A software development system and method based on internet information

Abstract

This invention relates to the field of application software development and discloses a software development system and method based on Internet information. One method includes parsing call identification information from API description data to generate an interface structure dependency graph; performing a layer-by-layer extended reachability analysis in the interface structure dependency graph, starting from the target interface, to generate candidate structure combinations; performing path-level consistency analysis on the candidate structure combinations based on the continuity of call paths and parameter passing consistency between interfaces to generate local structural constraint domains; continuously observing the state of structural units within the local structural constraint domains to form a local failure influence subgraph; selectively pruning the original structural paths based on the local failure influence subgraph, and performing secondary screening combined with a constraint verification mechanism to output updated local structural constraint results. This invention has the advantages of improving interface structure consistency and update efficiency.

Description

Technical Field

[0001] This invention relates to the field of application software development, specifically to a software development system and method based on Internet information. Background Technology

[0002] With the continuous development of software development models based on internet information, the development process often requires obtaining API description information from remote code repositories, online interface documentation platforms, and third-party service platforms to support interface calls and function integration. To improve development efficiency and interface reusability, existing development tools typically build local interface structure dependency mappings based on API description information and use these mappings as the basis for interface calls and composition generation. However, during the continuous evolution of interfaces, local changes such as field structure adjustments, parameter passing path changes, or interface dependency reorganization often occur, causing inconsistencies between the local interface structure dependency mapping and the actual remote interface dependency state. Existing update mechanisms typically adopt a global update method based on time periods or version identifiers, which makes it difficult to identify and update local changes in interface structure dependencies in a timely manner.

[0003] In this situation, on the one hand, local structural dependency mapping may still be used for interface calls and composition generation, causing the generated interface call relationships to remain correct during the compilation phase, but call anomalies or hidden errors may occur during the runtime phase due to changes in parameter passing paths or dependency relationships. On the other hand, due to the lack of ability to analyze the interface structural dependency relationships layer by layer and identify the scope of impact, it is impossible to determine the propagation range of local changes in structural dependencies, making it difficult to achieve differentiated processing in the update process, and still relying on the overall refresh method to complete the update, resulting in coarse update granularity, delayed response, and large resource consumption.

[0004] Therefore, it is essential to design a software development system and method based on Internet information that improves interface structure consistency and update efficiency. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a software development system and method based on Internet information, which has the advantages of improving interface structure consistency and update efficiency, and solves the problems mentioned in the background technology.

[0006] To achieve the aforementioned goals of improving interface structure consistency and update efficiency, this invention provides the following technical solution: a software development method based on Internet information, comprising the following steps: Parse the call identification information from the API description data, combine the field reference relationship and the parameter passing relationship between interfaces, merge the source dependencies of each interface in the call process, and generate an interface structure dependency graph with hierarchical association characteristics. Starting from the target interface, a layer-by-layer extended reachability analysis is performed in the interface structure dependency graph to form a set of reachable structural units. The input-output adaptation relationship between different structural units is verified by combining parameter constraint matching conditions to generate candidate structural combinations that meet the constraints. For candidate structure combinations, path-level consistency analysis is performed based on the continuity of call paths and the consistency of parameter passing between interfaces. Candidate combinations are sorted and filtered according to the consistency analysis results, and local structure constraint domains are generated. Continuous state observation is performed on structural units within the local structural constraint domain. When interface parameters change unexpectedly, return semantic offset, or response is abnormal, they are marked as failure nodes. Combined with structural dependencies, propagation analysis is performed on the affected upstream and downstream paths to form a local failure impact subgraph. Based on the local failure impact subgraph, the original structural path is selectively pruned. The reachability expansion and candidate combination generation process is re-executed on the remaining structural unit set. A second screening is performed in conjunction with the constraint verification mechanism, and the updated local structural constraint results are output.

[0007] Preferably, the process of generating an interface structure dependency graph with hierarchical association features is as follows: The API description data is parsed in a structured manner, and the path identifier, request method identifier, and version identifier are uniformly mapped into a call identifier vector; Based on the call identifier vector, the interfaces are initially classified, and the field dependency chain is constructed by combining the field reference relationship. Nested parameters, cross-interface reference fields and reusable fields are uniformly collected to form a field-level dependency subgraph. Based on the field-level dependency subgraph, the upstream and downstream relationships of the call are traced according to the parameter passing relationship between interfaces. The parameter input source and output destination are linked and expressed in a path-like manner to generate the call propagation path. Introduce source dependency aggregation rules in the call propagation path to identify and cluster interfaces that share upstream data sources or are in the same call context, forming an interface cluster with local consistency. Perform hierarchical encoding and structure mapping on the interface cluster, and uniformly abstract it into a multi-level structure representation system of interface layer, field layer and call layer, and output an interface structure dependency graph with hierarchical association characteristics.

[0008] Preferably, the process of generating an interface structure dependency graph with hierarchical association features is as follows: The API description data is parsed in a structured manner, and the path identifier, request method identifier, and version identifier are uniformly mapped into a call identifier vector; Based on the call identifier vector, the interfaces are initially classified, and the field dependency chain is constructed by combining the field reference relationship. Nested parameters, cross-interface reference fields and reusable fields are uniformly collected to form a field-level dependency subgraph. Based on the field-level dependency subgraph, the upstream and downstream relationships of the call are traced according to the parameter passing relationship between interfaces. The parameter input source and output destination are linked and expressed in a path-like manner to generate the call propagation path. Introduce source dependency aggregation rules in the call propagation path to identify and cluster interfaces that share upstream data sources or are in the same call context, forming an interface cluster with local consistency. Perform hierarchical encoding and structure mapping on the interface cluster, and uniformly abstract it into a multi-level structure representation system of interface layer, field layer and call layer, and output an interface structure dependency graph with hierarchical association characteristics.

[0009] Preferably, the process of generating candidate structure combinations that satisfy the constraints is as follows: Standardize the types of interface input and output parameters in the set of accessible structural units to unify the parameter expression format; Based on a unified parameter expression format, a parameter constraint mapping table is constructed to express field type consistency, mandatory constraints, value range, and structural nesting relationships in a regularized manner. Based on the parameter constraint mapping table, the connection relationship between structural units is matched and verified one by one to identify structural units with incompatible input and output. After removing incompatible structural units, the structural units that pass the verification are concatenated into a path to form a candidate call combination path; Based on the degree to which the parameter constraints are satisfied by the candidate combination paths, a comprehensive evaluation and ranking are performed to select the candidate structure combinations that meet the constraints, and the results are output.

[0010] Preferably, the process of performing path-level consistency analysis based on the continuity of call paths and the consistency of parameter passing between interfaces is as follows: The call paths in the candidate structure combinations are serialized to form a standardized interface call sequence; Based on the analysis of the parameter passing relationship between adjacent interfaces, the overall consistency status of the path is structurally represented, and the broken call points and abnormal jump points are marked. By combining consistency constraints, a unified measurement of the input and output deviations of parameters in the path is performed to generate a path consistency score.

[0011] Preferably, the process of generating local structural constraint domains is as follows: The path-level consistency score is weighted and fused in multiple dimensions, and a comprehensive ranking index is generated by combining the call depth weight and parameter stability weight. Based on comprehensive ranking indicators, candidate structure combinations are sorted in descending order to select highly reliable call structure combinations. Structural convergence analysis is performed on highly reliable combinations, and combinations with shared dependency paths or repetitive structural units are merged and compressed to construct local structural constraint boundaries. Based on the local structural constraint boundary, the maximum call depth and maximum dependency span of the structural units within the boundary are constrained and defined to generate a local structural constraint domain with a stable call structure boundary.

[0012] Preferably, the process of marking a node as a failure is as follows: A continuous monitoring mechanism is established for interface units within the local structural constraint domain to collect request and response data and parameter change information in real time, and to detect deviations from the current operating status based on the historical interface behavior baseline. When an unexpected addition or deletion of interface parameter structure, a shift in field semantics, or an abnormal response result is detected, the anomaly identification mechanism is triggered. The stability of anomalies is verified by combining continuous observation results within the time window, and the anomaly interface units that pass the stability verification are marked as failure nodes.

[0013] Preferably, the process of forming a local failure influence subgraph is as follows: Using the failure node as the propagation source, trace the upstream dependent interface and the downstream calling interface in the interface structure dependency graph; During the propagation process, critical path nodes affected by the failure are identified, and subgraphs of the affected nodes and associated paths are extracted to form a local failure propagation substructure. Cluster and merge the propagation substructures to output a subgraph showing the impact of local failures.

[0014] Preferably, the process of outputting the updated local structural constraint results is as follows: Failure paths in the local failure impact subgraph are removed, and path pruning is performed in the original structural dependency graph. Based on the trimmed structural map, a new reachability extension analysis is performed, starting from the target interface. The remaining structural units obtained from the expansion are re-matched with parameter constraints and verified with input-output adaptation to generate new candidate structural combinations. Perform path consistency analysis and scoring ranking again on the new candidate combinations; The consistency analysis results and scoring ranking results are fused with the updated reachable structure set and constraint screening results to output the corrected local structure constraint results.

[0015] This invention also discloses another technical solution: a software development system based on Internet information, comprising: Interface mapping module: performs call identifier parsing and multi-dimensional dependency fusion on API description data to construct an interface structure dependency graph; The reachable extension module: Starting from the target interface, it performs a layer-by-layer extension analysis on the structural dependency graph, generates a set of reachable structural units, and completes input-output constraint matching and filtering. Path filtering module: Performs path continuity and parameter consistency analysis on candidate structure combinations, sorts and filters them to form local structure constraint domains that satisfy consistency constraints; Failure perception module: continuously monitors the operating status of structural units, identifies interface anomalies or semantic offset nodes, and constructs a subgraph of the propagation of local failure effects; The adaptive reconfiguration module: Based on the failure impact subgraph, it trims and reconstructs the structural path, and re-executes the reachability analysis and constraint screening to output the updated structural results.

[0016] Compared with existing technologies, the present invention provides a software development system and method based on Internet information, which has the following beneficial effects: This invention accurately merges interface source dependencies and constructs an interface structure dependency graph with clear hierarchical relationships by parsing call identifiers, sorting out field references and interface parameter passing associations; it conducts layer-by-layer expansion reachability analysis with the target interface as the retrieval origin, delineates the set of reachable structural units, and strictly verifies the input-output adaptation relationship between units based on parameter constraint matching rules, efficiently screening compliant and usable candidate structural combinations; it completes path-level consistency verification from two dimensions: call path continuity and parameter passing consistency, and selects the best according to the analysis results, accurately delineating stable and reliable local structural constraint domains; at the same time, it implements full-cycle continuous status monitoring of all structural units within the constraint domain, which can promptly identify unexpected changes in interface parameters and semantic offsets in return. It can handle various fault scenarios such as response anomalies and mark failure nodes. Based on existing structural dependencies, it can complete the failure propagation simulation of upstream and downstream related paths and fully depict the local failure impact subgraph. According to the failure impact range, it can perform precise selective pruning of the original structural paths, iteratively carry out reachability expansion, candidate combination generation and constraint secondary verification on the remaining effective structural units, and dynamically update the output of local structural constraint results adapted to the current interface operation status. It can improve the standardization, development efficiency, structural stability, fault self-healing ability and dynamic adaptation performance of software development based on Internet interface resources, reduce the development risks and operation and maintenance costs caused by interface iteration changes, and adapt to the actual application needs of large-scale interface integration development in complex interconnected environments. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0019] Example 1: Please refer to Figure 1 A software development method based on Internet information according to an embodiment of the present invention includes the following steps: S1: Parse the call identification information from the API description data, combine the field reference relationship and the parameter passing relationship between interfaces, merge the source dependencies of each interface in the call process, and generate an interface structure dependency graph with hierarchical association characteristics.

[0020] The process of generating an interface structure dependency graph with hierarchical association features in S1 is as follows: The API description data is structured and parsed, and the path identifier, request method identifier, and version identifier are uniformly mapped into a call identifier vector. Interface definition information is obtained from external API description data sources. This information includes at least the interface path, request method type, and version identifier fields. The original description data is then structured and parsed. The path string is semantically split according to the delimiter rules. At the same time, the request method type is enumerated and encoded, and the version identifier is converted into a standard version serial number. The path identifier, request method identifier, and version identifier are uniformly mapped according to the preset vector encoding rules to form a call identifier vector that represents the unique semantic features of the interface, thereby achieving standardized alignment of APIs from different sources in a unified expression space. The interfaces are initially categorized based on call identifier vectors. Field dependency chains are constructed by combining field reference relationships. Nested parameters, cross-interface reference fields, and reused fields are uniformly collected to form a field-level dependency subgraph. Based on vector similarity or rule matching mechanisms, the interfaces are initially categorized so that interfaces with similar semantics or consistent call patterns are grouped into the same candidate set. The API request parameters and response fields are parsed field by field to extract the reference relationships between fields, including field nesting structure relationships, cross-interface field reuse relationships, and upstream and downstream field transmission relationships. For nested parameter structures, multi-level JSON or structure fields are parsed using a recursive expansion method to form a standardized set of field nodes. For cross-interface reference fields, cross-node association mapping is performed through field unique identifiers. Finally, all field nodes and their dependencies are organized into a field-level dependency subgraph. Based on the field-level dependency subgraph, the upstream and downstream relationships of the call are traced topologically according to the parameter passing relationship between interfaces. The parameter input source and output destination are linked in a chain and expressed in a path-like manner to generate the call propagation path. Based on the field-level dependency subgraph, the parameter passing relationship between interfaces is introduced to trace the upstream and downstream relationships of the interface call. Starting from the interface output field, the mapping relationship of its mapping relationship in other interface input fields is traced to build a cross-interface field passing link. Combined with the interface call chain log or contract definition information, the parameter passing path between multiple interfaces is organized sequentially, and the scattered field dependency relationship is linked into a continuous call propagation path. In the case of multi-branch propagation, the path is split and all valid propagation branches are retained to form a complete interface call propagation path, which is used to characterize the dynamic data flow relationship between interfaces. Source dependency aggregation rules are introduced into the call propagation path to identify and cluster interfaces that share upstream data sources or are in the same call context, forming interface clusters with local consistency. Source dependency aggregation rules are also introduced into the path space to perform cluster analysis on interfaces that share upstream data sources or are in the same call context. By calculating the proportion of common nodes, the coverage of shared fields, and the call context consistency index between paths, the similarity of propagation paths is evaluated. When multiple paths meet a preset threshold in the upstream dependency structure or semantic context, they are merged into the same interface cluster. Redundant paths are compressed, and key dependency paths are retained, thus forming an interface cluster structure with local consistency to express the stable composition relationship of the local call structure. Layered encoding and structural mapping are performed on interface clusters, unifying them into a multi-level structural representation system of interface layer, field layer, and call layer, outputting an interface structure dependency graph with hierarchical association characteristics. Layered encoding is performed on each interface cluster, dividing the interface system into three layers: interface layer, field layer, and call layer. The interface layer represents API entity nodes, the field layer represents internal parameters and return field structures of the interface, and the call layer represents the call propagation relationship between interfaces. Different level nodes are identified through unified encoding rules, and mapping relationships are established between layers. For example, field layer nodes belong to specific interface layer nodes, and call layer edges connect different interface layer nodes. Finally, the three-layer structure is unified and merged to form an interface structure dependency graph with hierarchical association characteristics. This graph can simultaneously express the static structure and dynamic call relationships of interfaces, providing a unified data foundation for reachability analysis and path consistency analysis.

[0021] S2: Starting from the target interface, perform a layer-by-layer extended reachability analysis in the interface structure dependency graph to form a set of reachable structural units. Combine the parameter constraint matching conditions to verify the input-output adaptation relationship between different structural units and generate candidate structural combinations that meet the constraints.

[0022] The process of forming an accessible set of structural units in S2 is as follows: Using the target interface as the initial retrieval node, the direct call dependency node is located in the interface structure dependency graph. Using the target interface as the starting retrieval object, the path identifier, request method identifier, and version identifier of the target interface are input into the interface structure dependency graph matching system. Same identifier feature localization is performed in the graph node set to extract the upstream interface node that has a direct call relationship with the target interface. All nodes that satisfy the call edge relationship are gathered to form the initial dependency node. Based on the hierarchical relationship of the interface structure dependency graph, the direct call dependency nodes are expanded layer by layer upwards to recursively obtain the upstream dependency interfaces. Starting from the initial dependency node, the upstream dependency edges in the interface structure dependency graph are recursively pushed up the hierarchy, extending the upstream interface nodes corresponding to the current layer node layer by layer upwards, and incorporating the expansion result of each layer into the expansion node set. This allows the expansion process to continuously advance to higher dependency sources in the hierarchical structure until all paths reach the termination state of having no upstream nodes or reach the preset expansion depth boundary, thereby forming a reachable path structure system covering multiple layers of dependency relationships. During the expansion process, constraints are synchronously checked on path nodes, and nodes that do not meet version compatibility or field dependency constraints are removed. At each stage of the layer-by-layer expansion, version compatibility constraint verification and field dependency structure verification are synchronously performed on the current path node to determine the compatibility matching status between the node interface version and the target calling environment, and to determine whether the node output field structure meets the input field dependency relationship of the downstream interface. When a node does not meet any constraint condition, the node is removed from the current path structure, and the subsequent expansion path starting from that node is cut off, so that the remaining path structure only retains the valid node sequence that meets both version and field constraint conditions. Within the effective path range, the reachability of structural units is comprehensively judged to form a corresponding reachability score. Within the effective path structure range, a multi-dimensional evaluation calculation is performed on each structural unit, taking the hierarchical depth of the structural unit in the dependency graph, path stability performance, and field matching coverage as evaluation input dimensions. The hierarchical depth is used to characterize the dependency distance of the structural unit relative to the target interface, the path stability is used to characterize the consistent occurrence of the structural unit in the historical call path, and the field matching coverage is used to characterize the structural matching integrity between input and output fields. The above dimensions are uniformly integrated and calculated to generate the reachability score result of the corresponding structural unit. Based on the accessibility score, structural units are screened and merged to form a set of accessible structural units centered on the target interface. After all structural units have completed the accessibility score calculation, the score results are sorted and screened. Structural units whose scores meet the set threshold range or are in the high-ranking interval are retained. Structural units with the same dependency level or highly overlapping paths are merged. Structural units with consistent semantics or highly similar dependency paths are integrated into a unified expression unit. Finally, all retained structural units and their dependency path relationships are summarized to form a set of accessible structural units centered on the target interface, which serves as the basic input structure for candidate combination generation.

[0023] The process of generating candidate structure combinations that satisfy the constraints in S2 is as follows: The input and output parameters of the interfaces in the accessible structural unit set are standardized to unify the parameter expression format. After the accessible structural unit set is input into the combination generation module, a unified parsing operation is performed on the input and output parameters contained in each interface. The parameter types, naming formats and data structure expressions in different interface definitions are normalized. For basic data types, strings, integers, floating-point numbers and boolean types are mapped to a standard type encoding system. For complex structural types, nested objects, array structures and composite types are decomposed hierarchically and converted into standard structural expression forms. For field names that have different semantics but similar functions, they are uniformly mapped according to the preset field semantic alignment rules, thereby forming standardized parameters with consistent parameter expressions. Based on a unified parameter expression format, a parameter constraint mapping table is constructed to express field type consistency, mandatory constraints, value range, and structural nesting relationships in a rule-based manner. After completing parameter standardization, a structured constraint expression system is established for each interface parameter. Field type consistency constraints, mandatory attribute constraints, value range constraints, and structural nesting constraints are modeled using unified rules. Field type consistency is used to express the data type matching relationship between input and output parameters. Mandatory attribute constraints are used to express whether a parameter must exist during the call. Value range constraints are used to limit the allowed numerical values ​​or enumeration ranges of parameters. Structural nesting constraints are used to describe the hierarchical dependency relationship between fields in complex objects. All constraint rules are uniformly organized into a parameter constraint mapping table and a mapping relationship is established with the interface parameter nodes, thereby forming a structured constraint expression system that can be used for automatic verification. Based on the parameter constraint mapping table, the connection relationships between structural units are matched and verified one by one to identify structural units with incompatible input and output. After the parameter constraint mapping table is constructed, a pairwise connection relationship matching analysis is performed on any two interface nodes in the set of reachable structural units. The output parameter set of the preceding interface is compared with the input parameter set of the following interface. The consistency of type, satisfaction of mandatory constraints, and compatibility of value range are verified item by item according to the constraint mapping table. When any constraint condition is not satisfied, the corresponding structural unit connection relationship is marked as incompatible and removed from the candidate connection relationship set, so that the remaining connection relationships all meet the structural consistency requirements of input and output parameters, thereby forming an effective structural unit connection that can be used for path splicing. After removing incompatible structural units, the structural units that pass the verification are path-based and spliced ​​to form candidate call combination paths. Based on the set of valid structural unit connections, the structural units are serialized and organized according to the calling direction relationship between interfaces. Interface nodes that meet the input-output matching relationship are chained together according to the order of calls, and multiple feasible connection branches are combined into candidate call path structures. During the path generation process, all feasible combination forms are retained for structural units with multiple path branches, so that different paths are expressed in the form of complete call chains, thereby forming candidate call combination paths that cover multiple call possibilities, and serving as the basic structural input for evaluation and screening. Based on the degree of satisfaction of parameter constraints by candidate combination paths, a comprehensive evaluation and ranking screening is performed, and candidate structure combinations that meet the constraints are output. A unified evaluation calculation process is performed on each path, and the degree of matching of parameter types involved in the path, the completeness of constraint satisfaction, and the continuity of the call chain are used as evaluation dimensions for comprehensive quantification, generating path evaluation results. Based on the evaluation results, all candidate paths are ranked and screened, and the set of paths that meet the evaluation threshold conditions is retained. This set is output as candidate structure combinations that meet the constraints.

[0024] S3: For candidate structure combinations, perform path-level consistency analysis based on the continuity of call paths and parameter passing consistency between interfaces, sort and filter candidate combinations according to the consistency analysis results, and generate local structure constraint domains.

[0025] The process of path-level consistency analysis in S3 based on the continuity of call paths and the consistency of parameter passing between interfaces is as follows: The call paths in the candidate structure combinations are serialized to form a standardized interface call sequence. The interface call relationships in each candidate combination are topologically expanded, and the call relationships are extracted into directed paths according to the execution order. For structures with branch calls, all executable path branches are generated by depth-first traversal or breadth-first traversal, and each path is linearly expanded. During the path serialization process, each interface node is converted into a standardized interface identifier. At the same time, the set of input parameter types, the set of output parameter types, and the call context information of the interface are encoded into structured node attributes to form a unified data expression format. Furthermore, by processing the sequential numbering of interface nodes in the path, the original unstructured call relationship is converted into a standardized interface call sequence with sequential semantics, thereby providing a unified input basis for consistency analysis. Based on the analysis of parameter passing relationships between adjacent interfaces using call sequence analysis, the overall consistency status of the path is structurally represented, and broken call points and abnormal jump points are marked. Adjacent interface nodes in the sequence are parsed in pairs to extract the output parameter set of the preceding interface and the input parameter set of the following interface. A parameter matching model is established based on parameter name, type, and semantic mapping relationship. For parameter pairs that can be successfully matched, they are recorded as valid passing relationships. For parameter pairs that cannot be matched or whose types are inconsistent, they are determined to be passing abnormal relationships. The parameter matching completeness and dependency satisfaction between adjacent interfaces are further calculated, and the path is structurally represented by node-level status identification. Among them, continuously matched interface segments are marked as stable call segments, interface positions with parameter breaks are marked as broken call points, and interface positions with non-sequential dependencies or jump parameter references are marked as abnormal jump points. Finally, a path structured representation containing status marking information is formed to describe the consistency distribution characteristics of the overall path. By combining consistency constraints, a unified measurement of parameter input-output deviations in the path is generated to produce a path consistency score. After obtaining the structured representation of the path, a single candidate call path is used as the analysis object. A consistency constraint system, including parameter type consistency constraints, required field constraints, value range constraints, and cross-interface dependency constraints, is introduced. Adjacent interfaces in the path are sequentially parsed, and the output parameters of the preceding interface and the input parameters of the following interface are extracted. The parameter correspondence is established based on field mapping relationships and semantic alignment rules. On this basis, consistency verification is performed on each set of corresponding parameters to identify deviation types such as inconsistent types, missing required fields, out-of-bounds value ranges, and broken dependencies. Different deviation types are assigned corresponding weights for quantification. The deviation results of all adjacent interfaces in the path are cumulatively weighted to obtain a path-level comprehensive deviation value. The influence of path length and parameter scale differences is eliminated through normalization. Finally, the normalized result is mapped to a unified path consistency score to characterize the overall consistency level of the path in the parameter input-output transmission process.

[0026] The process of generating local structural constraint domains in S3 is as follows: A multi-dimensional weighted fusion of path-level consistency scores is performed, and a comprehensive ranking index is generated by combining call depth weight and parameter stability weight. For each candidate combination, multi-dimensional evaluation factors are introduced, where call depth weight characterizes the complexity of the interface nesting hierarchy in the path, and parameter stability weight characterizes the fluctuation and consistency maintenance capability of parameters in the multi-interface transmission process. The path call depth is statistically analyzed, and a call depth weight coefficient is generated based on the maximum or average level depth of the interface nodes in the path. Simultaneously, fluctuation analysis is performed on the matching stability of parameters in the path across interfaces, and a stability weight coefficient is generated based on the parameter deviation distribution. The path-level consistency score, call depth weight, and parameter stability weight are weighted and fused to obtain a unified comprehensive ranking index, which characterizes the overall merits of candidate structure combinations in terms of consistency, structural complexity, and parameter stability. Candidate structure combinations are sorted in descending order based on comprehensive ranking indicators to select high-reliability call structure combinations. After generating comprehensive ranking indicators, all candidate structure combinations are uniformly sorted according to indicator values. The structure combinations with higher ranking indicators are prioritized in descending order. The ranking results are truncated and filtered based on preset threshold conditions. Structure combinations that rank at the top and have comprehensive indicators higher than the threshold are retained as the set of high-reliability call structure combinations. By eliminating low-ranking combinations, structural redundancy and low-consistency path interference are reduced, so that the retained structure combinations meet the expected constraints in terms of call consistency and structural stability. Structural convergence analysis is performed on high-confidence combinations. Combinations with shared dependency paths or repetitive structural units are merged and compressed to construct local structural constraint boundaries. For the selected set of high-confidence call structure combinations, the interface structural units and their dependency path information in each combination are extracted, and a structural similarity analysis model is constructed. The shared interface nodes, repetitive call paths, and public dependency chains between different combinations are compared and analyzed. When two or more structural combinations are detected to have highly overlapping structural units or shared dependency paths in the call path, structural convergence processing is performed on them. The repetitive parts are merged and compressed, and only the differentiated path branches are retained. Through this convergence process, the originally scattered structural combinations are integrated into a convergence set with a common core dependency structure. The boundary range of this convergence set is used as the local structural constraint boundary to limit the subsequent structural evolution space. Based on local structural constraint boundaries, the maximum call depth and maximum dependency span of structural units within the boundary are defined to generate a local structural constraint domain with stable call structure boundaries. Global statistical analysis is performed on all structural units within the boundary to calculate the maximum call depth and cross-interface dependency span of each structural unit in the call path. The maximum call depth represents the deepest nesting level of the structural unit in the path, and the maximum dependency span represents the maximum distance of cross-path dependency propagation. Based on the above statistical results, an upper limit for structural stability constraints is set. Structural units exceeding this upper limit are subject to boundary constraints to prevent them from participating in subsequent structural expansion. Finally, the set of structural units that satisfy both the call depth and dependency span constraints is defined as the local structural constraint domain, providing a stable constraint basis for structural optimization and path evolution.

[0027] S4: Perform continuous state observation on structural units within the local structural constraint domain. When interface parameters change unexpectedly, return semantic offsets, or response anomalies occur, they are marked as failure nodes. Combined with structural dependencies, propagation analysis is performed on the affected upstream and downstream paths to form a local failure impact subgraph.

[0028] The process of marking a node as a failure in S4 is as follows: A continuous monitoring mechanism is established for interface units within the local structural constraint domain to collect request and response data and parameter change information in real time, and to detect deviations from the current operating state based on historical interface behavior baselines. A continuous monitoring module is deployed for all interface units within the domain. This monitoring module uses interface call logs, request messages, and response results as data collection sources, and acquires interface operation data in real time according to fixed time intervals or event triggering mechanisms. It also records key indicators such as interface input parameter structure, output parameter structure, and response latency. At the same time, a behavior baseline model is constructed based on interface behavior data during historical stable operation. During real-time monitoring, the current collected data is compared and analyzed with the historical behavior baseline. The deviation calculation method is used to identify the changing trends of the interface in parameter structure and operation behavior, thereby realizing continuous deviation detection of the interface operation state. When unexpected additions or deletions to the interface parameter structure, shifts in field semantics, or abnormal response results are detected, an anomaly identification mechanism is triggered. Based on the deviation detection results, the changes in the interface parameter structure are classified and identified. Unexpected additions or deletions refer to the presence of fields in the input or output parameter set that are not present in the historical baseline or the absence of existing fields. Field semantic shifts refer to changes in the data meaning or mapping relationship of fields that have not changed in their names. Abnormal response results refer to significant inconsistencies between the interface return structure, status code, or key field values ​​and the historical stable distribution. When any type of anomaly reaches a preset trigger threshold, the anomaly identification mechanism is activated to mark the current interface unit as having an abnormal state and generate an anomaly type label and an anomaly intensity level for stability verification processing. By combining continuous observation results within a time window, anomalies are verified for stability. Anomaly interface units that pass the stability verification are marked as failure nodes. After the anomaly identification mechanism is triggered, a sliding time window is introduced to continuously observe the interface. The abnormal state of the same interface is repeatedly detected in multiple consecutive time slices, and the frequency and duration of anomalies are statistically analyzed. If the abnormal state persists in multiple consecutive time windows and does not recover to the historical behavior baseline range, the anomaly is determined to have stability characteristics, excluding occasional fluctuation interference. For interface units that pass the stability verification, they are marked as failure nodes, and the state of the node is updated in the local structural constraint domain, so that it is removed from the set of executable structural units or restricted from participating in subsequent path combinations, thereby realizing the identification and isolation of structural stability risk nodes.

[0029] The process of forming a local failure effect subgraph in S4 is as follows: Using the failed node as the propagation source, the upstream dependent interfaces and downstream calling interfaces are traced in the interface structure dependency graph. The marked failed node in the interface structure dependency graph is used as the starting propagation source. Based on the directed dependency relationship in the graph, the upstream nodes are traced in reverse to identify the interface units that provide input parameters or data dependencies for the failed node. At the same time, the downstream nodes are analyzed in a forward expansion to identify the subsequent calling interfaces that depend on the output results of the failed node. Thus, a bidirectional dependency expansion range centered on the failed node is constructed, forming the initial impact propagation range. During the propagation process, critical path nodes affected by the failure are identified, and subgraphs of the affected nodes and associated paths are extracted to form a local failure propagation substructure. Within the scope of bidirectional dependency extension, connectivity analysis is performed on all paths to screen out critical call path nodes that contain failure nodes and whose path continuity is disrupted or whose parameter transmission is interrupted. These nodes are then extracted in a structured manner based on path dependencies. Nodes that have direct or indirect dependencies on the failure nodes and their connections are retained, while redundant paths unrelated to failure propagation are removed, thus forming a local failure propagation substructure with the failure node as the core. Clustering and merging of the propagation substructures yields a local failure impact subgraph. Similarity analysis is performed on the nodes and edge relationships in the extracted local failure propagation substructures. Based on interface dependency overlap and path structure similarity, the propagation substructures are clustered, and substructures with the same upstream dependency or shared downstream propagation path are merged and compressed to eliminate duplicate propagation paths and redundant node connections. Finally, a compact local failure impact subgraph that expresses the complete failure propagation range is generated, which is used to characterize the local influence range and propagation structure characteristics of the failure node in the interface dependency system.

[0030] S5: Based on the local failure influence subgraph, the original structural path is selectively pruned. The reachability expansion and candidate combination generation process is re-executed on the remaining structural unit set. A second screening is performed in conjunction with the constraint verification mechanism, and the updated local structural constraint results are output.

[0031] The process of outputting the updated local structure constraint results in S5 is as follows: Failure paths in the local failure impact subgraph are removed, and path pruning is performed in the original structural dependency graph. Based on the generated local failure impact subgraph, a complete set of failure paths containing failure nodes and their propagation relationships is identified, and the paths in this set are marked as unusable paths. The edges and node connections corresponding to these failure paths are located in the original interface structural dependency graph, and path pruning is performed by deletion or disconnection to remove or isolate failure nodes and their propagation links from the overall dependency structure, thereby obtaining a corrected structural graph that no longer contains failure propagation effects. Based on the trimmed structural graph, a reachability extension analysis is performed again starting from the target interface. After the structural trimming is completed, the target interface is used as the starting point for analysis. In the corrected structural dependency graph, a layer-by-layer reachability extension is performed again. The interface dependencies are traversed in a depth-first or breadth-first manner to gradually obtain the set of interface units that can be reached under the current failure propagation constraints, thereby forming an updated set of reachable structural units, which is used to reflect the effective call space after failure removal. The remaining structural units obtained from the expansion are re-matched for parameter constraints and input / output adaptation to generate new candidate structural combinations. For the updated set of reachable structural units, the input and output parameters of each interface are re-parsed and standardized for comparison. Based on the consistency of parameter types, mandatory field constraints, and value range constraints, a pairwise matching and verification is performed to filter out connection relationships that do not meet the input / output adaptation conditions. Based on the valid connection relationships, the combination and reconstruction are carried out according to the interface call direction to generate new candidate structural combinations that meet the parameter constraints. The path consistency analysis and scoring ranking are performed again on the new candidate combinations; for the regenerated candidate structure combinations, the path structure is constructed according to the interface call order, and the parameter transmission consistency between adjacent interfaces in the path is detected. The path-level consistency deviation is calculated and the corresponding consistency score is generated. All candidate paths are uniformly ranked according to the consistency score, and candidate structure combinations with high consistency under the current structural constraints are selected to provide an evaluation basis for structural fusion. The consistency analysis results and scoring ranking results are fused with the updated reachable structure set and constraint screening results to output the corrected local structural constraint results. After completing the path consistency analysis and scoring ranking, the scoring results, ranking results, updated reachable structure unit set and constraint screening results are fused with multi-source information. The structural information from different sources is aligned for consistency and conflict resolution to form a unified structural evaluation and constraint expression model. Based on the fusion results, the final stable structural unit range and executable path set are determined, thereby outputting the corrected local structural constraint results, which are used to characterize the stable structural space after failure elimination and reconstruction analysis.

[0032] Example 2: Please refer to Figure 2 A software development system based on Internet information, comprising: Interface mapping module: performs call identifier parsing and multi-dimensional dependency fusion on API description data to construct an interface structure dependency graph; The reachable extension module: Starting from the target interface, it performs a layer-by-layer extension analysis on the structural dependency graph, generates a set of reachable structural units, and completes input-output constraint matching and filtering. Path filtering module: Performs path continuity and parameter consistency analysis on candidate structure combinations, sorts and filters them to form local structure constraint domains that satisfy consistency constraints; Failure perception module: continuously monitors the operating status of structural units, identifies interface anomalies or semantic offset nodes, and constructs a subgraph of the propagation of local failure effects; The adaptive reconfiguration module: Based on the failure impact subgraph, it trims and reconstructs the structural path, and re-executes the reachability analysis and constraint screening to output the updated structural results.

[0033] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0034] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An Internet information-based software development method, characterized by, Includes the following steps: Parse the call identification information from the API description data, combine the field reference relationship and the parameter passing relationship between interfaces, merge the source dependencies of each interface in the call process, and generate an interface structure dependency graph with hierarchical association characteristics. Starting from the target interface, a layer-by-layer extended reachability analysis is performed in the interface structure dependency graph to form a set of reachable structural units. The input-output adaptation relationship between different structural units is verified by combining parameter constraint matching conditions to generate candidate structural combinations that meet the constraints. For candidate structure combinations, path-level consistency analysis is performed based on the continuity of call paths and the consistency of parameter passing between interfaces. Candidate combinations are sorted and filtered according to the consistency analysis results, and local structure constraint domains are generated. Continuous state observation is performed on structural units within the local structural constraint domain. When interface parameters change unexpectedly, return semantic offset, or response is abnormal, they are marked as failure nodes. Combined with structural dependencies, propagation analysis is performed on the affected upstream and downstream paths to form a local failure impact subgraph. Based on the local failure impact subgraph, the original structural path is selectively pruned. The reachability expansion and candidate combination generation process is re-executed on the remaining structural unit set. A second screening is performed in conjunction with the constraint verification mechanism, and the updated local structural constraint results are output.

2. The software development method based on Internet information according to claim 1, characterized in that, The process of generating an interface structure dependency graph with hierarchical association features is as follows: The API description data is parsed in a structured manner, and the path identifier, request method identifier, and version identifier are uniformly mapped into a call identifier vector; Based on the call identifier vector, the interfaces are initially classified, and the field dependency chain is constructed by combining the field reference relationship. Nested parameters, cross-interface reference fields and reusable fields are uniformly collected to form a field-level dependency subgraph. Based on the field-level dependency subgraph, the upstream and downstream relationships of the call are traced according to the parameter passing relationship between interfaces. The parameter input source and output destination are linked and expressed in a path-like manner to generate the call propagation path. Introduce source dependency aggregation rules in the call propagation path to identify and cluster interfaces that share upstream data sources or are in the same call context, forming an interface cluster with local consistency. Perform hierarchical encoding and structure mapping on the interface cluster, and uniformly abstract it into a multi-level structure representation system of interface layer, field layer and call layer, and output an interface structure dependency graph with hierarchical association characteristics.

3. The software development method based on Internet information according to claim 2, characterized in that, The process of forming an accessible set of structural units is as follows: Using the target interface as the initial retrieval node, locate the directly invoked dependent node in the interface structure dependency graph; Based on the hierarchical relationship of the interface structure dependency graph, the direct call dependency nodes are expanded upwards layer by layer to recursively obtain the upstream dependency interfaces; During the expansion process, constraints are checked on the path node synchronization, and nodes that do not meet version compatibility or field dependency constraints are removed. Within the effective path range, the accessibility of structural units is comprehensively judged to form a corresponding accessibility score; Structural units are filtered and merged based on accessibility scores to form a set of accessible structural units centered on the target interface.

4. The software development method based on Internet information according to claim 3, characterized in that, The process of generating candidate structure combinations that satisfy the constraints is as follows: Standardize the types of interface input and output parameters in the set of accessible structural units to unify the parameter expression format; Based on a unified parameter expression format, a parameter constraint mapping table is constructed to express field type consistency, mandatory constraints, value range, and structural nesting relationships in a regularized manner. Based on the parameter constraint mapping table, the connection relationship between structural units is matched and verified one by one to identify structural units with incompatible input and output. After removing incompatible structural units, the structural units that pass the verification are concatenated into a path to form a candidate call combination path; Based on the degree to which the parameter constraints are satisfied by the candidate combination paths, a comprehensive evaluation and ranking are performed to select the candidate structure combinations that meet the constraints, and the results are output.

5. A software development method based on Internet information according to claim 4, characterized in that, The process of performing path-level consistency analysis based on the continuity of call paths and the consistency of parameter passing between interfaces is as follows: The call paths in the candidate structure combinations are serialized to form a standardized interface call sequence; Based on the analysis of the parameter passing relationship between adjacent interfaces, the overall consistency status of the path is structurally represented, and the broken call points and abnormal jump points are marked. By combining consistency constraints, a unified measurement of the input and output deviations of parameters in the path is performed to generate a path consistency score.

6. The software development method based on Internet information according to claim 5, characterized in that, The process of generating local structural constraint domains is as follows: The path-level consistency score is weighted and fused in multiple dimensions, and a comprehensive ranking index is generated by combining the call depth weight and parameter stability weight. Based on comprehensive ranking indicators, candidate structure combinations are sorted in descending order to select highly reliable call structure combinations. Structural convergence analysis is performed on highly reliable combinations, and combinations with shared dependency paths or repetitive structural units are merged and compressed to construct local structural constraint boundaries. Based on the local structural constraint boundary, the maximum call depth and maximum dependency span of the structural units within the boundary are constrained and defined to generate a local structural constraint domain with a stable call structure boundary.

7. A software development method based on Internet information according to claim 6, characterized in that, The process of marking a node as a failure is as follows: A continuous monitoring mechanism is established for interface units within the local structural constraint domain to collect request and response data and parameter change information in real time, and to detect deviations from the current operating status based on the historical interface behavior baseline. When an unexpected addition or deletion of interface parameter structure, a shift in field semantics, or an abnormal response result is detected, the anomaly identification mechanism is triggered. The stability of anomalies is verified by combining continuous observation results within the time window, and the anomaly interface units that pass the stability verification are marked as failure nodes.

8. A software development method based on Internet information according to claim 7, characterized in that, The process of forming a local failure effect subgraph is as follows: Using the failure node as the propagation source, trace the upstream dependent interface and the downstream calling interface in the interface structure dependency graph; During the propagation process, critical path nodes affected by the failure are identified, and subgraphs of the affected nodes and associated paths are extracted to form a local failure propagation substructure. Cluster and merge the propagation substructures to output a subgraph showing the impact of local failures.

9. A software development method based on Internet information according to claim 8, characterized in that, The process of outputting the updated local structural constraint results is as follows: Failure paths in the local failure impact subgraph are removed, and path pruning is performed in the original structural dependency graph. Based on the trimmed structural map, a new reachability extension analysis is performed, starting from the target interface. The remaining structural units obtained from the expansion are re-matched with parameter constraints and verified with input-output adaptation to generate new candidate structural combinations. Perform path consistency analysis and scoring ranking again on the new candidate combinations; The consistency analysis results and scoring ranking results are fused with the updated reachable structure set and constraint screening results to output the corrected local structure constraint results.

10. A software development system based on Internet information, applied to the software development method based on Internet information as described in any one of claims 1-9, characterized in that, include: Interface mapping module: performs call identifier parsing and multi-dimensional dependency fusion on API description data to construct an interface structure dependency graph; The reachable extension module: Starting from the target interface, it performs a layer-by-layer extension analysis on the structural dependency graph, generates a set of reachable structural units, and completes input-output constraint matching and filtering. Path filtering module: Performs path continuity and parameter consistency analysis on candidate structure combinations, sorts and filters them to form local structure constraint domains that satisfy consistency constraints; Failure perception module: continuously monitors the operating status of structural units, identifies interface anomalies or semantic offset nodes, and constructs a subgraph of the propagation of local failure effects; The adaptive reconfiguration module: Based on the failure impact subgraph, it trims and reconstructs the structural path, and re-executes the reachability analysis and constraint screening to output the updated structural results.

Images