is tagged with "Quality Gated Trend Alert Allowed," the status is set to "Approved but Alert Issued," the metric information that triggered the alert is added to the comments section, a tracking issue is created and associated with the merge request, and code merging is allowed. Send early warning notifications to project leaders and module responsible persons, requiring them to submit rectification plans within 5 working days; when the decision is a trend-predictive interception, tag the merge request with "Quality Governance Trend Interception", set the status to "Blocked - Trend Risk", add detailed interception descriptions in the comments section, create a high-priority issue, block code merging, and send high-priority notifications to the submitter, module leader, project manager, and quality team; for trend-predictive interception decisions, provide a manual approval exemption process. After approval, the system automatically removes the merge blocking status and updates the merge request status to "Approved - Exempted".

[0014] The quality access control-based integration test version control linkage system is implemented through a quality access control-based integration test version control linkage method, including: The time series construction module is used to structure and store the version identifier, measured values ​​of module quality indicators, and access control decision results after each access control detection is completed, and to extract the time series of each quality indicator of the module. The trend feature calculation and type determination module is used to calculate the trend slope and oscillation coefficient of the time series of each quality indicator and determine the trend type. The predictive risk identification module is used to determine the type label of quality indicators, calculate the predicted value and prediction confidence interval based on the trend slope, and determine the risk outcome. The differentiated access control decision module is used to compare the current measured value with the access control threshold, and generate access control decisions according to priority based on the risk level judgment result and trend type. The version control linkage execution module is used to automatically execute corresponding tag setting, status change, comment addition, issue creation and notification sending operations through the version control system interface according to the type of access control decision.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention, by storing the test results in a structured manner into a historical database after each access control test, constructs a time series of quality indicators by module, establishing a complete historical record of quality evolution for each module; by introducing two-dimensional features of trend slope and oscillation coefficient to quantify the rate and regularity of indicator changes, and subdividing trend types into five types: continuous deterioration, oscillating deterioration, continuous improvement, continuous stability, and recent abrupt change, compared to the traditional single-point judgment relying solely on current measured values, it can more accurately distinguish between two fundamentally different deterioration modes: stable continuous deterioration and fluctuating overall deterioration, providing a more accurate basis for subsequent differentiated decision-making; by calculating predicted values ​​through linear extrapolation based on trend slope and constructing a prediction confidence interval by combining historical residual standard deviation, the uncertainty of the prediction is incorporated into risk assessment, enabling early identification of potential risks that currently meet the standards but are foreseeably falling below the access control threshold in the future; By establishing a four-level priority decision-making system—traditional rejection, trend-predictive interception, trend-warning approval, and traditional approval—the system breaks through the inherent logic of traditional access control that allows access as soon as the current standard is met. It proactively intervenes when the quality indicators are currently at the qualified level but the trend is approaching the threshold. Trend-warning approval provides a response of allowing access with a warning for moderate risks. The automatic escalation mechanism for three consecutive warnings effectively prevents warnings from being ignored for a long time, achieving precise hierarchical handling of different risk levels. By automatically synchronizing access control decision results to the version control system via an interface, the system automatically performs linked operations such as tagging, status setting, comment addition, issue creation, and notification sending for the four decision types. It also provides a complete manual approval exemption process for trend-predictive interception, eliminating the manual transmission link between access control decision-making and version control, ensuring that decision results are visible in real time, and forming a complete closed-loop management from quality inspection to version control. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method framework structure of the present invention; Figure 2 This is a schematic diagram of the system framework structure of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1 This application provides an integrated test version control linkage method based on quality access control, including: Step 1: Collect access control history data from the code module multiple times to construct a time series of quality indicators; After each code module submission passes the quality access control check, the result data of this access control check is automatically stored in a structured manner. Then, on a module-by-module basis, the numerical sequence of the quality indicators of the module as a result of version changes is extracted from the stored historical records to construct the quality indicator time series of the module. This time series records the evolution trajectory of the module quality over time and serves as the data foundation for subsequent trend analysis.

[0019] Quality metrics include code coverage, test pass rate, number of static scan defects, and average cyclomatic complexity. Access control history collection is automatically triggered based on the CI / CD process when developers submit code to the version control system, without requiring manual intervention. In addition, the data is stored immediately after each access control test to ensure the integrity of historical data.

[0020] After each access control test involving a specific code module is completed, the system stores the following fields in a structured manner in the access control history database: a unique version identifier such as version number v2.3.15, a timestamp of the submission time, a list of code modules involved in the modification of this version such as PaymentService and OrderService, the measured values ​​of various quality indicators of this module in this test, including coverage percentage, test pass percentage, number of defects, cyclomatic complexity, the final decision of the access control system as pass or rejection, and if rejection, the specific reason for rejection is recorded as which indicator did not meet the standard.

[0021] The storage record example format is as follows: Record number REC-20240115-001, version ID v2.3.15, submission time January 15, 2024, 14:30:25, involving modules PaymentService and OrderService. PaymentService has a coverage of 82%, a pass rate of 95%, 2 defects, and a cyclomatic complexity of 8.5. OrderService has a coverage of 78%, a pass rate of 92%, 3 defects, and a cyclomatic complexity of 7.2. The access control decision is passed.

[0022] For each code module, the system retrieves all historical access control records containing that module from the access control history database, sorts them by submission time from oldest to newest, and extracts the values ​​for each quality indicator across all versions to form a time series for that indicator. The length of the time series is the most recent N valid records; here, the default value for N is set to 15, meaning data from the most recent 15 versions is used. This parameter can be adjusted according to project requirements. If the total historical records for a module are less than 15, then all historical records are used.

[0023] It's important to note that "valid records" here refers to records showing that the module actually underwent code changes or participated in test execution within that version. If the module remained completely unchanged in a version, with no code modifications or tests run, such records are not considered valid and will not be included in the time series.

[0024] By storing historical data from each access control detection in a structured manner, a complete record view of quality indicators changing over time was established. A module-level time series construction method was adopted, ensuring that each module has its own independent quality evolution history, providing a data foundation for subsequently identifying quality change trends in different modules. The advantage of this approach is that each module can be monitored separately, preventing quality fluctuations in other modules from affecting the judgment of the current module.

[0025] If a module has fewer than 3 valid historical records, it indicates that the module is too new or has been rarely modified, and there is insufficient data for trend analysis. In this case, the module will skip trend determination and only perform routine current value detection. If a quality indicator has more than 30% missing values ​​in the time series, such as 6 missing values ​​out of 15 records, the reliability of this indicator is too poor and it will not be included in trend analysis.

[0026] By immediately storing the test results in a structured format into a historical database after each access control test, and extracting the measured values ​​of various quality indicators in chronological order by module, a time series of quality indicators for that module is constructed. The core value of this approach lies in connecting the isolated access control test results of the past into a continuous trajectory of quality evolution, allowing the system to have a complete memory of the quality change history of each module. At the same time, through an effective record filtering mechanism, version records not involved in the module are excluded, ensuring the data quality of the time series and providing a reliable data foundation for subsequent trend calculations.

[0027] Step 2: Analyze the time series of quality indicators, calculate the trend slope and oscillation coefficient, and determine the trend type; Based on the quality indicator time series constructed in Step 1, two key trend feature values ​​are calculated for each quality indicator in each module: one is the trend slope, used to quantify the average rate of change of the indicator with version iterations; the other is the oscillation coefficient, used to measure the frequency of fluctuations during the indicator's evolution. Then, these two feature values ​​are combined to determine the trend type, which includes continuous deterioration, oscillating deterioration, continuous improvement, continuous stability, and recent abrupt change. This classification result directly affects subsequent access control decisions.

[0028] The specific method for calculating the trend slope and oscillation coefficient is as follows: S1: For a single quality indicator, assign a position number to each data point in the time series. The position number of the first data point is 1, the second is 2, and so on, with the position number of the Nth data point being N. This position number actually represents the chronological order of the versions. S2: Calculate the average of the location numbers, denoted as the location number mean. The specific calculation method is as follows: add all the location numbers from 1 to N, that is, calculate the sum of 1 plus 2 plus 3 up to N. This sum is equal to N multiplied by N plus 1 and then divided by 2. Then divide this sum by the total number of data points N to obtain the location number mean. S3: Calculate the average value of the quality indicators, denoted as the indicator mean. The specific calculation method is as follows: sum the quality indicator values ​​of all N data points in the time series to obtain the total quality indicator value, and then divide the total value by the total number of data points N to obtain the indicator mean. S4: For each data point in the time series, calculate its positional deviation and value deviation: the positional deviation is equal to the position number of the data point minus the mean of the position numbers; the value deviation is equal to the quality index value of the data point minus the mean of the index values. S5: For each data point in the time series, multiply the positional deviation of the data point by the value deviation of the data point to obtain the deviation product of the data point; add the deviation products of all data points in sequence to obtain the sum of the deviation products, denoted as the sum A; S6: For each data point in the time series, multiply the positional deviation of the data point by itself to obtain the square of the positional deviation of the data point; sum the squares of the positional deviations of all data points in sequence to obtain the sum of the squares of the positional deviations, denoted as the sum B; S7: The trend slope equals the sum A divided by the sum B; if the slope is positive, it indicates that the quality indicator is on an upward trend over time; if it is negative, it indicates that the quality indicator is on a downward trend; the meaning of the trend slope is how many units the quality indicator changes on average with each version iteration. S8: Perform the calculations from S1 to S7 for each quality indicator to obtain the trend slope corresponding to each quality indicator. The specific method for calculating the oscillation coefficient is as follows: S01: For a single quality indicator, calculate the difference between adjacent data points in the time series. For a data point with position number i and a data point with position number i+1, the adjacent difference is equal to the quality indicator value of the data point with position number i+1 minus the quality indicator value of the data point with position number i. If this difference is positive, it means that the quality has increased from version i to version i+1; if it is negative, it means that the quality has decreased. S02: Determine the sign of each difference. A difference greater than 0 is recorded as positive, a difference less than 0 is recorded as negative, and a difference equal to 0 retains the sign of the previous difference. This results in a sequence of signs, such as: negative, negative, negative, positive, negative, negative, positive, negative, negative. S03: Count the number of alternations of signs. Starting from the beginning of the sign sequence, each time a sign changes, i.e. from positive to negative or from negative to positive, increment the count by 1. S04: The oscillation coefficient is equal to the number of sign alternations divided by the number of adjacent differences; for example, if there are 14 adjacent differences in 15 data points, and the alternation occurs 3 times, the oscillation coefficient is equal to 3 divided by 14, which is approximately 0.21. The oscillation coefficient ranges from 0 to 1. If the oscillation coefficient is close to 1, it means that the sign changes almost every time, and the indicator fluctuates frequently, sometimes good and sometimes bad; if it is close to 0, it means that the sign remains basically unchanged, and the indicator changes steadily in one direction. The oscillation coefficient is used to determine whether the quality indicator changes steadily in one direction or fluctuates repeatedly, sometimes good and sometimes bad. S05: Perform the calculations from S01 to S04 for each quality indicator to obtain the oscillation coefficients corresponding to each quality indicator. The specific method for determining the trend type is as follows: First, the trend judgment threshold T1 and mutation judgment threshold T2 corresponding to each quality indicator are set as follows: the coverage rate is set to 0.5% for T1 and 5% for T2; the pass rate is set to 0.5% for T1 and 5% for T2; the number of defects is set to 0.3 for T1 and 3 for T2; and the cyclomatic complexity is set to 0.1 for T1 and 1.0 for T2. For a single quality indicator, if the trend slope corresponding to the quality indicator is negative, and the absolute value of the slope is greater than the corresponding T1 threshold, and the oscillation coefficient is less than 0.3, it is judged as a continuous deterioration trend type, which means that the quality indicator continues to decline, and the decline process is relatively stable, rather than fluctuating. When the trend slope corresponding to the quality indicator is negative, and the absolute value of the slope is greater than the corresponding T1 threshold, and the oscillation coefficient is greater than or equal to 0.3, it is judged as an oscillating deterioration trend type, which means that the overall quality indicator is declining, but there are frequent fluctuations in the process, sometimes rising and sometimes falling. When the slope of the trend corresponding to the quality indicator is positive and the absolute value of the slope is greater than the corresponding T1 threshold, it is determined to be a continuous improvement trend type, which means that the quality indicator is continuously rising and the quality is improving. When the absolute value of the trend slope corresponding to the quality indicator is less than or equal to the corresponding T1 threshold, it is determined to be a continuous stable trend type, which means that the quality indicator does not have an obvious upward or downward trend and remains stable. When the absolute value of the difference between the last two data points of the quality indicator time series is greater than the corresponding T2 threshold, and the absolute value of the long-term trend slope is less than the corresponding T1 threshold, it is determined to be a recent sudden trend type, which means that the quality indicator has been stable for a long time, but the most recent version has suddenly changed significantly. The above judgment is performed on each quality indicator to obtain the trend type corresponding to each quality indicator. It should be noted that the determination of recent mutation has the highest priority. As long as a quality indicator meets the conditions for recent mutation, it is determined to be a recent mutation, overriding other determination results.

[0029] Based on the time series constructed in step one, the trend slope is calculated for each indicator using a linear trend estimation method, quantifying the average rate of change of the indicator in each version iteration. Simultaneously, the oscillation coefficient is calculated by statistically analyzing the number of alternations of adjacent difference signs, characterizing the stability of the indicator's direction of change. Combining these two methods, and matching downwards according to the highest priority of recent abrupt changes, five trend type labels are derived: continuous deterioration, oscillating deterioration, continuous improvement, continuous stability, and recent abrupt changes. This two-dimensional judgment method, which quantifies both the rate of change and the regularity of change, can more accurately distinguish between two fundamentally different deterioration patterns—stable continuous deterioration and fluctuating but overall deteriorating—compared to the one-dimensional judgment that only considers the slope, providing a more granular basis for subsequent differentiated decision-making.

[0030] Step 3: Extrapolate the trend slope to predict the quality status of future versions and identify foreseeable risks; Based on the trend slopes of the various quality indicators calculated in step two, we first determine whether each quality indicator belongs to the type where higher values ​​generally indicate better quality or lower values ​​generally indicate better quality. Then, we predict the values ​​of each quality indicator for the current module in the next K version iterations. Finally, we determine whether there is a risk that each quality indicator meets the quality standards in the current version but will foreseeably exceed the access control threshold in the future. This step is the key basis for achieving predictive access control intervention, enabling the system to have the ability to anticipate unforeseen circumstances.

[0031] The prediction method is based on the assumption of linear extrapolation, which assumes that the quality indicators will continue their current trend in the short term. The default value for the prediction step size K is set to 3 iterations, corresponding to a time span of about 4 to 6 weeks for most projects; this parameter is configurable. The prediction result includes the predicted value and the prediction confidence interval, where the confidence interval reflects the uncertainty of the prediction. The specific method for determining the type of quality indicator is as follows: T1: For a single quality indicator, query the name or identifier of the indicator from the pre-configured quality indicator type mapping table; The quality metric type mapping table includes metrics with higher numerical values ​​indicating better quality and metrics with lower numerical values ​​indicating better quality. Specifically, code coverage, test pass rate, static scan defect count, and average cyclomatic complexity all belong to the lower quality category. T2: If the indicator is in the list of the "higher the value, the better the quality" type, then mark the indicator type as "higher the value, the better the quality". T3: If the indicator is in the list of types where lower values ​​mean better quality, then mark the indicator type as "lower values ​​mean better quality". T4: If the indicator is not found in the mapping table, obtain the access control threshold expression for the quality indicator. If the access control threshold expression is greater than or equal to a certain value, the indicator type is determined to be that the higher the value, the better the quality; if the access control threshold expression is less than or equal to a certain value, the indicator type is determined to be that the lower the value, the better the quality. T5: Perform judgments from T1 to T4 on each quality indicator to obtain the type label corresponding to each quality indicator; The specific method for calculating the predicted values ​​of future quality indicators is as follows: P1: For a single quality indicator, obtain the value of the last data point in the time series of that quality indicator in step one, that is, the measured value of the quality indicator corresponding to the data point with position number N in the time series, and use it as the measured value of the current version, and mark it as the current measured value; P2: Obtain the trend slope of this quality indicator calculated in step two; P3: Set the prediction step size K, which indicates how many iterations to predict. The default K is 3. The predicted value is calculated as follows: the predicted value equals the current measured value plus the product of the trend slope and the prediction step size K, that is, the predicted value equals the current measured value plus the trend slope multiplied by K; P4: Perform calculations from P1 to P3 for each quality indicator to obtain the predicted values ​​for each quality indicator. This calculation method predicts the value that the quality metric might reach after K iterations, based on the current trend.

[0032] The specific method for calculating the prediction confidence interval is as follows: C1: For a single quality indicator, obtain all data points of the time series of that quality indicator in step one; C2: Calculate the residual for each data point in the time series of this quality indicator. The calculation of the residual involves two steps: First, calculate the trendline prediction value for the data point. The trendline prediction value is equal to the mean of the indicator value calculated in step two, plus the product of the trend slope and the position deviation of the data point. That is, the trendline prediction value equals the mean of the indicator value plus the trend slope multiplied by the position deviation of the data point. Second, the residual is equal to the actual quality indicator value of the data point minus the trendline prediction value of the data point. C3: Calculate the standard deviation of the residuals for all data points of this quality indicator. The specific calculation method is as follows: multiply the residual of each data point by itself to obtain the square of the residual of that data point, add up the squares of the residuals of all data points and divide by the total number of data points N to obtain the mean of the squares of the residuals, and then take the square root of the mean of the squares of the residuals to obtain the standard deviation of the residuals; C4: The lower bound of the prediction confidence interval is equal to the predicted value minus twice the standard deviation of the residuals, and the upper bound of the prediction confidence interval is equal to the predicted value plus twice the standard deviation of the residuals. Here, twice the standard deviation is used, which, according to the statistical theory of normal distribution, roughly corresponds to a 95% confidence level. C5: Perform calculations from C1 to C4 for each quality indicator to obtain the prediction confidence intervals for each quality indicator. The confidence interval reflects the range of uncertainty in the forecast, avoiding over-reliance on a single forecast value.

[0033] The specific methods for determining predictable risks are as follows: First, obtain the access control thresholds corresponding to each quality indicator, such as code coverage access control threshold of 70%, test pass rate access control threshold of 85%, static scan defect number access control threshold of less than or equal to 5, and average cyclomatic complexity access control threshold of less than or equal to 10. The specific methods for determining predictable risks are as follows: First, obtain the access control thresholds corresponding to each quality indicator, such as code coverage access control threshold of 70%, test pass rate access control threshold of 85%, static scan defect number access control threshold of less than or equal to 5, and average cyclomatic complexity access control threshold of less than or equal to 10. R1: For a single quality indicator, when the type of the quality indicator is marked as "the higher the value, the better the quality", and the predicted value is lower than the access threshold of the quality indicator, and the lower bound of the prediction confidence interval is also lower than the access threshold of the quality indicator, the indicator is judged to have foreseeable risks. R2: When the quality indicator is labeled as "the higher the value, the better the quality", and the predicted value is higher than the access control threshold, but the lower bound of the prediction confidence interval is lower than the access control threshold, the indicator is judged to have a low risk warning. R3: When the quality indicator is marked as "the higher the value, the better the quality", and the predicted value is higher than the access control threshold, and the lower bound of the prediction confidence interval is also higher than the access control threshold, the indicator is judged to meet the prediction standard. R4: When the quality indicator is tagged as "the lower the value, the better the quality", and the predicted value is higher than the access control threshold of the quality indicator, and the upper bound of the prediction confidence interval is also higher than the access control threshold of the quality indicator, the indicator is judged to have foreseeable risks. R5: When the quality indicator is tagged as lower value means better quality, and the predicted value is lower than the access control threshold, but the upper bound of the prediction confidence interval is higher than the access control threshold, the indicator is judged to have a low risk warning. R6: When the quality indicator is categorized as "lower value, better quality", and the predicted value is lower than the access control threshold, and the upper bound of the prediction confidence interval is also lower than the access control threshold, the indicator is judged to meet the prediction standard. R7: Perform the judgments from R1 to R6 on each quality indicator to obtain the risk judgment results corresponding to each quality indicator; It should be noted that when any one of the multiple quality indicators of a module is determined to have a foreseeable risk, the module as a whole is marked as having a predictable quality risk, and corresponding access control intervention measures need to be triggered in subsequent steps.

[0034] By using a quality indicator type mapping table, it is clarified whether each indicator belongs to the category of "higher quality is better" or "lower quality is better," ensuring the correct direction of subsequent risk assessment. Then, linear extrapolation is performed based on the trend slope to calculate the predicted value after K iterations. Next, a prediction confidence interval is constructed using the historical residual standard deviation, incorporating the uncertainty of the prediction into the risk assessment. Finally, risk assessment rules are designed for the two indicator types, distinguishing between three risk levels: foreseeable risk, low-risk warning, and predicted compliance. The core innovation of this step lies in upgrading the one-dimensional judgment of whether the current value meets the standard to a predictive judgment of whether the future value may exceed the threshold, enabling the access control system to have the ability to detect risks in advance.

[0035] Step 4: Based on the combined test results of various quality indicators in the current version and the risk assessment results from Step 3, implement differentiated access control decisions; The actual test results of traditional access control are integrated with the risk assessment results of various quality indicators in step three, and a final decision is made in descending order of priority. There are four decision types, with priority from high to low: traditional rejection, trend predictive interception, trend warning approval, and traditional approval. Among them, trend predictive interception breaks the inherent logic of traditional access control that allows access as soon as the current standard is met. It enables the system to proactively intervene when the quality indicators are currently at the qualified level, but the trend clearly points to falling below the threshold, thus blocking the risk in advance.

[0036] The four decision types are executed according to priority, from highest to lowest priority: traditional rejection, trend predictive interception, trend early warning access, and traditional access. The trend predictive interception decision breaks the inherent logic of traditional access control that "allows access if the current standard is met", which is a key breakthrough in predictive quality management.

[0037] The specific determination method for the decision-making rule is as follows: First priority, traditional rejection decision: D1: Compare the measured values ​​of each quality indicator in the current version with their corresponding access control thresholds. For indicators where higher values ​​indicate better quality, if the measured value is lower than the access control threshold, the indicator is deemed to have failed the test. For indicators where lower values ​​indicate better quality, if the measured value is higher than the access control threshold, the indicator is deemed to have failed the test. D2: If any quality indicator is determined to be substandard in actual measurement, the overall access control decision will be traditional rejection, and access will be directly refused, regardless of the trend determination result in step three; Second priority, trend-predictive interception decision: D3: When D1 and D2 determine that all measured values ​​of quality indicators meet their respective access control thresholds, i.e., the traditional rejection is not triggered, the second priority judgment is entered. D4: From the risk assessment results in step three, select all quality indicators that are judged to have foreseeable risks and form a risk indicator list; D5: For each indicator in the risk indicator list, further check whether its trend type in step two is continuous deterioration. If the indicator's trend type is continuous deterioration, then the indicator meets the trend interception condition; D6: For each indicator that meets the trend interception conditions, calculate its safety margin. For indicators where higher values ​​generally indicate better quality, the safety margin equals the current measured value minus the access control threshold; for indicators where lower values ​​generally indicate better quality, the safety margin equals the access control threshold minus the current measured value. D7: Calculate the critical interception threshold, which is equal to twice the absolute value of the indicator's trend slope. If the safety margin is less than the critical interception threshold, the indicator triggers a trend-predictive interception. D8: If at least one quality indicator triggers a trend-predictive blocking, the overall access control decision will be determined as a trend-predictive blocking, and the version will be blocked from being released. The submitter is required to provide a root cause analysis of the degradation of the indicator and an improvement plan. The version can only be released after manual approval. Third priority, trend early warning through decision-making: D9: When D8 determines that the trend prediction interception condition is not met, i.e. the second priority is not triggered, the third priority judgment is entered; D10: Check whether at least one quality indicator in the risk assessment results of step three is judged to have a low-risk warning; D11: Check whether at least one quality indicator in the trend type determination results of step two is determined to be an oscillating deterioration or a recent sudden trend type; D12: If either D10 or D11 is met, the overall access control decision will be deemed as a trend warning pass, and the version will be allowed to proceed. However, the system will automatically send a trend warning notification to the project manager and module manager, requiring them to submit a rectification plan within five working days. D13: Records the current trend warning pass event. If the same quality indicator triggers a trend warning pass three times consecutively, it will automatically be upgraded to trend predictive interception when the indicator triggers a warning again on the fourth time. The fourth priority, traditionally decided upon through decision-making: D14: If none of the above first, second, and third priority decisions are triggered, it means that all quality indicators of the current version have met the standards in actual testing, and there is no trend risk or warning. In this case, the overall access control decision is to pass the traditional version and the version is released normally without any additional conditions.

[0038] By integrating the actual test results of access control with the actual risk assessment results, a four-level priority decision system is established: traditional rejection, trend predictive interception, trend warning access, and traditional access. Among them, trend predictive interception accurately judges the approaching risk by comparing the safety margin with the critical interception judgment value, and realizes proactive interception when the current actual test is qualified but the trend is approaching the threshold. Trend warning provides a flexible response of allowing access with warning for moderate risks. The automatic escalation mechanism for three consecutive warnings prevents warnings from being ignored for a long time.

[0039] Step 5: Automatically synchronize the access control decision results generated in Step 4 to the version control system via the interface, and execute the corresponding version control operations to achieve seamless linkage between quality access control and version control. Based on the four different access control decision types, perform corresponding operations such as tagging, status setting, comment addition, issue creation, and notification sending to ensure that the decision results are promptly fed back to relevant personnel and implemented in the version management process. The specific method of version control linkage operation is as follows: When the access control decision is "Traditional Pass", tag the merge request with "Quality Access Control Pass", set the merge request status to "Approved", add a brief explanation in the comments section that the quality access control test has passed, all indicators meet the standards and the trend is healthy, and allow code merging; When the access control decision is traditional rejection, tag the merge request with quality access control rejection, set the merge request status to blocked, list the name of the unmet quality indicator, the current measured value and the corresponding access control threshold in the comments section, the specific format is indicator name: measured value XX, required threshold XX, gap XX, block code merging, send email and instant message notification to the submitter, and ask them to fix it and resubmit. When the access control decision is approved for a trend warning, tag the merge request with "Quality Access Control Trend Warning" and set the merge request status to "Approved but with Warning". Add trend warning information to the comments section, including the name of the quality indicator that triggered the warning, the current measured value, the trend type, the trend slope, the predicted value, and the prediction confidence interval. Create a tracking issue and associate it with the merge request. The issue title format is "Quality Trend Warning: Module Name - Indicator Name". Allow code merging. Send a warning notification to the project leader and the module leader, requiring them to submit a rectification plan within 5 working days. Record this warning event in the warning counter. When the access control decision is a trend predictive interception, perform the following operations: Tag the merge request as "Quality Access Control Trend Interception", set the merge request status to "Blocked - Trend Risk", add detailed interception descriptions in the comments section, including the name of the quality indicator that triggered the interception, the current measured value, the access control threshold, the safety margin, the critical interception judgment value, the trend type, the trend slope, the predicted value, and the prediction confidence interval. Create a high-priority issue with the title format "[Urgent] Trend Predictive Interception: Module Name - Indicator Name", prevent code merging, and send a high-priority notification to the submitter, module owner, project manager, and quality team. The notification should clearly state that a root cause analysis report and improvement plan must be submitted and the request can only be released after manual approval. For trend-predictive blocking decisions, a manual approval exemption process is provided: After the submitter replies to the question and provides a root cause analysis report and improvement plan, the project technical manager or quality committee approver reviews it. If approved, the approver replies with approval comments in the question and manually sets a trend risk exemption flag. After the system detects the flag, it automatically updates the merge request status from blocked - trend risk to approved - exempted, adds an exemption description in the comments section including the person who exempted, the exemption time, and the exemption reason, removes the merge blocking status, and allows the submitter to manually perform the merge operation.

[0040] The access control decision results from step four are automatically synchronized to the version control system via an interface. For each of the four decision types, corresponding operations are performed, including tagging, status setting, comment addition, issue creation, and notification sending. A complete manual approval exemption process is also provided for trend-based predictive blocking. This linkage mechanism eliminates the manual transmission link between access control decisions and version control, ensuring that decision results are visible in real time and automatically implemented in the version management process, forming a closed-loop management system from quality inspection to version control.

[0041] Multiple access control historical records are constructed as time series. Trend type is determined through two-dimensional features: trend slope and oscillation coefficient. Then, by combining indicator type labeling and linear extrapolation prediction, foreseeable risks are identified. Finally, the measured conclusions and predicted risks are integrated into a four-level differentiated decision, which is automatically linked to the version control system to implement the decision results. These five steps are interconnected, upgrading the previously isolated single-point access control detection into a complete quality control closed loop with historical perception, trend prediction, and version linkage capabilities.

[0042] Example 2: Please refer to Figure 2 As shown, this embodiment also provides an integrated test version control linkage system based on quality access control. This system is implemented through the aforementioned integrated test version control linkage method based on quality access control, specifically including: The time series construction module is used to store the access control detection results in a structured manner into the historical database after each code module participates in access control detection, and to extract the time series of various quality indicators from the historical records on a module-by-module basis. The trend feature calculation and type determination module is used to calculate the trend slope and oscillation coefficient for each quality indicator based on the time series output by the time series construction module, and determine the trend type of each quality indicator according to the preset threshold. The predictive risk identification module is used to determine the type of each quality indicator, calculate the predicted value and prediction confidence interval of each quality indicator, and determine the risk level of each quality indicator based on the comparison relationship between the indicator type, predicted value, prediction confidence interval and access control threshold. The differentiated access control decision module is used to integrate the measured results of various quality indicators in the current version with the risk judgment results of the predictive risk identification module, and output one of four access control decisions according to the preset priority order: traditional rejection, trend predictive interception, trend warning pass, or traditional pass. The version control linkage execution module is used to automatically synchronize the access control decision results output by the differentiated access control decision module to the version control system through an interface, and execute the corresponding version control linkage operation for the four decision types.

[0043] The criteria for trend predictive blocking in the differentiated access control decision module are as follows: all measured values ​​of quality indicators meet the access control threshold, and at least one quality indicator simultaneously meets the three conditions of being judged as having foreseeable risks, the trend type being continuous deterioration, and the safety margin being less than the critical blocking judgment value. The critical blocking judgment value is equal to twice the absolute value of the trend slope of the indicator. If the same quality indicator triggers trend warnings three times consecutively, the fourth time will automatically be upgraded to trend predictive blocking.

[0044] The version control linkage module also provides a manual approval exemption process for trend predictive blocking decisions: after the submitter provides a root cause analysis report and improvement plan, the approver approves it and sets a trend risk exemption flag. When the version control linkage module detects the flag, it automatically updates the merge request status from blocked - trend risk to approved - exempted. It adds an exemption description in the comment section, including the person who exempted the exemption, the exemption time, and the reason for the exemption, and removes the merge blocking status, allowing the submitter to perform the merge operation.

[0045] Example 3: As Example 3 of the present invention, in specific implementation, compared with Example 1 and Example 2, the technical solution of this example is to combine the solutions of Example 1 and Example 2.

[0046] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for integrated test version control linkage based on quality access control, characterized in that, include: Step 1: After each access control test is completed by the code module, the access control test results are stored in the historical database in a structured manner. The numerical sequence of each quality indicator as the version changes is extracted from the historical records by module, and the time series of quality indicators of each module is constructed. Step 2: Based on the time series of quality indicators, calculate the trend slope and oscillation coefficient for each quality indicator. Determine the trend type of each quality indicator based on the trend slope and oscillation coefficient. The trend types include continuous deterioration, oscillating deterioration, continuous improvement, continuous stability, and recent sudden change. Step 3: Based on the trend slope of each quality indicator, first determine whether each quality indicator belongs to the type of higher quality is better or lower quality is better. Then, calculate the predicted value and prediction confidence interval of each quality indicator after K iterations. Finally, based on the type label, predicted value, prediction confidence interval and preset access threshold, determine the risk level of each quality indicator as having foreseeable risk, having low risk warning or predicting compliance. Step 4: Compare the measured values ​​of each quality indicator in the current version with the access control threshold. Combine the risk level judgment results and trend type, and determine the differentiated access control decision according to the priority from high to low. The access control decision types include traditional rejection, trend predictive interception, trend warning access, and traditional access. Step 5: Automatically synchronize the generated access control decision results to the version control system via the interface, and execute the corresponding version control linkage operation for each of the four decision types.

2. The integrated test version control linkage method based on quality access control according to claim 1, characterized in that, The specific method for constructing the time series of quality indicators for each module is as follows: The historical access control records for this module are queried from the access control history database and sorted from oldest to newest by submission time. For each quality indicator, the measured values ​​from each version are extracted. The most recent N valid records are used to construct a time series, where N is 15. If there are fewer than 15 records, all historical records are used. Valid records refer to records in which the corresponding module has indeed undergone code changes or participated in test execution in that version. If a module has fewer than 3 valid historical records, the trend determination is skipped and only the current value is checked. If a quality indicator has more than 30% missing values ​​in the time series, it is not included in the trend analysis, thus obtaining the time series of quality indicators for each module.

3. The integrated test version control linkage method based on quality access control according to claim 1, characterized in that, The specific method for calculating the trend slope and oscillation coefficient is as follows: Assign a location number to each data point in the time series, calculate the mean of the location numbers and the mean of the index values, calculate the location deviation and the value deviation for each data point, sum the product of the location deviations and value deviations of all data points to get the total A, sum the squares of the location deviations of all data points to get the total B, and the trend slope is equal to the total A divided by the total B. For each pair of adjacent data points in the quality indicator time series, calculate the difference and generate a sign sequence based on the sign of the difference. A difference greater than 0 is recorded as a positive sign, a difference less than 0 is recorded as a negative sign, and a difference equal to 0 uses the sign of the previous difference. The total number of symbol alternations in a sequence of symbols; The oscillation coefficient is equal to the total number of sign alternations divided by the number of adjacent differences.

4. The integrated test version control linkage method based on quality access control according to claim 3, characterized in that, The specific method for determining the trend type of each quality indicator is as follows: For each quality indicator, a trend judgment threshold T1 and a sudden change judgment threshold T2 are preset; when the trend slope is negative and its absolute value is greater than T1, and the oscillation coefficient is less than 0.3, it is judged as continuous deterioration; When the trend slope is negative and its absolute value is greater than T1, and the oscillation coefficient is greater than or equal to 0.3, it is judged as oscillating deterioration; when the trend slope is positive and its absolute value is greater than T1, it is judged as continuous improvement; when the absolute value of the trend slope is less than or equal to T1, it is judged as continuous stability; when the absolute value of the difference between the last two data points of the time series is greater than T2, and the absolute value of the trend slope is less than T1, it is judged as a recent mutation, and the judgment of recent mutation has the highest priority.

5. The integrated test version control linkage method based on quality access control according to claim 1, characterized in that, The specific method for determining the type of quality indicator is as follows: A pre-defined quality indicator type mapping table is used, where code coverage and test pass rate belong to the type where higher values ​​mean better quality, while static scan defect count and average cyclomatic complexity belong to the type where lower values ​​mean better quality. When an indicator to be judged falls into the corresponding list, it is assigned a corresponding mark; otherwise, it is automatically identified according to the access control threshold expression of the indicator. If the access control threshold is greater than or equal to or greater than the form, it is judged as the type where higher values ​​mean better quality; if it is less than or equal to or less than the form, it is judged as the type where lower values ​​mean better quality.

6. The integrated test version control linkage method based on quality access control according to claim 3, characterized in that, The specific method for calculating the predicted values ​​and prediction confidence intervals of each quality indicator after K iterations is as follows: The predicted value is equal to the current measured value plus the product of the trend slope and the prediction step size K. The default value of K is 3. For each data point in the quality index time series, the predicted value of its trend line is equal to the mean of the index value plus the product of the trend slope and the corresponding position deviation. The residual is equal to the actual quality index value of the data point minus the predicted value of the trend line. Calculate the standard deviation of the residuals for all data points. The lower bound of the prediction confidence interval is equal to the predicted value minus twice the standard deviation of the residuals, and the upper bound is equal to the predicted value plus twice the standard deviation of the residuals.

7. The integrated test version control linkage method based on quality access control according to claim 5, characterized in that, The specific method for determining the risk level of each quality indicator is as follows: If the quality indicator is categorized as "higher value, better quality," and the predicted value is below the access control threshold, and the lower bound of the prediction confidence interval is also below the access control threshold, then a foreseeable risk is identified. If the predicted value is above the access control threshold but the lower bound of the prediction confidence interval is below the access control threshold, then a low-risk warning is identified. If both the predicted value and the lower bound of the prediction confidence interval are above the access control threshold, then the prediction is considered satisfactory.

8. The integrated test version control linkage method based on quality access control according to claim 1, characterized in that, The specific method for determining differentiated access control decisions is as follows: First priority: If the current measured value of any quality indicator does not meet the corresponding access control threshold, then traditional rejection is executed. Second priority: If the measured values ​​of all quality indicators meet the access control threshold, and any quality indicator simultaneously meets the following conditions: the risk level judgment result indicates foreseeable risk, the trend type is continuous deterioration, and the safety margin is less than twice the absolute value of the trend slope of the indicator, then trend predictive blocking is executed. Among these, for indicators of the type where higher values ​​indicate better quality, the safety margin is equal to the current measured value minus the access control threshold; for indicators of the type where lower values ​​indicate better quality, the safety margin is equal to the access control threshold minus the current measured value. Third priority: If the second priority is not triggered and the risk judgment result of any quality indicator indicates a low-risk warning, or the trend type of any quality indicator is oscillating deterioration or recent abrupt change, then trend warning is passed, and the number of warnings is recorded. If the same quality indicator triggers trend warnings three times consecutively, then when the indicator triggers a warning again for the fourth time, it is automatically upgraded to trend predictive blocking. Fourth priority: If none of these conditions are triggered, then traditional pass is executed.

9. The integrated test version control linkage method based on quality access control according to claim 8, characterized in that, The specific methods for executing the corresponding version control linkage operations for the four decision types are as follows: When the decision is "Traditional Approval," the merge request is tagged with "Quality Gated Pass," the status is set to "Approved," and code merging is allowed. When the decision is "Traditional Rejection," the merge request is tagged with "Quality Gated Rejection," the status is set to "Blocked," the names, measured values, and gate thresholds of the non-compliant indicators are listed in the comments section, code merging is blocked, and the submitter is notified. When the decision is "Trend Warning Approval," the merge request is tagged with "Quality Gated Trend Warning Allow," the status is set to "Approved but Warning Issued," the indicator information that triggered the warning is added to the comments section, a tracking issue is created and associated with the merge request, code merging is allowed, and a warning notification is sent to the project manager and module manager, requiring them to submit a rectification plan within 5 working days. When the decision is a trend predictive interception, tag the merge request with Quality Gated Trend Interception, set the status to Blocked - Trend Risk, add detailed interception description in the comments section, create a high-priority issue, block the code merge, and send a high-priority notification to the committer, module owner, project manager, and quality team. For trend-predictive blocking decisions, a manual approval exemption process is provided. Once approved, the system automatically removes the merge block status and updates the merge request status to "Approved - Exempted".

10. An integrated test version control linkage system based on quality access control, characterized in that, This method is implemented through the integrated test version control linkage method based on quality access control as described in any one of claims 1-9, including: The time series construction module is used to structure and store the version identifier, measured values ​​of module quality indicators, and access control decision results after each access control detection is completed, and to extract the time series of each quality indicator of the module. The trend feature calculation and type determination module is used to calculate the trend slope and oscillation coefficient of the time series of each quality indicator and determine the trend type. The predictive risk identification module is used to determine the type label of quality indicators, calculate the predicted value and prediction confidence interval based on the trend slope, and determine the risk outcome. The differentiated access control decision module is used to compare the current measured value with the access control threshold, and generate access control decisions according to priority based on the risk level judgment result and trend type. The version control linkage execution module is used to automatically execute corresponding tag setting, status change, comment addition, issue creation and notification sending operations through the version control system interface according to the type of access control decision.

Images