75 results about "Function definition" patented technology

The invention discloses an intention understanding tool calling method based on a large model function call technology, and the method comprises the following steps: S1, obtaining a natural language instruction of a user, and carrying out the preprocessing of the natural language instruction; s2, based on the structured instruction fragment and the context dialogue information, constructing a semantic vector for intention recognition, and inputting the semantic vector into a large language model; s3, screening function description information from a function definition library according to an intention recognition result; s4, extracting parameter content from the semantic vector under the constraint of a parameter limiting condition; s5, constructing a call request format conforming to a function call interface specification, and executing a corresponding tool call task; and S6, receiving a function calling result, generating a natural language response text, and outputting the natural language response text to a user interface to complete an instruction response closed loop. According to the method, full-automatic semantic mapping from a natural language instruction to tool function calling is realized, and the analysis accuracy and the calling success rate are improved.

Various embodiments described herein support or provide operations including receiving input data that comprises a function definition; generating a schema that describes output data in an output format expected in a function output; and executing a function based on the function definition. The execution of the function includes generating a prompt that includes the schema and the input data; using a machine learning model to generate a response based on the prompt, the machine learning model comprising at least one large language model; and generating the function output by converting the response into a programming data object in the output format described by the schema.

A generative artificial intelligence system includes a retrieval augmented generation (RAG) assistant that utilizes function calling to facilitate multi-source data retrieval to enhance user queries transmitted to a large language model (LLM). The RAG assistant Transmits, to the LLM, a function selection instruction prompt that includes conversation history data, a function list including function definitions that each correspond to a data source, and instructions directing the LLM to return a function call to at least one function defined on the function list identified as relevant to the conversation history data based on a corresponding function descriptor. In response to receiving a function selection response from the LLM that includes the function call, the RAG assistant selects and executes a conditional operation based on a name of the at least one function.

The invention relates to the technical field of artificial intelligence and software development, in particular to a context-aware code generation and real-time optimization method and system based on a large language model.The method comprises the following steps that a compiler front-end technology is utilized, input codes are analyzed, and an abstract syntax tree is constructed; function definition, variable declaration and statement block structure information are extracted through an abstract syntax tree, so that syntax and semantic structures of codes are obtained, and basic data support is provided for subsequent code generation and optimization; the method has the beneficial effects that the context graph covering grammar, semantics and runtime behaviors is constructed by fusing static grammar analysis (AST construction) and dynamic execution tracking (such as variable value change and function call time sequence), and the problem of context perception deficiency in a traditional code generation technology is solved. And the compatibility, the maintainability and the expandability of the generated code and the project architecture are improved.

The invention belongs to the field of software testing, and relates to an ROS 2 software package defect detection method based on a large language model and coverage rate guidance. According to the method, function information is extracted from a to-be-tested program, and a regular expression is used for analyzing function definition, parameter characteristics and other information; the method comprises the following steps: presetting verification standards related to a to-be-tested program, and extracting physical constraints and document constraints from the verification standards; integrating the extracted function information, physical constraint and document constraint into accurate information, filling a cue word template with the accurate information, inputting the cue word template into a large language model, and generating an initial test seed; a mutation strategy is introduced, and diversified mutation operation is performed on the generated seeds, so that test cases with higher diversity are generated; optimizing the test case by adopting a coverage rate guiding mechanism, and guiding to generate a new test case by monitoring the coverage rate in real time; through a defect verification mechanism, dominant defects such as program crash and the like can be detected, and hidden defects such as violation of physical constraints or violation of document constraints and the like can be identified.

The invention discloses an interface scanning method and device based on a GO language application, equipment and a medium, and relates to the technical field of software development, and the method comprises the steps: in the compiling process of a target GO language application, if it is detected that an enabling instruction of a preset GO language tool is triggered, obtaining a target code file in the target GO language application, generating an abstract syntax tree corresponding to the target code file; the abstract syntax tree is traversed, interface nodes are determined from all nodes of the abstract syntax tree, routing paths and function definitions corresponding to the interface nodes are extracted, and the function definitions are specific function implementation in the nodes; searching incoming parameters and return parameters corresponding to the function definitions in the interface nodes based on the function definitions; and generating an interface scanning document according to each interface node, the routing path, the function definition, the corresponding incoming parameter and the corresponding return parameter. According to the method and the device, the API document collection efficiency in the GO language application can be improved.

The embodiment of the invention discloses a test script generation method, system and device, a medium and a product, and relates to the technical field of tests.The method comprises the steps that target structured DBC data is obtained; processing the target structured DBC data by utilizing a target test script generation model to obtain at least one CAPL test script fragment; the target test script generation model is obtained by training according to historical structured DBC data and corresponding historical CAPL test script fragments; the CAPL test script fragment comprises at least one of variable declaration, function definition and event processing; and according to the at least one CAPL test script fragment, determining a target CAPL test script so as to quickly realize verification of the CAN bus communication condition.

The invention discloses a dynamic analysis-combined memory security vulnerability fuzzy test method. The method comprises the following steps of inserting pointer metadata for dynamic analysis into a source code file of a to-be-tested program, performing memory security detection, and performing related type definition and function definition of a memory allocation packaging function; analyzing the source code by using a compiler, generating an abstract syntax tree, traversing the abstract syntax tree, performing instrumentation on nodes of the abstract syntax tree, and maintaining a shared memory variable of a minimum pointer boundary distance and a memory allocation peak value; compiling the source code after instrumentation to obtain an assembly code; performing coverage rate instrumentation on the assembly code, and skipping the previously inserted code for dynamic analysis; assembling and linking the instrumented assembly codes to obtain an executable file; performing fuzzy testing on the executable file; and running the fuzzy test to output a program input file which enables the target program to have a memory security error. According to the method, memory vulnerabilities in the target program can be found, and the error detection capability of fuzzy testing is improved.

A compute server of a distributed cloud computing network receives an inference request. The inference request triggers execution of code that is related to an AI application that interacts with the inference request and causes its input to be run through an AI model. Helper code executing on the compute server transmits an inference request to the AI model, where this inference request includes a prompt and one or more function definitions, and where the AI model is executing in the same datacenter as the compute server. The helper code receives a result of the AI model executing the inference request, where it includes structured data for calling a function. The helper code calls the function using the structured data. The helper code receives a response to the called function. The code processes a response to the inference request based at least on the received response to the called function.

The invention discloses a vehicle body domain controller hardware-in-the-loop simulation method under an AUTOSAR architecture, and the main conception of the method is that based on the standardized interface definition of an AUTOSAR basic software layer, an HIL cabinet hardware channel standardized mapping relation library is established, and thus a hardware corresponding relation configuration file template is constructed; establishing a subsystem simulation model for each functional module and constructing a three-layer decoupling architecture in combination with different functional specifications to realize dynamic binding of test software and a hardware channel; then, based on an actual hardware interface and software function definition, a simulation test bench is built, and verification is completed; and finally, generating an automatic test case based on the function specification, and carrying out the function test of each subsystem of the vehicle body domain controller in the current test environment. According to the invention, the rack development period is obviously shortened, the flexibility and expandability of the test environment are improved, and the full-life-cycle test verification capability optimization is realized.

The invention discloses a combination instrument test method and device and a medium, and the method comprises the steps: obtaining a function definition of a to-be-tested combination instrument, carrying out the analysis and classification of the function definition based on a preset function module division rule, and generating a plurality of test function module sets; for each test function module set, generating a corresponding automatic test script according to the display type and the association relationship of the display signals in the test function module set; automatic test scripts corresponding to the selected target function modules are loaded and run in sequence, and corresponding display signals are sent to the to-be-tested combination instrument according to the sequence, value and time sequence logic defined in the automatic test scripts; and observing and recording an actual display state of the to-be-tested combination instrument after executing the corresponding automated test script, and comparing and analyzing the actual display state and an expected display state predefined in the test case to generate a test judgment result for the target function module.

The invention relates to a computer-readable recording medium storing a program, a data processing apparatus, and a data processing method. The recording medium stores a program for causing a computer to execute: receiving user definition information including function definition information defining an objective function of a combinatorial optimization problem and neighborhood definition information defining a neighborhood state when a solution of the problem is searched by a local search; generating an optimization system using the user-defined information and basic information indicating a process irrelevant to the definition of the function and the definition of the neighborhood state; receiving coefficients in the function; and executing an optimization system to calculate a change amount of a value of the function when a state represented by a state variable in the function transitions to a neighborhood state using the function definition information, the neighborhood definition information, and the coefficient, select a first state from the neighborhood state based on the change amount, and update the state variable to the first state.

The invention provides a hybrid computing task classification method and device, equipment and a medium, and the method comprises the steps: carrying out the grammar analysis of an input hybrid computing code, generating a tree structure reflecting a code structure, and extracting subtask units and unit information in the hybrid computing code based on the tree structure; extracting feature words of the unit information according to the unit information; obtaining a subtask type of a predefined hybrid computing task and a corresponding keyword; and respectively mapping the keywords and the feature words into semantic vectors, and calculating semantic similarity to determine the subtask type of the subtask unit. The tree structure is obtained through the abstract syntax tree algorithm, so that the hybrid computing code is analyzed, the function definition and the dependency relationship thereof are accurately extracted, and subjectivity and errors possibly brought by manual intervention are avoided.

The invention belongs to the technical field of industrial dynamic formula execution, and discloses an analysis-free high-availability and high-performance formula execution method. The method comprises the following steps: S1, formula extraction and function body construction: receiving a calculation formula input by a user, and taking the calculation formula as a function body core calculation logic; s2, function packaging: adding a function head and a return value statement for a function body, and constructing a complete function definition character string; wherein the function head comprises a function name and a parameter list; and S3, function persistence and dynamic import execution: writing a function character string into an independent Python module file, and calling and executing through a dynamic import mechanism. According to the method, function packaging is adopted, the repeated lexical analysis and grammar analysis process which must be carried out when a traditional event () method is executed each time is eliminated, and the problems that a formula is forcibly required to only comprise functions of input parameters, calculation expressions and return values, and the bottleneck of industrial real-time application and the fragility of a whole framework are solved.

The invention is suitable for the technical field of GPU (Graphics Processing Unit), and provides an SPIR-V-based inter-thread shared memory processing method, which comprises the following steps of: S1, reading a high-level language code source code to a character buffer area, and then converting the source code in the character buffer area into an intermediate code in an SPIR-V form; s2, traversing an instruction sequence of an intermediate code, and for variable declarations in the instruction sequence, accumulating the sum of occupied spaces of all shared memory variables; s3, calculating a one-dimensional index and a shared memory address of the thread group; and S4, traversing the instruction sequence of the intermediate code, and for instructions in the function declaration and the function definition, if processing of a shared memory is involved, calculating a real address and generating different GPU instructions according to instruction types. The invention provides a universal method for translating related codes, used for processing a shared memory of a working group, in SPIR-V codes into basic instructions of all GPU architectures, and the instructions used in the method are basic instructions of all GPU instruction sets, so that the method has hardware platform universality.

Technologies for providing attestation for function as a service flavors include a compute device including circuitry configured to obtain function definition data indicative of a set of operations to be performed in a function and a set of hardware resources to be utilized by the function, execute a benchmark operation to produce benchmark data indicative of a measured performance of the function, and sign the function definition data and the benchmark data to produce function flavor data. The circuitry is also configured to provide the function flavor data to one or more other compute devices for validation that the function, when executed on the hardware resources, provides the measured performance and write, to a distributed ledger, the function flavor data.

The invention discloses a track section automatic turn-back attribute and safe area generation method and device and a medium. The method comprises the steps that topological data and function definition data of a track traffic line are input; on the basis of the incidence relation between the function definition data and the topological data, through a predefined attribute calculation rule, automatic turn-back attributes of related track sections in the line are automatically calculated, and the automatic turn-back attributes are used for identifying function roles of the track sections in automatic turn-back; on the basis of the automatic turn-back attributes of all the track sections, one or more continuous automatic turn-back safety function areas are automatically generated through a path search algorithm; and outputting the automatic turn-back attributes and the automatic turn-back safety function area information of all the track sections. Compared with the prior art, the method has the advantages that the track section attributes are automatically calculated, the safety function area is generated, and automatic and accurate conversion from the line data to the automatic turn-back control data is achieved.

The invention discloses an automatic vehicle-mounted protocol analysis and function verification method and device, equipment and a medium, and belongs to the technical field of communication. The method includes performing a configuration on a pre-selected packet capture filter to capture an original packet; performing multi-round data cleaning on the original data packet, storing the original data packet as a preset format file, and performing data division; establishing a function definition mapping relation for each field; loading the preset format file by using an automatic analysis tool and analyzing the data packets one by one so as to extract related fields from the analyzed data packets, and then executing format conversion; and matching the related fields subjected to format conversion with a pre-established function definition model so as to identify the specific meaning of each related field and the corresponding vehicle-mounted function thereof. Through automatic analysis and mapping function definition, manual intervention is remarkably reduced, the efficiency and accuracy of data analysis are improved, the cost of data analysis and function verification is reduced, and the resource utilization rate is increased.

The invention aims to provide a three-dimensional function extreme value dynamic visualization teaching software algorithm based on a gradient rising method, which helps students to intuitively understand the principle and effect of the gradient rising method by dynamically displaying the gradient rising process, and improves the teaching effect. The three-dimensional function definition module defines a three-dimensional function f (x, y, z), and the function has a clear extreme point. The gradient calculation module calculates the gradient f (x, y, z) of the function f (x, y, z). The gradient ascending algorithm module utilizes a gradient ascending method to iteratively update parameters and gradually approaches extreme points of the function. And the dynamic visualization module gradually displays the track change after each iteration through an animation updating function, so that students can intuitively observe the dynamic change in the gradient rising process. And the animation storage module stores the dynamic visualization result as an animation file in a GIF format, so that teaching demonstration and student review are facilitated.

The invention provides an NCCL test method and device and electronic equipment. The NCCL test method comprises the steps that function definition parameters are acquired; according to the function definition parameters, defining the test times of the main function and the number of graphics cards; calling a test function, wherein the test function is used for testing the NCCL high-performance communication library; and based on the test function, executing a test operation to obtain a test result. According to the method, the NCCL high-performance communication library is tested through the test function, so that the technical problems of obvious insufficiency in the aspects of communication efficiency, expansibility, code complexity, compatibility, optimization and maintenance cost and the like in the prior art are solved, and the test performance is improved.

This invention discloses a data encoding system and method based on simplicated complex homogeneous linear codes. The system includes: a parameter initialization module for initializing the construction parameters of an n-ary multivariable function F; a function definition module for generating the multivariable function F based on the construction parameters; and a definition set extraction module for constructing a definition set D. F ; Linear code generation module, used to generate linear codes according to the definition set D F This invention employs an inner product mapping to construct a basic p-ary linear code corresponding to the defined set; a homogenization processing module is used to homogenize the multivariable function F to obtain an (n+1)-ary homogeneous function; and a homogenized code generation module is used to generate a homogenized definition set based on the homogenized function, and then generate a homogenized p-ary linear code based on this set. This invention solves the problems of existing p-ary linear code construction techniques being complex in structure, having uncontrollable parameters, or being difficult to analyze weight distribution, leading to difficulties in efficient deployment in engineering.

The application discloses a hardware-in-loop simulation method of a vehicle body domain controller under an AUTOSAR architecture, and the main idea of the application is that a standardized mapping relationship database of a HIL cabinet hardware channel is established based on standardized interface definition of an AUTOSAR basic software layer, and a hardware corresponding relationship configuration file template is constructed; in combination with different function specifications, a subsystem simulation model is built for each function module, a three-layer decoupling architecture is constructed, and dynamic binding of test software and the hardware channel is realized; then, based on actual hardware interfaces and software function definitions, a simulation test bench is built and verification is completed; finally, automatic test cases are generated based on the function specifications, and each subsystem function test of the vehicle body domain controller is carried out in the current test environment. The application significantly shortens the development cycle of the test bench, improves the flexibility and expandability of the test environment, and optimizes the whole life cycle test verification capability.

The invention discloses a log generation method and device, electronic equipment and a storage medium, and the log generation method comprises the steps: receiving a second dictionary table corresponding to a first software code of an upgrade version sent by a cloud, the second dictionary table is obtained by acquiring the change content of the first software code, which is changed compared with the second software code of the original version, and updating the first dictionary table corresponding to the second software code according to the change content, and the change comprises at least one of modification, addition and deletion; the change content comprises at least one of code files, function declarations and function definitions, and table entries in the dictionary table comprise combination names obtained after the function declarations and the code files where the function declarations are located are combined and identifiers corresponding to the combination names; in the process of generating the log generated when the first software code is operated, the combined name corresponding to the identifier is replaced based on the identifier in the second dictionary table, the generated log is obtained, and then the size of the log is compressed.

A data processing program causing a computer to execute: receiving user definition information including: function definition information defining an objective function of a combinatorial optimization problem and neighborhood definition information defining neighborhood states at a time of searching for a solution to the problem by a local search; generating an optimization system using the user definition information and base information indicating processing independent of: a definition of the function and a definition of the neighborhood states; receiving a coefficient in the function; and executing the optimization system to: calculate a change amount of a value of the function when a state represented by state variables in the function transitions to the neighborhood states, using the function definition information, the neighborhood definition information, and the coefficient, select a first state among the neighborhood states based on the change amount, and update the state variables to the first state.

The invention discloses an embedded system. The embedded system comprises a CPU (Central Processing Unit), an interface function reconfiguration controller and a first connector, the input end of the interface function reconfiguration controller is connected with each functional interface of the CPU in a one-to-one correspondence manner, the output end of the interface function reconfiguration controller is connected with the input end of the first connector in a one-to-one correspondence manner, and the output end of the first connector is used for being connected with external equipment; the interface function reconfiguration controller is configured to receive an interface function reconfiguration instruction; determining a target interface and a target function definition of an interface function reconstruction controller according to the interface function reconstruction instruction; and reconstructing the function definition of the target interface of the interface function reconstruction controller according to the target function definition. In the embodiment of the invention, the interface function reconfiguration controller is introduced into the embedded system, and the interface function definition is dynamically configured through the interface function reconfiguration controller, so that the pin function of the connector is changed, the limitation of interface curing of a traditional module is broken through, and the resource waste is reduced.

The embodiment of the invention provides a mobile terminal application program exception attribution method and device based on user experience measurement. The method comprises the steps that full-link data of a mobile terminal application program are collected in real time; according to the full-link data and the user experience measurement system, calculating a user experience measurement comprehensive evaluation result; whether the difference value between the comprehensive evaluation result and the comprehensive reference value is larger than or equal to a preset measurement threshold value is judged; if yes, performing streaming computation on the in-event stage data and the post-event stage data according to a user experience measurement system, and obtaining real-time evaluation results of the application program in an evaluation feedback dimension, a function definition dimension, a participation dimension and a performance and security dimension; and performing cross aggregation analysis on the real-time evaluation result and the data in the prior stage, and determining an abnormal reason of the application program of the mobile terminal. According to the method, real-time acquisition of full-link data can be realized, and based on the comprehensive evaluation result and the real-time evaluation result of each dimension, the abnormality reason is automatically positioned by backward deduction layer by layer, so that efficient abnormality solving is promoted.

Disclosed are implementations, including a method for interacting with a data storage that includes first data comprising data values for rendering in a renderable view and second data comprising specifications for determining the values in the first data. The method includes processing the data values of the first data to yield processed first data including semantic interpretation of the data values, processing the specifications of the second data to yield processed second data including functional definitions of the data values of the first data, accepting a natural language input, and processing the natural language input using the first processed data and the second processed data to respond to the natural language input. The data storage may include a spreadsheet, and the renderable view may include an array of cells of the spreadsheet in which at least some of the cells include renderings of respective data values.

Embodiments of the present application provide a kind of function control program's architecture method, device, storage medium and computer equipment.The method comprises: obtaining the function definition table of setting;Function definition table is input into parsing tool, and function configuration table and software development kit function file are generated;According to software development kit function file, the function interface of function control program is set;According to software development kit function file, the function / value interface parameter model of function control program is generated;According to function configuration table, the software development kit of function control program is generated.The technical scheme provided in the embodiments of the present application, function definition table is regarded as only input, unified function interface is designed based on parsing tool, can keep function interface uniform, consistent and scalable, improve the efficiency of application development.

The invention provides a static analysis and large language model-based open source software functionality measurement method, which comprises the following steps of: 1) receiving and analyzing a software source code, and constructing a function and class call atlas by virtue of a compiling tool; 2) based on the important attributes of the functions and the calling relation between the functions, generating a structured document for the functions by means of LLMs; 3) based on the important attributes of the classes, the incidence relation between the classes and functions and the calling relation between the classes, generating structured documents for the classes by means of LLMs; 4) based on the generated function document, generating a structured document by means of LLMs; 5) based on the generated function document and module document, generating a structured warehouse document by means of LLMs; 6) analyzing the difference between two pieces of software and providing software model selection suggestions by means of LLMs based on the extracted function definition, the generated module document and the warehouse document, and 7) answering questions of a user by means of LLMs based on the generated function document, class document, module document and warehouse document.

The application relates to the technical field of APP development, in particular to an industrial application software code testing method based on a large model, which comprises the following steps: receiving an industrial application software source file, parsing function definition statements and interface calling instructions in the file, extracting logical connection relationships between various business modules, constructing a function calling topological graph reflecting code unit directivity, and mapping code character sequences of each node in the function calling topological graph to a high-dimensional numerical coordinate system through a large model. In the application, conditional branching and parameter type reverse deduction are performed on a key path, constraint conditions are solved, and regression test case data capable of triggering a specific business scenario are automatically generated. In this way, the selection of a test focus, the discovery of a key path and the generation process of test data are automated and intelligentized, so that test resources can be accurately invested in function links with the highest value and the highest risk in software, and the test efficiency is improved.