A dynamic mapping and project resource scheduling method and system based on a meta-protocol

Abstract

The application discloses a kind of based on meta-protocol dynamic mapping and project resource scheduling method and system, the method is by defining unified meta-protocol specification, the core of configuration and verification algorithm and extended information, generate standard meta-protocol file.Based on this, automatically match or generate algorithm adaptation plug-in, and detect repair running dependence;With data interaction rule and historical fault record, generate and test intelligent call agent, complete the access of algorithm.In the stage of data interaction, the trained parameter prediction model is used to dynamically optimize the conversion parameter, and the precise data mapping and iterative optimization are executed.In the resource scheduling stage, the running environment demand, resource consumption trend and real-time state are comprehensively considered, the resource demand peak is actively predicted by time series prediction model, and dynamic resource allocation and isolation are executed based on the prediction result.The application realizes algorithm access automation, data interaction intelligence and resource scheduling prediction, significantly improves the integration efficiency, data accuracy and system stability.

Description

Technical Field

[0001] This invention belongs to the field of computer algorithm integration technology, specifically relating to a dynamic mapping and project resource scheduling method and system based on meta-protocols. Background Technology

[0002] In the process of deepening business operations within enterprise-level project management platforms, algorithm integration serves as a crucial link in supporting business data processing, improving business efficiency, empowering intelligent decision-making, and facilitating business expansion. It is essential to achieve a collaborative closed loop between integrated algorithms and existing business modules. However, current algorithm integration faces challenges such as poor compatibility, chaotic data interaction formats, lack of unified integration rules, and a deficiency in dynamic adaptation and scheduling.

[0003] (1) Poor algorithm integration compatibility and high adaptation costs

[0004] Due to differences in technology stacks, the algorithms to be integrated require the development of dedicated hard-coded interfaces for business modules. When adding new algorithms, existing interfaces also need to be reconstructed. This approach not only consumes a lot of development resources but also prolongs the platform iteration cycle, severely restricting the efficiency of algorithm deployment.

[0005] (2) Data interaction format is chaotic and collaboration efficiency is low.

[0006] The lack of unified data interaction rules between algorithms and business modules makes data transmission prone to parsing errors; in high-frequency interaction scenarios, redundant data format conversion can easily cause data response delays, affecting real-time business decisions.

[0007] (3) There are no unified rules for integration and the standardization is insufficient.

[0008] There is a lack of unified standards in algorithm integration and environment configuration, requiring manual verification of algorithm dependencies and manual configuration of the runtime environment; there is no version management mechanism for algorithm iteration, which can easily lead to errors in business processing results due to calling old versions of algorithms.

[0009] (4) Lack of dynamic adaptation and scheduling, resulting in slow iterative response.

[0010] Business iterations require synchronous algorithm updates. Existing solutions require manual modification of the mapping relationship between the algorithm and business modules, as well as the input validation code. This not only easily leads to business failures but also fails to achieve one-time adaptation and multiple reuses, resulting in low business iteration response efficiency.

[0011] Existing containerized deployments only solve the problem of environment isolation. Remote interface calls are not required for local scenarios, and static protocol integration lacks dynamic adaptation. Therefore, they cannot meet the integrated requirements of adaptive access of multiple types of algorithms, standardized cross-module interaction, and dynamic adaptation to business iteration.

[0012] Therefore, providing a dynamic mapping and project resource scheduling method and system based on meta-protocols to solve the above problems is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0013] To address the aforementioned technical problems, the present invention aims to provide a dynamic mapping and project resource scheduling method based on meta-protocols. This method effectively solves problems such as poor access compatibility, chaotic data interaction, insufficient integration standardization, and lack of dynamic adaptation and scheduling in the existing multi-language algorithm integration process. It enables zero-preset adaptive access of multiple types of algorithms in the platform, intelligent enhanced data interaction, predictive resource scheduling, and self-repair of version conflicts, thereby improving the intelligence and efficiency of algorithm integration.

[0014] The technical solution provided by this invention is as follows:

[0015] A dynamic mapping and project resource scheduling method based on meta-protocols includes the following steps:

[0016] According to the preset meta-protocol specification, configure the target algorithm to generate the initial meta-protocol configuration file and verify it; after the verification is successful, generate the standard meta-protocol file.

[0017] The standard meta-protocol file is parsed to extract its core and extended information, which is then stored in the meta-information database.

[0018] Based on the algorithm language type in the core information, match or generate the corresponding algorithm adaptation plugin, and based on the runtime environment requirements in the core information, automatically detect and repair runtime environment dependencies.

[0019] Based on the data interaction rules in the core information and the historical fault records in the extended information, a call proxy corresponding to the algorithm adaptation plugin is generated, and the call proxy is tested; after the test is passed, the call proxy is registered to the proxy management library to complete the access of the target algorithm;

[0020] In response to a data interaction request for the target algorithm that has been accessed, the corresponding data interaction rules are obtained from the metadata database. Based on the data interaction rules, the trained parameter prediction model, and the data features of the current interaction data, data transformation and dynamic mapping are performed, and mapping deviations are recorded to iteratively optimize subsequent mappings.

[0021] In response to a resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data conversion and dynamic mapping process, a preset time series prediction model is used to predict resource requirements, and dynamic resource scheduling is performed based on the prediction results and the real-time resource status.

[0022] Preferably, the core information includes at least: the unique identifier of the target algorithm, the algorithm language type, the data interaction rules, the runtime environment requirements, and the version number;

[0023] The extended information includes at least: algorithm interaction frequency, historical fault records, and resource consumption trend data.

[0024] Preferably, the step of matching or generating corresponding algorithm adaptation plugins based on the algorithm language type in the core information, and automatically detecting and repairing runtime environment dependencies based on the runtime environment requirements in the core information, includes the following steps:

[0025] Based on the algorithm language type, a matching algorithm adaptation plugin is retrieved from a preset adaptation plugin library;

[0026] If the search is successful, the matching algorithm adaptation plugin is loaded; if the search fails, the target algorithm's interface is parsed through a syntax tree based on the algorithm's language type and the data interaction rules, a preset language template is matched, and data verification logic defined by the data interaction rules is injected to generate a new algorithm adaptation plugin.

[0027] After loading or generating the algorithm adaptation plugin, it checks whether the current runtime environment meets the runtime environment requirements and performs automatic repair for missing or mismatched dependencies.

[0028] Preferably, the step of generating a call proxy corresponding to the algorithm adaptation plugin based on the data interaction rules in the core information and the historical fault records in the extended information, and testing the call proxy; after passing the test, registering the call proxy to the proxy management library to complete the access of the target algorithm includes the following steps:

[0029] Invoke the proxy template corresponding to the algorithm language type, inject the calling rules and data verification logic defined by the data interaction rules into it, and inject fault tolerance logic based on the fault type indicated by the historical fault record to generate executable calling proxy code;

[0030] Based on the aforementioned call proxy code, parameter interception, data compliance, and connectivity tests are performed.

[0031] If the test fails, the call proxy will be regenerated; if the test passes, the call proxy will be registered in the proxy management library, thus completing the access to the target algorithm.

[0032] Preferably, the step of responding to a data interaction request for the accessed target algorithm, obtaining the corresponding data interaction rules from the metadata database, and performing data transformation and dynamic mapping based on the data interaction rules, the trained parameter prediction model, and the data features of the current interaction data, while recording the mapping deviation to iteratively optimize subsequent mappings, includes the following steps:

[0033] In response to a data interaction request, extract the data features of the current interaction data;

[0034] The data features are input into the trained parameter prediction model, which outputs the optimal data transformation parameters.

[0035] According to the data interaction rules, the corresponding mapping rules are matched from the dynamic mapping library, and based on the mapping rules and the optimal data conversion parameters, the current interaction data is converted in type and adapted in format to generate adapted data that conforms to the data interaction rules.

[0036] Calculate the mapping deviation rate of this data transformation, and iteratively update the mapping rules in the dynamic mapping library based on the mapping deviation rate.

[0037] Preferably, the trained parameter prediction model is obtained through the following steps:

[0038] With the goal of minimizing the transformation bias, the linear regression model fitted by the least squares method is obtained by iteratively training using historical mapping data and gradient descent.

[0039] The data features include at least data type encoding and interaction scenario encoding.

[0040] Preferably, the step of responding to a resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data transformation and dynamic mapping process, using a preset time-series prediction model to predict resource requirements, and performing dynamic resource scheduling based on the prediction results and real-time resource status, includes the following steps:

[0041] In response to a resource scheduling request, resource scheduling decision data is generated based on the operating environment requirements and the historical consumption patterns represented by the resource consumption trend data, combined with the real-time resource status and the real-time resource requirements of the data conversion process.

[0042] The preset time-series prediction model is invoked to predict the peak resource demand for future periods based on historical consumption data in the resource scheduling decision data.

[0043] Based on the peak resource demand and the real-time resource status, a resource allocation scheme is generated;

[0044] The target algorithm and data transformation process are allocated and isolated using a resource isolation mechanism, with dedicated computing resources and operating environment.

[0045] Preferably, the preset time series prediction model adopts the moving average method, which calculates the predicted value of resource demand for the same period in the future based on the actual resource consumption value of the same period in the past N cycles, where N is a positive integer.

[0046] Preferably, the resource demand prediction formula of the preset time series prediction model is:

[0047] ;

[0048] in, This represents the projected resource demand for future periods. For the past The same period of each cycle The actual value of resource consumption, The number of periods used to calculate the average value is a positive integer.

[0049] A dynamic mapping and project resource scheduling system based on meta-protocols, comprising:

[0050] The verification and protocol generation module is used to configure the target algorithm according to the preset meta-protocol specification to generate the initial meta-protocol configuration file, and to verify it; after the verification is successful, a standard meta-protocol file is generated.

[0051] The information extraction and storage module is used to parse the standard meta-protocol file, extract the core information and extended information therein, and store them in the meta-information database;

[0052] The plugin adaptation and dependency management module is used to match or generate corresponding algorithm adaptation plugins based on the algorithm language type in the core information, and to automatically detect and repair runtime environment dependencies based on the runtime environment requirements in the core information.

[0053] The proxy generation and access module is used to generate a call proxy corresponding to the algorithm adaptation plugin based on the data interaction rules in the core information and the historical fault records in the extended information, and to test the call proxy; after the test is passed, the call proxy is registered to the proxy management library to complete the access of the target algorithm;

[0054] The data mapping and optimization module is used to respond to the data interaction request for the connected target algorithm, obtain the corresponding data interaction rules from the metadata database, and perform data transformation and dynamic mapping based on the data interaction rules, the trained parameter prediction model and the data characteristics of the current interaction data, while recording the mapping deviation to iteratively optimize the subsequent mapping.

[0055] The resource prediction and scheduling module is used to respond to resource scheduling requests for the target algorithm. Based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data conversion and dynamic mapping process, it uses a preset time series prediction model to predict resource requirements and performs dynamic resource scheduling based on the prediction results and real-time resource status.

[0056] This invention discloses a dynamic mapping and project resource scheduling method based on meta-protocols. The method involves configuring the core and extended information of the target algorithm according to a preset meta-protocol specification, performing verification, and generating a standard meta-protocol file upon successful verification. This standard meta-protocol file is then parsed, and the core and extended information are extracted and stored in a meta-information database. Based on the algorithm language type in the core information, a corresponding algorithm adaptation plugin is matched or generated. Simultaneously, based on runtime environment requirements, runtime environment dependencies are automatically detected and repaired. Based on the data interaction rules in the core information and historical fault records in the extended information, a call proxy corresponding to the algorithm adaptation plugin is generated. After the call proxy passes testing, it is registered in a proxy management database, thus completing the access of the target algorithm. In response to data interaction requests, data transformation and dynamic mapping are performed based on data interaction rules, a trained parameter prediction model, and the data characteristics of the current interaction data. In response to resource scheduling requests, resource demand is predicted using a preset time-series prediction model, based on runtime environment requirements and resource consumption trend data, combined with the real-time resource status of the current node and the real-time resource demand during data transformation and dynamic mapping. Dynamic resource scheduling is then performed in conjunction with the real-time resource status.

[0057] The present invention has the following beneficial effects:

[0058] 1. Through the meta-protocol driven adaptive access mechanism, combined with the adaptation plugin library and the automatic generation function of the call proxy, the intelligent plugin generation unit can automatically adapt algorithms without pre-built plugins without manual intervention, and complete the parsing, adaptation and registration of multi-language algorithms. This effectively reduces the technical threshold of algorithm integration and significantly improves the access efficiency and compatibility of cross-language algorithms.

[0059] 2. By defining unified data interaction rules through meta-protocols, and by training prediction models with historical data and dynamically adjusting mapping rules, data conversion and interaction processing between algorithms and business modules can be completed automatically, effectively reducing data format adaptation deviations and significantly improving internal data flow efficiency.

[0060] 3. By formulating comprehensive integration specifications, we can achieve full lifecycle management of meta-protocols, pre-allocate resources based on time-series prediction models, reduce the failure rate caused by insufficient resources, and improve the stability of algorithm operation; by automatically resolving cross-layer version conflicts through priority matching algorithms, we can shorten version update time, avoid business interruptions caused by conflicts, effectively improve the maintainability of algorithm integration and platform scalability, and improve resource utilization and business iteration response rate.

[0061] This invention also provides a dynamic mapping and project resource scheduling system based on meta-protocols. Since it shares the same technical concept and solves the same technical problem as this method, it should have the same beneficial effects, and will not be described in detail here. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 This is a flowchart of a dynamic mapping and project resource scheduling method based on meta-protocol provided in an embodiment of the present invention;

[0064] Figure 2 This is a flowchart of step S3 provided in an embodiment of the present invention;

[0065] Figure 3 This is a flowchart of step S4 provided in an embodiment of the present invention;

[0066] Figure 4 This is a flowchart of step S5 provided in an embodiment of the present invention;

[0067] Figure 5 This is a flowchart of step S6 provided in an embodiment of the present invention;

[0068] Figure 6 This is a schematic diagram of a dynamic mapping and project resource scheduling system based on meta-protocol provided in an embodiment of the present invention. Detailed Implementation

[0069] 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.

[0070] The embodiments of this invention are written in a progressive manner.

[0071] This invention provides a method and system for dynamic mapping and resource scheduling of meta-protocols. It primarily addresses the technical problems of high cost of multilingual algorithm integration, inconsistent data interaction formats, lack of dynamic resource scheduling, and unsystematic meta-protocol management. This invention achieves non-intrusive and rapid integration of multilingual algorithms by providing standardized meta-protocol definition and verification, multilingual zero-preset adaptive automated integration, dynamic data feature extraction and intelligent mapping, full lifecycle management of meta-information, and predictive resource scheduling. It offers a systematic solution to problems such as repetitive development, poor compatibility, and complex operation and maintenance in traditional algorithm integration, improving algorithm integration efficiency and resource utilization, optimizing the efficiency and accuracy of cross-module data flow, and reducing algorithm operation and maintenance costs and data interaction error rates.

[0072] For ease of description, the following terms in this document have the following meanings:

[0073] Dynamic mapping library: refers to a database or rule base that stores data mapping rules, used to store and manage data transformation rules.

[0074] Mapping rules: These are specific rule entries that describe how data fields are converted from source format to target format.

[0075] Mapping rule table: refers to the tabular storage format of rules in the dynamic mapping library.

[0076] like Figure 1 As shown, a dynamic mapping and project resource scheduling method based on meta-protocols includes the following steps:

[0077] S1. Configure the target algorithm according to the preset meta-protocol specification to generate the initial meta-protocol configuration file, and verify it; after the verification is passed, generate the standard meta-protocol file;

[0078] S2. Parse the standard meta-protocol file, extract its core and extended information, and store it in the meta-information database;

[0079] S3. Based on the algorithm language type in the core information, match or generate the corresponding algorithm adaptation plugin, and based on the runtime environment requirements in the core information, automatically detect and repair runtime environment dependencies;

[0080] S4. Based on the data interaction rules in the core information and the historical fault records in the extended information, generate a call proxy corresponding to the algorithm adaptation plugin, and test the call proxy; after the test is passed, register the call proxy to the proxy management library to complete the access of the target algorithm;

[0081] S5. In response to the data interaction request for the target algorithm that has been connected, obtain the corresponding data interaction rules from the metadata database, and perform data transformation and dynamic mapping based on the data interaction rules, the trained parameter prediction model and the data characteristics of the current interaction data, while recording the mapping deviation to iteratively optimize the subsequent mapping;

[0082] S6. In response to the resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data transformation and dynamic mapping process, the resource requirements are predicted using a preset time series prediction model, and dynamic resource scheduling is performed based on the prediction results and the real-time resource status.

[0083] Steps S1 to S6 detail the implementation of the dynamic mapping and project resource scheduling method based on the meta-protocol. Step S1 standardizes the description language of all algorithms using the meta-protocol specification. Step S3 utilizes the matching and generation mechanism of the algorithm adaptation plugin and the invocation proxy in Step S4, abandoning the traditional approach of developing coding interfaces for each algorithm. Plugins and proxies are automatically configured, lowering the barrier and cost of algorithm integration. Step S5 performs data conversion and dynamic mapping based on defined data interaction rules, solving the problem of low efficiency due to parsing errors caused by inconsistent data formats and improving the reliability of algorithm integration. Step S6 uses resource consumption trend data and time-series prediction models to transform the passive response mode into an active prediction mode, improving resource allocation utilization. With the meta-information database in Step S2 as the core hub, the processes of algorithm access, data conversion, dynamic mapping, and resource scheduling are completed, achieving a closed-loop management system for algorithm integration.

[0084] In one embodiment, the implementation process of the dynamic mapping and project resource scheduling method based on meta-protocols mainly includes the following functional stages:

[0085] The verification and protocol generation phase is implemented through a meta-protocol definition layer. This layer is responsible for the standardized definition and standard generation of meta-protocols, and includes components such as a meta-protocol configuration unit, a validity verification tool, and a standard file generator. The meta-protocol configuration unit collects core information about the target algorithm, such as unique identifiers, language types, data rules, environmental requirements, version numbers, and extended information, to form initial configuration data. Extended information includes interaction frequency, historical fault records, and resource consumption trends. The validity verification tool performs mandatory field integrity and field validity checks on the initial configuration; if any fail, feedback is provided to the user for correction. After successful verification, the standard file generator encapsulates the valid information into standard format files such as JSON, providing a unified intelligent decision-making basis for subsequent multi-language adaptation and resource prediction scheduling.

[0086] In the information extraction and storage stage, the meta-protocol parsing unit of the multi-language adaptive access layer and the preset meta-information database are used. The meta-protocol parsing unit uses a basic JSON parsing tool to read standard files, extract core information such as unique identifier, development language, data rules, environment requirements, and version number, ensures compliance through simple format verification, performs meta-protocol parsing, and stores the parsed information directly into the meta-information database through the storage unit.

[0087] In the plugin adaptation and dependency management phase, the system utilizes an intelligent plugin generation unit and a plugin iteration unit within the multi-language adaptive access layer, along with an adaptation plugin library. The intelligent plugin generation unit retrieves corresponding plugins from the pre-built multi-language interaction logic package's adaptation plugin library, automatically matching and loading the appropriate plugin. If a matching plugin is found, it is automatically loaded into the runtime environment, providing language-specific interactive processing capabilities. If no plugin is found, the intelligent plugin generation unit is triggered, calling a syntax tree analysis tool to parse the algorithm interface syntax, extract input parameter types and return value formats, match the pre-built language template mapping library, and generate an adaptation plugin. After loading into the runtime environment, logs are recorded regarding plugin generation time and initialization status, resolving incompatibility issues in interaction logic during multi-language algorithm access. After plugin loading, the plugin iteration unit automatically scans the current server environment, detects runtime environment dependencies, compares the environment requirements in the meta-protocol, identifies whether hardware resources such as CPU, memory, and storage are insufficient, and whether software dependencies are missing or mismatched. It generates a structured dependency problem list with dependency types, current status, requirement standards, and remediation solutions, providing precise data for automated remediation.

[0088] In the proxy generation and access phase, the system is implemented through a multi-language adaptive access layer, comprising a proxy generation unit, a proxy testing unit, an execution proxy registration unit, a service activation unit, and a proxy management library. The proxy generation unit retrieves language-specific proxy templates, injects meta-protocol data interaction rules (such as parameter formats and validation standards), and invocation rules (such as timeout settings and retry mechanisms). Finally, it incorporates targeted fault-tolerance logic based on historical fault records in the meta-protocol extended information. These steps generate executable proxy code for the adapted method interface. This proxy acts as middleware between the algorithm and business modules, resolving inconsistencies in interface calls. The proxy testing unit performs multi-dimensional tests, including parameter interception, data compliance, and connectivity. Parameter interception tests verify the proxy's ability to capture business parameters and convert algorithm-readable formats. Data compliance tests check if the algorithm output conforms to meta-protocol rules. Connectivity tests verify communication stability through multiple loop requests. Test logs are synchronized to a visualization interface in real time, providing a basis for troubleshooting and ensuring the reliability of the proxy function. The test result processing branch automatically analyzes the reasons for failure and triggers proxy regeneration if the proxy fails the test. If the test passes, it proceeds to the proxy registration stage, achieving closed-loop control of proxy quality. The proxy registration and service activation units perform proxy registration and service activation operations, registering the core information of proxies that pass the test to the proxy management library, such as interface address, adapted language, and meta-protocol version. An index association is established in the meta-information library for algorithm identifiers, proxy information, and extended information, and these are included in the callable resource pool. The visual interface updates the algorithm access status to successful, indicating that the algorithm is now capable of being called by business modules, completing the closed-loop access process.

[0089] The data transformation and dynamic mapping phase is achieved through a standardized data interaction layer. This layer includes a data feature extraction unit, a dynamic mapping library, a data transformation engine, an intelligent mapping optimization unit, and a mapping rule iteration library. It optimizes data type transformation logic based on the data interaction rules in the meta-protocol and historical data interaction records; it automatically records mapping deviation data and iteratively updates the rules in the dynamic mapping library, achieving self-optimization of dynamic data mapping and reducing data interaction error rates and latency. Specifically, dynamic data mapping uses the data feature extraction unit to parse the data stream, extract feature tags such as data type, dimension, and interaction scenario, match them with the transformation rules defined in the meta-protocol, and simultaneously call the intelligent mapping optimization unit to predict the optimal transformation parameters based on historical mapping data, performing format adaptation to resolve cross-module data interaction format deviation issues.

[0090] In the resource prediction and scheduling phase, this is achieved through a meta-information and scheduling layer. This layer includes a resource prediction and scheduling unit and a version conflict resolution unit, which manage the entire lifecycle of the meta-protocol, implement meta-protocol version control, and pre-allocate computing resources based on resource requirements and historical resource consumption data in the meta-protocol using a time-series prediction model. During version updates, cross-layer dependency conflicts are automatically detected and resolved using a priority matching algorithm, completing resource matching and scheduling. Specifically, dynamic resource scheduling involves the resource scheduling unit allocating initial resources based on algorithm environment requirements and real-time server status, while simultaneously invoking the resource prediction and scheduling unit to predict future peak resource demand based on resource consumption trends in the meta-protocol extended information, reserving redundant resources in advance. During meta-protocol updates, the version management unit automatically synchronizes to the associated layer, and the version conflict resolution unit detects cross-layer dependency conflicts. If the data dynamic mapping and resource scheduling process runs stably, the algorithm becomes a standardized service; if manual intervention is triggered, the process terminates and can be restarted after processing.

[0091] Preferably, the core information includes at least: the unique identifier of the target algorithm, the algorithm language type, the data interaction rules, the runtime environment requirements, and the version number;

[0092] Extended information should include at least: algorithm interaction frequency, historical fault records, and resource consumption trend data.

[0093] In practical applications, a unique identifier is assigned to each algorithm globally to accurately distinguish different algorithms and ensure uniqueness in subsequent agent registration, resource scheduling, and other processes. The language type specifies the development and execution language of the algorithm, determining the matching logic for compatible plugins and agent generation. Data interaction rules strictly define the data format, type, and precision of the algorithm's input / output to ensure consistency in cross-module data transmission. The runtime environment requirements specify the hardware resources and software dependencies required for the algorithm's operation. The version number uniquely identifies different versions of the meta-protocol, providing fundamental support for version iteration and update control of the meta-protocol. Extended information includes daily average interaction counts, historical fault records, and resource consumption trends, providing data for subsequent intelligent plugin generation, agent optimization, and resource prediction and scheduling.

[0094] like Figure 2 As shown, preferably, based on the algorithm language type in the core information, a corresponding algorithm adaptation plugin is matched or generated, and based on the runtime environment requirements in the core information, runtime environment dependencies are automatically detected and repaired, including the following steps:

[0095] A1. Based on the algorithm language type, retrieve matching algorithm adaptation plugins from the preset adaptation plugin library;

[0096] A21. If the search is successful, load the matching algorithm adaptation plugin;

[0097] A22. If the retrieval fails, based on the algorithm language type and data interaction rules, the interface of the target algorithm is parsed through the syntax tree, matched with the preset language template, and the data verification logic defined by the data interaction rules is injected to generate a new algorithm adaptation plugin.

[0098] A3. After loading or generating the algorithm adaptation plugin, check whether the current runtime environment meets the runtime environment requirements, and perform automatic repair for missing or mismatched dependencies.

[0099] Steps A1 to A3 are the specific implementation details of step S3. Steps A1 to A22 use the algorithm language type as the basis for judgment, and retrieve matching algorithm adaptation plugins from the adaptation plugin library. If the platform already supports mainstream languages ​​such as Python and Java, the corresponding algorithm adaptation plugin is directly connected. If it is a less common new language or a custom language, the structure of the algorithm interface (such as function name, parameter type, etc.) is parsed through the syntax tree, then matched with the preset language template, and data verification logic based on the data interaction rules is injected to create a new adaptation plugin. Since both loading existing plugins and generating new plugins are based on the data interaction rules in the meta-protocol, it means that the verification behavior of matching or generating plugins on the algorithm interface is consistent. Step A3 actively detects the current server's running environment after the plugin is ready and compares it with the running environment requirements in the meta-protocol. For missing or mismatched dependencies (such as Python environment libraries, Java environment libraries), it attempts to perform automated repair (such as automatically downloading the environment libraries corresponding to the language).

[0100] In one embodiment, step A22 is implemented as follows:

[0101] If the retrieval fails, a new algorithm adaptation plugin will be generated according to the following process:

[0102] ① Call the syntax tree parsing tool corresponding to the target algorithm language;

[0103] ② Parse the abstract syntax tree of the target algorithm interface and extract core interface information such as input parameter types, return value formats, and function name specifications;

[0104] ③ Match the preset language template;

[0105] ④ Inject the data verification logic defined in the data interaction rules of the meta-protocol to generate an algorithm adaptation plugin that can be directly called;

[0106] In practical application, the target algorithm is the project cost accounting interface of an enterprise-level project management platform developed in a niche language (without pre-built plugins). The input parameters are project_id:str, budget_data:dict, actual_expense:list, and the return value is cost_deviation:float. After extracting the core interface information through syntax tree parsing, it matches the Python language template and injects the verification logic with project_id length ≥ 8 characters and cost_deviation range [-50%, 50%] based on the meta-protocol data interaction rules. This generates an adapter plugin that can seamlessly interact with the platform's Python business modules.

[0107] The above solution transforms the complex, specialized, and error-prone task of localizing algorithms into a highly automated, standardized, and reliable backend process. Business developers or algorithm providers only need to focus on how to correctly describe the algorithm using meta-protocols, without needing to delve into the technical details and environment configurations for cross-language calls. This greatly liberates productivity and accelerates the value transfer process of algorithms from development to deployment.

[0108] like Figure 3 As shown, preferably, based on the data interaction rules in the core information and the historical fault records in the extended information, a call proxy corresponding to the algorithm adaptation plugin is generated, and the call proxy is tested; after the test is passed, the call proxy is registered to the proxy management library to complete the access of the target algorithm, including the following steps:

[0109] B1. Call the proxy template corresponding to the algorithm language type, inject the calling rules and data verification logic defined by the data interaction rules into it, and inject fault tolerance logic based on the fault type indicated by the historical fault records to generate executable calling proxy code;

[0110] B2. Based on the call proxy code, perform parameter interception, data compliance, and connectivity testing;

[0111] B31. If the test fails, the proxy will be regenerated.

[0112] B32. If the test passes, the proxy will be registered with the proxy management library to complete the access of the target algorithm.

[0113] Steps B1 to B3 are the specific implementation details of step S4. Step B1 calls the proxy corresponding to the language type, injecting the calling rules and data verification logic defined in the core information of the meta-protocol. Fault tolerance logic is injected based on the fault types indicated by historical fault records in the extended information, ultimately generating the calling proxy code. This calling proxy code carries existing operational experience and past failure experience, exhibiting high maturity and robustness. Step B2, after generating the calling proxy code, performs testing: parameter interception testing verifies whether the proxy can correctly intercept business requests and preprocess them according to rules (such as packaging and conversion); data compliance testing verifies whether the proxy can ensure that the data input to the algorithm and the data returned by the algorithm strictly conform to the format and constraints defined by the meta-protocol. This is the key gatekeeper for data consistency; connectivity testing verifies whether the proxy can establish a stable and effective communication link with the backend algorithm instance. The embedded testing process prevents defective proxies from entering the production environment, ensuring the quality of service interfaces from the source. Redundancy measures are set up in step B31. When the tested call code fails a certain test, the call proxy is regenerated, with a maximum of 3 retries. If it still fails after 3 retries, the system automatically records the reason for failure and synchronizes it to the metadata database, while sending an alarm notification and terminating the current algorithm access process. Step B32 uploads the tested call proxies to the proxy management database, which can be used by external businesses at any time and securely.

[0114] The above solution encapsulates algorithm capabilities into standardized service components that are highly reliable, manageable, and immune to memory (fault tolerance), which significantly reduces the failure rate of algorithm services during operation, improves the overall system's SLA (Service Level Agreement), and makes the operation and maintenance management of algorithms centralized, transparent, and efficient.

[0115] like Figure 4 As shown, preferably, in response to a data interaction request for a target algorithm that has been connected, the corresponding data interaction rules are obtained from the metadata database, and based on the data interaction rules, the trained parameter prediction model, and the data characteristics of the current interaction data, data transformation and dynamic mapping are performed, while the mapping deviation is recorded to iteratively optimize subsequent mappings, including the following steps:

[0116] C1. In response to a data interaction request, extract the data features of the current interaction data;

[0117] C2. Input the data features into the trained parameter prediction model and output the optimal data transformation parameters;

[0118] C3. According to the data interaction rules, match the corresponding mapping rules (hereinafter referred to as 'mapping rules') from the dynamic mapping library, and based on the mapping rules and the optimal data conversion parameters, perform type conversion and format adaptation on the current interaction data to generate adapted data that conforms to the data interaction rules;

[0119] C4. Calculate the mapping deviation rate of this data transformation, and iteratively update the mapping rules in the dynamic mapping library based on the mapping deviation rate.

[0120] Preferably, the trained parameter prediction model is obtained through the following steps:

[0121] With the goal of minimizing the transformation bias, the linear regression model fitted by the least squares method is obtained by iteratively training using historical mapping data and gradient descent.

[0122] Among them, data characteristics include at least data type encoding and interaction scenario encoding.

[0123] Steps C1 to C4 are the specific implementation details of step S5. Through steps C1 to C3, a trained parameter prediction model is introduced. Before executing the mapping, real-time data features of the current data are extracted and input into the model to dynamically predict the optimal parameters for this transformation. This dynamic transformation design improves the granularity, accuracy, and scenario adaptability of the data transformation. Step C2, model selection, uses a trained least-squares fitted linear regression model, characterized by its simplicity, fast training and prediction speed, and strong interpretability, making it very suitable for online real-time prediction scenarios. Feature engineering explicitly uses data type encoding (e.g., int, float, string) and interaction scenario encoding (e.g., real-time query, batch analysis, model training), which is crucial for transforming business semantics into learnable features for the model. The optimization objective is to minimize the transformation deviation, directly addressing the core business requirement (data accuracy). Step C4 ensures that after each mapping execution, the system automatically calculates the mapping deviation rate as a quantitative assessment of the transformation quality. The evaluation results are not discarded but used to iteratively update the rules in the dynamic mapping library.

[0124] The above solution transforms data interaction from a fault-prone, rigid, and high-maintenance mechanical operation into a highly automated, adaptive, and increasingly intelligent operation. It significantly reduces the semantic gap and format friction in data integration, ensuring high-fidelity and high-efficiency transmission of data between business modules and algorithms, and between algorithms themselves.

[0125] In one embodiment, raw data access and feature extraction are the starting point for dynamic data mapping, triggered by a data input request initiated by a business module or a data output request initiated by the target algorithm. The business module sends raw business data to the data feature extraction unit, while the target algorithm sends its output data to this unit. The data feature extraction unit then calls the feature metadata definitions preset based on the metadata protocol in the metadata database to perform structured parsing of the data, ultimately outputting feature data including field names, data types, historical interaction records, and checksums, providing a standardized input basis for subsequent mapping rule matching.

[0126] Mapping rule matching and transformation preparation, as an intermediate link in dynamic data mapping, involves automatically matching corresponding transformation rules based on the mapping rule index preset by the meta-protocol in the dynamic mapping library after feature data is input; historical interaction records are input to the intelligent optimization unit to predict the optimal mapping parameters; then the matched rules and feature data are packaged and sent to the data transformation engine, which retrieves the target format template defined by the meta-protocol from the meta-information library to complete the rule and template preparation before transformation, laying a compliant foundation for accurate data adaptation.

[0127] Based on the matching mapping rules, the data transformation engine performs operations such as field completion, type conversion, and validation correction on the feature data to generate adapted data that conforms to the meta-protocol specification, which is then output to the business module and the target algorithm respectively. At the same time, the transformation engine synchronizes the real-time resource requirements during the transformation process to the resource scheduling and resource prediction scheduling unit, realizing the linkage between data processing and resource supply.

[0128] During operation, the platform periodically collects the mapping deviation rate of data interaction. If the deviation rate of a certain mapping rule remains high, the rule will be adjusted directly.

[0129] The system employs a JSON-formatted mapping rule table for dynamic mapping. This table has a concise structure, clearly defining the source data fields, target data fields, and transformation methods. For example, it converts a source field to a target field by converting a string to a number and retaining one decimal place. During data transformation, the transformation is performed directly according to the rule table. The deviation is calculated using a direct comparison method, comparing the transformed data with preset standard data. The formula is as follows:

[0130] ;

[0131] In the formula, For numerical data, the absolute deviation rate. For the converted data, Standard data preset for the meta-protocol;

[0132] To check if string data matches exactly, the ratio of the number of deviation fields to the total number of fields is calculated. The formula for calculating the overall mapping deviation rate is as follows:

[0133] ;

[0134] In the formula, The sum of the deviations of all fields. This represents the total number of fields in a single data interaction. Monthly mapping deviation records are compiled. If the deviation rate of a certain mapping rule consistently exceeds 5%, adjustments are made according to the following strategies: ① Deviation rate 5%-10%: Fine-tune mapping parameters; ② Deviation rate >10%: Replace the mapping rule; ③ After adjustment, mark the rule version and the reason for the adjustment, and synchronize them to the dynamic mapping library and metadata library to form a traceable iterative closed loop.

[0135] The trained least squares fitted linear regression model is used as the transformation parameter prediction tool, as shown in the following formula:

[0136] ;

[0137] In the formula, To determine the optimal conversion parameters for the output, such as the number of decimal places to retain and string processing identifiers, Discretized encoding of data type characteristics, such as int=1, string=2, etc. Discretized encoding of interactive scene features, such as real-time query = 1, batch processing = 2, etc. 、 For feature weights, the weights for different data types and different interaction scenarios can be preset. For bias terms;

[0138] The loss function formula is as follows:

[0139] 2 ;

[0140] in, This represents the number of training samples over the past 7 days. For the first The actual optimal transformation parameters for each sample These are the parameters for model prediction.

[0141] Among them, the mean square error can be quantified to transform the deviation, and the gradient calculation is continuous and smooth, which is suitable for the iterative optimization objective of the gradient descent method.

[0142] The gradient descent method is used to iteratively optimize the parameters, and the parameter update formula is as follows:

[0143] ;

[0144] ;

[0145] Among them, settings The learning rate is fixed at 0.05 and requires no manual adjustment. The index for the training samples, i.e., a single historical data conversion record, and For the ordinal identifier of the input feature, such as =1,2, corresponding to data types and interaction scenarios respectively; stop after 5 iterations when Loss ≤ 0.1; training data is the mapping records of the past 7 days; input features are data types (such as int / string) and interaction scenarios (such as real-time query / batch processing); output is the optimal conversion parameters (such as the number of digits to retain for numerical precision and the string processing method); the training objective is to minimize the conversion bias.

[0146] like Figure 5 As shown, preferably, in response to a resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data transformation and dynamic mapping process, a preset time series prediction model is used to predict resource requirements, and dynamic resource scheduling is performed based on the prediction results and the real-time resource status, including the following steps:

[0147] D1. In response to resource scheduling requests, resource scheduling decision data is generated based on historical consumption patterns represented by operating environment requirements and resource consumption trend data, combined with real-time resource status and real-time resource requirements during data conversion.

[0148] D2. Call the preset time series prediction model and predict the peak resource demand in future periods based on historical consumption data in the resource scheduling decision data;

[0149] D3. Generate a resource allocation plan based on peak resource demand and real-time resource status;

[0150] D4. Allocate and isolate dedicated computing resources and operating environments for the target algorithm and data transformation process through resource isolation mechanisms.

[0151] Preferably, the preset time series forecasting model adopts the moving average method, which calculates the predicted value of resource demand for the same period in the future based on the actual resource consumption value of the same period in the past N cycles, where N is a positive integer.

[0152] Preferably, the resource demand forecasting formula for the preset time series forecasting model is:

[0153] ;

[0154] in, This represents the projected resource demand for future periods. For the past The same period of each cycle The actual value of resource consumption, The number of periods used to calculate the average value is a positive integer.

[0155] Steps D1 to D4 are the specific implementation details of step S6. After receiving the scheduling request, steps D1 and D2 analyze the historical consumption patterns represented by resource consumption trends and call the time-series prediction model to predict the peak resource demand in future periods. This allows for advance prediction of the resource pressure generated by the algorithm task at a certain moment, enabling resource planning and reservation in advance before the bottleneck actually occurs. This fundamentally avoids performance jitter, task queuing, or failure caused by instantaneous resource shortages, significantly improving service stability and response determinism. In particular, the resource scheduling decision data in step D1 takes into account the runtime environment requirements in the meta-protocol, historical consumption patterns, real-time system load, and real-time overhead during data conversion, thus considering how resources should be scheduled from a global perspective and avoiding resource waste. The preset time-series prediction model in step D2 adopts the moving average method, which is a classic and simple time series forecasting and smoothing technique. Its core idea is to predict the value of the next period or describe the long-term trend of the sequence by calculating the average of multiple recent consecutive observations in the time series, while filtering out short-term random fluctuations. This method extracts stable and repeatable consumption patterns from the noisy historical resource consumption data of the algorithm and filters out occasional sudden high loads, which makes the prediction results more stable and reliable, and avoids excessive or insufficient resource reservation due to individual outliers. Step D4 allocates dedicated resources and environment to the algorithm, data transformation and dynamic mapping through a resource isolation mechanism, which reduces the mutual interference of resource components and improves the SLA of the overall system.

[0156] The above solution upgrades resource supply from an "on-demand (allocation already occurred)" model to an "allocation based on prediction" model. This significantly improves the resource utilization, system throughput, and task success rate of the entire algorithm service platform, while ensuring the service quality of high-priority or critical algorithm services. It solves the most troublesome resource contention uncertainty problem when algorithms run in production environments.

[0157] In one embodiment, resource scheduling data acquisition and preparation, as a prerequisite for resource scheduling, is linked with dynamic data mapping. The triggering condition is the initialization of the target algorithm or resource requirements reported by the data conversion engine. The resource scheduling unit extracts core data from the metadata database, including algorithm environment requirements based on pre-defined metadata protocols, historical resource consumption, real-time resource status of the current node, and resource requirements of the data conversion stage. It then calls a time-series prediction model to predict future consumption data based on historical consumption data and aggregates this data into a resource scheduling decision package, serving as the quantitative basis for subsequent resource allocation.

[0158] Dynamic resource allocation and operation support services are provided for dynamic data mapping and algorithm execution. The resource scheduling unit generates differentiated resource allocation schemes based on decision packages and distributes the schemes to the resource management module. The module divides dedicated resource areas through a resource isolation mechanism, allocates suitable hardware resources and software environments to the target algorithm and data conversion engine, monitors resource usage in real time, and automatically allocates resources based on predicted data to support efficient data processing and stable algorithm operation.

[0159] Version synchronization and metadata updates serve as a closed-loop mechanism to ensure consistency between dynamic data mapping and resource scheduling. When the version management unit detects a meta-protocol version iteration, it triggers version synchronization, generating a version log containing updated content and effective time, as well as corresponding mapping rule update packages and scheduling basis update packages. Upon detecting cross-layer dependency conflicts, it automatically upgrades the dynamic mapping library version and simultaneously synchronizes the mapping rule update package to the dynamic mapping library and the scheduling basis update package to the resource scheduling unit. This achieves coordinated upgrades of data mapping rules and resource scheduling basis, ensuring that the entire chain adheres to a unified meta-protocol version.

[0160] The preset time series forecasting model uses the moving average method as the time series forecasting approach to predict resource demand, as shown in the following formula:

[0161] ;

[0162] In practical applications, the period unit is either an hour or a day, depending on the type of algorithm task: daily tasks are in days, while real-time tasks are in hours. The value ranges from 3 to 30, and the system automatically adapts to the fluctuation of historical consumption data. The value is adjusted to balance prediction sensitivity and stability.

[0163] in, This represents the predicted resource demand for future time period t. For (N takes the value of 7) the same time period in the past 7 days The actual resource consumption is calculated, and the average consumption over the past 7 days is used as the predicted value for the same period in the future. Basic resources are allocated based on the predicted value. If a failure occurs, the process is restarted and resources are reallocated, and the failure record is updated to the metadata database. Conflicts are resolved using a simple priority sorting method, which considers factors such as business dependency, environmental compatibility, and version stability, and filters versions from high to low priority. No complex algorithm calculations are required, and the conflict is synchronized to the relevant layer after resolution.

[0164] The resource isolation mechanism is implemented using Docker container technology, and the process is as follows:

[0165] ① Create a dedicated Docker container to allocate an independent running environment for the target algorithm and data transformation process;

[0166] ② Configure container resource limits;

[0167] ③ Enable network isolation;

[0168] ④ Mount a dedicated data volume;

[0169] In practical applications, a Docker container is allocated for the project risk assessment algorithm developed in Python, configured with 2 CPU cores and 4GB of memory. The network is only open to the 8080 communication port with the business module, and a dedicated data volume / data / risk-assess is mounted to ensure that the algorithm's operating resources do not conflict with other modules.

[0170] like Figure 6 As shown, a dynamic mapping and project resource scheduling system based on meta-protocols includes:

[0171] The verification and protocol generation module is used to configure the target algorithm according to the preset meta-protocol specification to generate the initial meta-protocol configuration file, and to verify it; after the verification is successful, a standard meta-protocol file is generated.

[0172] The information extraction and storage module is used to parse the standard meta-protocol file, extract the core and extended information, and store it in the meta-information database;

[0173] The plugin adaptation and dependency management module is used to match or generate corresponding algorithm adaptation plugins based on the algorithm language type in the core information, and automatically detect and repair runtime environment dependencies based on the runtime environment requirements in the core information.

[0174] The proxy generation and access module is used to generate a call proxy corresponding to the algorithm adaptation plugin based on the data interaction rules in the core information and the historical fault records in the extended information, and to test the call proxy; after the test is passed, the call proxy is registered to the proxy management library to complete the access of the target algorithm.

[0175] The data mapping and optimization module is used to respond to data interaction requests for the connected target algorithm, obtain the corresponding data interaction rules from the metadata database, and perform data transformation and dynamic mapping based on the data interaction rules, the trained parameter prediction model and the data characteristics of the current interaction data, while recording the mapping deviation to iteratively optimize subsequent mappings;

[0176] The resource prediction and scheduling module is used to respond to resource scheduling requests for the target algorithm. Based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data transformation and dynamic mapping process, it uses a preset time series prediction model to predict resource requirements and performs dynamic resource scheduling based on the prediction results and real-time resource status.

[0177] This invention also discloses a dynamic mapping and project resource scheduling system based on meta-protocols. It aims to achieve dynamic mapping and resource scheduling in a modular manner. The system employs a verification and protocol generation module to configure target algorithms according to preset meta-protocol specifications, generate an initial meta-protocol configuration file, and verify it. Upon successful verification, a standard meta-protocol file is generated. An information extraction and storage module parses the standard meta-protocol file, extracts its core and extended information, and stores it in a meta-information database. A plugin adaptation and dependency management module matches or generates corresponding algorithm adaptation plugins based on the algorithm language type in the core information and automatically detects and repairs runtime environment dependencies based on the runtime environment requirements in the core information. Finally, a proxy generation and access module generates a proxy corresponding to the algorithm adaptation plugin based on the data interaction rules in the core information and historical fault records in the extended information. The system utilizes a call proxy to test the call proxy. After passing the test, the call proxy is registered to the proxy management library, completing the access to the target algorithm. Through the data mapping and optimization module, in response to data interaction requests for the accessed target algorithm, the system retrieves the corresponding data interaction rules from the metadata library. Based on the data interaction rules, the trained parameter prediction model, and the data characteristics of the current interaction data, it performs data transformation and dynamic mapping, while recording mapping deviations to iteratively optimize subsequent mappings. Through the resource prediction and scheduling module, in response to resource scheduling requests for the target algorithm, the system uses the runtime environment requirements and resource consumption trend data stored in the metadata library, combined with the real-time resource status of the current node and the real-time resource requirements during the data transformation and dynamic mapping process, to predict resource requirements using a preset time-series prediction model. Based on the prediction results and the real-time resource status, the system performs dynamic resource scheduling.

[0178] In one embodiment, the implementation logic of this solution is illustrated by using an enterprise-level project management platform integrated with a Python-based project schedule deviation prediction algorithm. This deviation prediction algorithm predicts schedule deviations based on project task completion rate, resource allocation ratio, and milestone achievement status, and achieves the aforementioned process through the collaborative implementation of the following functional modules:

[0179] Validation and Protocol Generation: Configure core information according to the meta-protocol specification, including language type Python, data interaction rules (input task_data: JSON, output deviation: float), runtime environment requirements Python 3.x (e.g., 3.9), Pandas 1.x (e.g., 1.5), Scikit-learn 1.x (e.g., 1.2), and version number V1.x (e.g., V1.0). Extended information includes moderate interaction frequency, historical fault records, task completion rate, number of data format errors, and resource consumption trend with peak values ​​of low to moderate resource configuration. Generate a standard JSON format meta-protocol file after validity verification.

[0180] Information extraction and storage: Parse the standard meta-protocol file, extract core and extended information and store it in the meta-information database, and establish an association index of algorithm identifier, scenario and resource requirements.

[0181] Plugin adaptation and dependency management: Retrieves and loads Python adaptation plugins from the adaptation plugin library. If a missing Pandas dependency is detected, the system automatically downloads, installs, and repairs it.

[0182] Proxy generation and integration: Call the Python proxy template, inject the task_data with the task_id field, the completion rate range [0,100], the fault tolerance logic of retrying 3 times after timeout, generate the proxy calling code, and after passing the parameter interception, data compliance, and connectivity tests, register the proxy to the proxy management library to complete the algorithm integration.

[0183] Data Mapping and Optimization: The business module initiates a data interaction request, extracts the current data features as JSON type and real-time prediction scenario, inputs the trained parameter prediction model, and outputs the optimal conversion parameters as numerical precision with 1 decimal place and null padding; matches the dynamic mapping library rules, converts Java format task data into a Python algorithm-compatible format, calculates the mapping deviation rate to be less than 1%, and updates the mapping rules synchronously.

[0184] Resource Prediction and Scheduling: In response to resource scheduling requests, based on the recent low to medium resource consumption trends of multiple periods in the metadata database and the availability of real-time idle CPUs, the algorithm predicts medium resource allocation for peak resource demand using the moving average method; generates an allocation scheme, allocates dedicated resources and isolates them through Docker containers, ensuring no resource shortages during operation, and outputs stable progress deviation prediction results.

[0185] One or more embodiments in this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of this application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments in this application should be included within the protection scope of this application.

[0186] If a flowchart is used in this application, it is used to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0187] The foregoing has provided a detailed description of a method and system for dynamic mapping and project resource scheduling based on meta-protocols, as provided in this application. The above description of the disclosed embodiments enables those skilled in the art to implement or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A dynamic mapping and project resource scheduling method based on meta-protocols, characterized in that, Includes the following steps: According to the preset meta-protocol specification, configure the target algorithm to generate the initial meta-protocol configuration file and verify it; after the verification is successful, generate the standard meta-protocol file. The standard meta-protocol file is parsed to extract its core and extended information, which is then stored in the meta-information database. Based on the algorithm language type in the core information, match or generate the corresponding algorithm adaptation plugin, and based on the runtime environment requirements in the core information, automatically detect and repair runtime environment dependencies. Based on the data interaction rules in the core information and the historical fault records in the extended information, a call proxy corresponding to the algorithm adaptation plugin is generated, and the call proxy is tested. After the test is passed, the calling proxy is registered to the proxy management library to complete the access of the target algorithm; In response to a data interaction request for the target algorithm that has been accessed, the corresponding data interaction rules are obtained from the metadata database. Based on the data interaction rules, the trained parameter prediction model, and the data features of the current interaction data, data transformation and dynamic mapping are performed, and mapping deviations are recorded to iteratively optimize subsequent mappings. In response to a resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data conversion and dynamic mapping process, a preset time series prediction model is used to predict resource requirements, and dynamic resource scheduling is performed based on the prediction results and the real-time resource status.

2. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 1, characterized in that, The core information includes at least: the unique identifier of the target algorithm, the algorithm language type, data interaction rules, runtime environment requirements, and version number; The extended information includes at least: algorithm interaction frequency, historical fault records, and resource consumption trend data.

3. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 1, characterized in that, The process of matching or generating corresponding algorithm adaptation plugins based on the algorithm language type in the core information, and automatically detecting and repairing runtime environment dependencies based on the runtime environment requirements in the core information, includes the following steps: Based on the algorithm language type, a matching algorithm adaptation plugin is retrieved from a preset adaptation plugin library; If the search is successful, the matching algorithm adaptation plugin is loaded; if the search fails, the target algorithm's interface is parsed through a syntax tree based on the algorithm's language type and the data interaction rules, a preset language template is matched, and data verification logic defined by the data interaction rules is injected to generate a new algorithm adaptation plugin. After loading or generating the algorithm adaptation plugin, it checks whether the current runtime environment meets the runtime environment requirements and performs automatic repair for missing or mismatched dependencies.

4. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 3, characterized in that, Based on the data interaction rules in the core information and the historical fault records in the extended information, a call proxy corresponding to the algorithm adaptation plugin is generated, and the call proxy is tested. After the test is passed, the call proxy is registered to the proxy management library to complete the access of the target algorithm, including the following steps: Invoke the proxy template corresponding to the algorithm language type, inject the calling rules and data verification logic defined by the data interaction rules into it, and inject fault tolerance logic based on the fault type indicated by the historical fault record to generate executable calling proxy code; Based on the aforementioned call proxy code, parameter interception, data compliance, and connectivity tests are performed. If the test fails, the proxy will be regenerated. If the test passes, the agent will be registered with the agent management library to complete the access to the target algorithm.

5. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 1, characterized in that, In response to a data interaction request for the accessed target algorithm, the corresponding data interaction rules are obtained from the metadata database. Based on the data interaction rules, the trained parameter prediction model, and the data features of the current interaction data, data transformation and dynamic mapping are performed, while recording the mapping deviation to iteratively optimize subsequent mappings. This includes the following steps: In response to a data interaction request, extract the data features of the current interaction data; The data features are input into the trained parameter prediction model, which outputs the optimal data transformation parameters. According to the data interaction rules, the corresponding mapping rules are matched from the dynamic mapping library, and based on the mapping rules and the optimal data conversion parameters, the current interaction data is converted in type and adapted in format to generate adapted data that conforms to the data interaction rules. Calculate the mapping deviation rate of this data transformation, and iteratively update the mapping rules in the dynamic mapping library based on the mapping deviation rate.

6. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 5, characterized in that, The trained parameter prediction model is obtained through the following steps: With the goal of minimizing the transformation bias, the linear regression model fitted by the least squares method is obtained by iteratively training using historical mapping data and gradient descent. The data features include at least data type encoding and interaction scenario encoding.

7. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 1, characterized in that, The step of responding to a resource scheduling request for the target algorithm, based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data transformation and dynamic mapping process, uses a preset time-series prediction model to predict resource requirements, and performs dynamic resource scheduling based on the prediction results and real-time resource status, including the following steps: In response to a resource scheduling request, resource scheduling decision data is generated based on the operating environment requirements and the historical consumption patterns represented by the resource consumption trend data, combined with the real-time resource status and the real-time resource requirements of the data conversion process. The preset time-series prediction model is invoked to predict the peak resource demand for future periods based on historical consumption data in the resource scheduling decision data. Based on the peak resource demand and the real-time resource status, a resource allocation scheme is generated; The target algorithm and data transformation process are allocated and isolated using a resource isolation mechanism, with dedicated computing resources and operating environment.

8. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 7, characterized in that, The preset time series prediction model adopts the moving average method, which calculates the predicted value of resource demand for the same period in the future based on the actual resource consumption value of the same period in the past N cycles, where N is a positive integer.

9. The dynamic mapping and project resource scheduling method based on meta-protocol as described in claim 8, characterized in that, The resource demand prediction formula of the preset time series prediction model is: ； in, This represents the projected resource demand for future periods. For the past The same period of each cycle The actual value of resource consumption. The number of periods used to calculate the average value is a positive integer.

10. A dynamic mapping and project resource scheduling system based on meta-protocols, characterized in that, include: The verification and protocol generation module is used to configure the target algorithm according to the preset meta-protocol specification to generate the initial meta-protocol configuration file and verify it. After successful verification, a standard meta-protocol file is generated. The information extraction and storage module is used to parse the standard meta-protocol file, extract the core information and extended information therein, and store them in the meta-information database; The plugin adaptation and dependency management module is used to match or generate corresponding algorithm adaptation plugins based on the algorithm language type in the core information, and to automatically detect and repair runtime environment dependencies based on the runtime environment requirements in the core information. The proxy generation and access module is used to generate a call proxy corresponding to the algorithm adaptation plugin based on the data interaction rules in the core information and the historical fault records in the extended information, and to test the call proxy; After the test is passed, the calling proxy is registered to the proxy management library to complete the access of the target algorithm; The data mapping and optimization module is used to respond to the data interaction request for the connected target algorithm, obtain the corresponding data interaction rules from the metadata database, and perform data transformation and dynamic mapping based on the data interaction rules, the trained parameter prediction model and the data characteristics of the current interaction data, while recording the mapping deviation to iteratively optimize the subsequent mapping. The resource prediction and scheduling module is used to respond to resource scheduling requests for the target algorithm. Based on the runtime environment requirements and resource consumption trend data stored in the metadata database, combined with the real-time resource status of the current node and the real-time resource requirements of the data conversion and dynamic mapping process, it uses a preset time series prediction model to predict resource requirements and performs dynamic resource scheduling based on the prediction results and real-time resource status.

Images