Chip data processing method, device and equipment and readable storage medium

Abstract

The invention relates to the technical field of chip data processing, and discloses a chip data processing method, device and equipment and a readable storage medium, and the chip data processing method comprises the steps: obtaining a chip device data set associated with a target chip device, carrying out the analysis processing of the chip device data set, and obtaining analysis result data; determining target area information related to the target time sequence diagram according to the analysis result data, and generating corresponding time sequence diagram image data; performing image analysis processing on the time sequence diagram image data, and determining segmentation information; performing section division on the target time sequence diagram based on the segmentation information to obtain time sequence fragment association information; and inputting the time sequence fragment association information into a time sequence analysis model for reasoning analysis, and generating a time sequence analysis result. The test preparation efficiency is remarkably improved, the risk of man-made misjudgment is reduced, standardized data is provided for the test process, and upgrade of integrated circuit test is promoted.

Description

Technical Field

[0001] This application relates to the field of chip data processing technology, and in particular to a chip data processing method, apparatus, device and readable storage medium. Background Technology

[0002] Currently, in integrated circuit signal integrity testing, testers rely on datasheets provided by chip manufacturers for test preparation. These datasheets are typically presented as electronic documents such as PDFs, containing a wealth of text and graphics, especially timing diagrams that define signal timing relationships and describe key parameters such as read cycles, write cycles, address setup / hold times, and data output delays. In actual testing, engineers must manually consult the datasheets, locate the relevant timing diagrams, and, in conjunction with the actual waveforms captured by an oscilloscope, manually compare and identify the start and end positions of each logic cycle and their corresponding parameter values ​​to complete the test item configuration and result determination. This process, as a prerequisite for automated testing, is widely used in functional verification scenarios for high-speed digital circuits.

[0003] However, existing technologies heavily rely on manual intervention, resulting in significant efficiency bottlenecks and quality risks. On one hand, inconsistent formatting of datasheets from different manufacturers and diverse timing diagram layouts lead to lengthy manual searches and analyses, requiring an average of approximately two person-days per project for data extraction and entry. On the other hand, identifying graphical semantic information such as instruction cycles, data cycles, and boundary arrows is susceptible to subjective experience, leading to misjudgments or omissions, incorrect test parameter configurations, and consequently affecting the accuracy and repeatability of test results. Furthermore, with the increasing speed of chip interfaces and the growing complexity of protocols, the number and structure of timing diagrams have increased, making traditional manual methods insufficient to meet the demands of efficient and highly reliable testing. Summary of the Invention

[0004] In view of this, embodiments of this application provide a chip data processing method, apparatus, device, and readable storage medium, which can effectively solve the technical problems in the prior art, such as low test preparation efficiency, large human judgment error, and poor cross-vendor document format compatibility caused by relying on manual consultation of chip datasheets and manual identification of timing diagram logic cycles.

[0005] In a first aspect, embodiments of this application provide a chip data processing method, including: Obtain a set of chip device data associated with the target chip device, and parse the set of chip device data to obtain parsing result data; Based on the parsing results, target region information related to the target time series map is determined, and corresponding time series map image data is generated; The time series image data is subjected to image analysis processing to determine the segmentation information used to divide the time series regions; Based on the segmentation information, the target time sequence graph is divided into segments to obtain time sequence segment association information; The associated information of the time series segments is input into the time series analysis model for semantic reasoning analysis to generate a structured time series parsing result corresponding to the target time series diagram.

[0006] In some embodiments, the method further includes: The structured timing analysis results are stored to form a timing analysis record corresponding to the target chip device.

[0007] In some embodiments, obtaining a chip device data set associated with the target chip device and parsing the chip device data set to obtain parsing result data includes: Read the chip device data set associated with the target chip device to form the data to be parsed; Perform text parsing processing on each page in the data to be parsed to obtain page text data; The spatial coordinates of each text object are determined based on the page text data, and the page text data and the spatial coordinates are combined to form the parsing result data.

[0008] In some embodiments, determining the target region information related to the target time series map based on the parsing result data and generating the corresponding time series map image data includes: Based on preset text matching rules, target text objects related to the target time sequence diagram are identified in the parsed result data to obtain a set of target text objects; The corresponding page identifier is determined based on the target text object set, and the initial region information where the target time sequence diagram is located is determined by combining the spatial coordinate position in the parsing result data; Rendering is performed based on the initial region information to form the target region information, and time-series image data corresponding to the target region information is generated.

[0009] In some embodiments, the step of performing image analysis processing on the time series image data to determine segmentation information for dividing the time series regions includes: Convert the time series image data into grayscale image data; Threshold segmentation is performed on the grayscale image data to obtain binary image data; Edge detection processing is performed based on the binary image data to obtain edge image data; Extract a set of candidate contours from the edge image data; The candidate contour set is filtered according to preset area conditions to obtain the target contour set; The location information corresponding to the target contour set is determined as segmentation information for dividing the temporal region.

[0010] In some embodiments, the step of segmenting the target time series graph based on the segmentation information to obtain time series segment association information includes: Based on the segmentation information, candidate marker objects representing temporal boundaries are extracted, and the position information of each candidate marker object is obtained; All candidate marker objects are sorted according to the location information, and adjacent candidate marker objects are paired based on the sorting results to obtain multiple sets of temporal boundary pairs; Based on the position of each of the time-series boundaries in the target time-series diagram, the corresponding time-series segments are determined, and time-series segment association information is generated for each of the time-series segments.

[0011] In some embodiments, the step of inputting the temporal segment association information into a temporal analysis model for semantic reasoning analysis to generate a structured temporal parsing result corresponding to the target temporal graph includes: Based on the time-series segment association information, time-series text data and the identification information of each time-series segment are extracted to form time-series analysis data; The time series analysis data is combined with preset analysis constraint information to construct time series analysis request data; The time series analysis request data is input into the time series analysis model to obtain the time series analysis result data; Based on the time series analysis results, the data is organized according to a predetermined data structure to obtain a structured time series parsing result corresponding to the target time series diagram.

[0012] Secondly, embodiments of this application provide a chip data processing apparatus, comprising: The data parsing module is used to acquire a set of chip device data associated with the target chip device, and to parse the set of chip device data to obtain parsing result data. The data processing module is used to determine the target region information related to the target time series map based on the parsing result data, and generate the corresponding time series map image data; The image processing module is used to perform image analysis and processing on the time series image data to determine segmentation information for dividing the time series regions; The region segmentation module is used to segment the target time series map based on the segmentation information to obtain time series segment association information; The analysis and processing module is used to input the associated information of the time series segments into the time series analysis model for time series analysis and processing, so as to generate a structured time series parsing result corresponding to the target time series diagram.

[0013] Thirdly, embodiments of this application provide a terminal device, the terminal device including a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the chip data processing method of the first aspect described above.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium, wherein when the computer program is executed on a processor, it implements the chip data processing method of the first aspect described above.

[0015] The embodiments of this application have the following beneficial effects: by parsing and merging chip datasets and combining image analysis and semantic reasoning, end-to-end automated processing from unstructured documents to structured parsing results is achieved, effectively avoiding the low efficiency and error-prone problems caused by manually consulting manuals, recognizing waveforms, and extracting parameters; by performing region positioning, segmentation, and partitioning of timing diagrams, the accuracy and stability of key information extraction are improved; by using a timing analysis model to perform semantic understanding of the partitioning results and output standardized structured results, the level of intelligence in chip timing information processing is significantly improved, making it suitable for rapid configuration and reuse in high-frequency, multi-model testing scenarios, and greatly shortening the test preparation cycle. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of a chip data processing method according to an embodiment of this application is shown; Figure 2 Another flowchart of the chip data processing method according to an embodiment of this application is shown; Figure 3 This illustrates yet another flowchart of a chip data processing method according to an embodiment of this application; Figure 4 A schematic diagram of a chip data processing method according to an embodiment of this application is shown. Detailed Implementation

[0018] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0019] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] In the following text, the terms "comprising," "having," and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more combinations thereof. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance.

[0021] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0022] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features described herein can be combined with each other.

[0023] Considering the technical problems in existing technologies, such as low test preparation efficiency, large human judgment errors, and poor cross-vendor document format compatibility caused by relying on manual consultation of chip datasheets and manual identification of timing diagram logic cycles, a chip data processing method is proposed. This method performs image and text analysis on the chip device data set to locate and generate image data of the target timing diagram. Combined with image analysis, timing segmentation information is determined, and timing segments are divided based on this segmentation information. Finally, a timing analysis model is used to process the information of each segment to generate structured timing analysis results.

[0024] The following describes the data processing method for this chip using specific examples.

[0025] Figure 1 A flowchart of a chip data processing method according to an embodiment of this application is shown. Exemplarily, the chip data processing method includes the following steps: Step S100: Obtain the chip device data set associated with the target chip device, parse the chip device data set, and obtain the parsing result data.

[0026] Among them, the chip device data set refers to the technical documents provided by the chip manufacturer to describe its functional characteristics and timing behavior, usually in PDF format; parsing and processing refers to extracting the text content and its spatial distribution information on the page from the document to form structured intermediate data that can be used for subsequent analysis.

[0027] For example, the system receives the target chip datasheet file uploaded by the user and performs multimodal parsing operations on the file, including text layer reading and layout information extraction, identifying the text content and the position coordinates of the display area in each page, associating and storing semantic information and spatial information, and generating parsing result data containing both text and position attributes; this data serves as the basis for subsequent location of the target time sequence map region and realization of image-text mapping.

[0028] In an optional embodiment, step S100 includes the following sub-steps: S101, Read the chip device data set associated with the target chip device to form the data to be parsed.

[0029] Reading refers to loading electronic documents stored locally or remotely into the system's operating environment to construct accessible data objects; data to be parsed refers to document instances that have been loaded but are still in their original format, possessing the ability to access page indexes and content streams.

[0030] As an example, the datasheet of the target chip device is received through a file interface, and the document loading process is initiated. The file is then processed for format recognition and decapsulation to establish a document object structure in memory, enabling page-by-page access and content extraction. This process ensures that the overall structure of the document is preserved intact, avoiding content loss due to encoding differences or compression methods. For example, in actual processing, after detecting that the input file is in standard PDF format, decoding is completed and an internal representation with page indexes is generated for subsequent page-by-page traversal.

[0031] S102, perform text parsing processing on each page in the data to be parsed to obtain the page text data.

[0032] Text parsing refers to the process of extracting human-readable text content from each page of the data to be parsed, excluding non-text elements such as graphics and lines; page text data refers to a collection of text fragments organized by page, including titles, paragraphs, annotations, etc.

[0033] As an example, all pages in the data to be parsed are traversed, and character-level scanning and layout analysis are performed on each page to identify independent text blocks and extract their display content. The obtained text is arranged in reading order, and redundant line breaks and whitespace characters are removed to form standardized page text data. For example, when "Figure 7: Write Cycle Timing" or a similar expression is extracted on a certain page, it can be preliminarily determined that the page is the page to be processed. This process realizes the transformation from a layout document to searchable text.

[0034] S103, determine the spatial coordinates of each text object based on the page text data, and combine the page text data with the spatial coordinates to form the parsing result data.

[0035] Among them, spatial coordinates refer to the two-dimensional area occupied by each text object on the page, usually defined by two coordinate points, the upper left and lower right corners, reflecting its physical distribution; parsing result data refers to a data structure that integrates text content and its position information, supporting image and text linkage analysis.

[0036] As an example, while extracting the text data of the page, the system obtains the actual display area coordinates of each text block, records its boundary range based on the page coordinate system, and binds the coordinate information with the corresponding text content. Finally, the text and coordinate entries of all pages are summarized to form a unified parsing result data. For example, for the annotation item with the text content "tDH", the system records the coordinate range of the rectangular area where it is located, so that it can be accurately located and cropped in the subsequent image processing stage.

[0037] Step S200: Determine the target region information related to the target time series map based on the parsing result data, and generate the corresponding time series map image data.

[0038] Among them, target region information refers to the precise display range of the target time series map on the document page, which is usually represented by a rectangular bounding box; time series map image data refers to the high-resolution bitmap file generated based on the rendering of this region.

[0039] As an example, by utilizing the textual semantics and layout information in the parsed data, the pages that may contain the target time series diagram and their approximate location range are identified, and the image content of the corresponding area is extracted by combining coordinate information; by improving the rendering accuracy, a clear bitmap image is generated to ensure that the lines, arrows and annotation text in the image can be accurately identified.

[0040] In one alternative embodiment, such as Figure 2 As shown, step S200 includes the following sub-steps: S201, in the parsed result data, identify the target text objects related to the target time sequence diagram based on the preset text matching rules, and obtain the target text object set.

[0041] Among them, the text matching rule refers to the string or regular expression pattern used to identify the target time sequence diagram related descriptions, covering common title naming habits; the target text object refers to the text block that conforms to the rule, usually in the form of "TimingDiagram", "Time Sequence Diagram", "Read / Write Cycle", etc.

[0042] For example, all text entries in the parsed result data are traversed, and each text content is compared with a preset rule set to filter out text objects containing specific keywords or conforming to the naming structure. The rules include, but are not limited to, logical combinations such as "containing 'Timing' and adjacent to 'Diagram' or 'Cycle'" and "starting with 'Figure' and containing 'Read' or 'Write'". When a match is successful, the text object and its coordinate information are recorded to form a set of target text objects. For example, if the text content "Figure 6-3:SPI Write Timing Diagram" is identified on a certain page, it is determined to be a target text object and added to the set because it simultaneously meets the conditions of the prefix "Figure" and the keyword "Timing". This process serves as a trigger mechanism for image localization, effectively narrowing the search range.

[0043] S202, determine the corresponding page identifier based on the target text object set, and determine the initial region information corresponding to the target time series diagram by combining the spatial coordinate position in the parsing result data.

[0044] Among them, the page identifier refers to the page number where the target text object is located, which is used to locate the specific page; the initial area information refers to the approximate display range of the sequence diagram inferred based on the position of the target text object, which is usually a rectangular area.

[0045] For example, the page number to which each target text object belongs is extracted to determine the target page to be processed; then the coordinate position of the text object in the page is analyzed, and the actual coverage of the timing diagram is determined in combination with the distribution characteristics of the surrounding text blocks; for example, a rectangular area surrounding multiple related annotation items is delineated by extending a certain distance in the vertical direction and extending to the page margin or adjacent blank area in the horizontal direction; for example, when “Read Timing” is identified as being located in the upper middle position of the page, and waveform description text and signal names are continuously distributed below it, the area will be extended downwards to form an initial area to include as much of the complete waveform diagram as possible.

[0046] S203, based on the initial region information, perform rendering processing to form target region information, and generate time-series image data corresponding to the target region information.

[0047] Rendering refers to the process of converting a specified area in a PDF into a bitmap image; target area information is the precise cropping range after further adjustment of the initial area, supporting high-quality image output.

[0048] As an example, based on the initial area information, a magnification factor is set to re-render a specified page area; the magnification factor is set to 2.0, which is twice the original resolution, so that the image DPI reaches approximately 144, significantly improving the recognizability of small symbols (such as arrows and dashed lines); the original color mode is preserved during the rendering process, and the output is a PNG image with RGB channels; the resulting image contains only the content of the target area, with neat edges, which is convenient for subsequent processing; for example, after completing the magnified rendering, a clear image with a resolution of 1200×800 pixels is obtained, which fully presents the clock signal, data lines and left and right arrow markings, meeting the accuracy requirements of subsequent contour detection and symbol recognition.

[0049] Step S300: Perform image analysis processing on the time series image data to determine the segmentation information used to divide the time series regions.

[0050] Among them, segmentation information refers to the spatial location data that characterizes the main range of the time series diagram, serving as the basic boundary for the horizontal logical period division.

[0051] For example, the generated time series image data is received and subjected to a multi-stage image processing flow, which sequentially completes color space conversion, background noise suppression, edge feature extraction and contour filtering, and finally determines the target contour representing the complete time series image representation region; the bounding rectangle region corresponding to this contour constitutes the reference range for subsequent segmentation; for example, by detecting dense line areas and regular geometric structures in the image, non-waveform areas such as headers and annotation tables are excluded, and the central waveform area is accurately located, thereby obtaining stable and reliable segmentation information.

[0052] In an optional embodiment, step S300 includes the following sub-steps: S301 converts time series image data into grayscale image data.

[0053] Grayscale image data refers to an image representation that contains only brightness information and no color channels. It is usually described by single-channel pixel values ​​(0-255) to describe the brightness of each point. This conversion aims to simplify the computational burden of subsequent processing and highlight the structural features of the graphics.

[0054] As an example, a color space transformation is performed on the input color or RGB format timing image, converting the three-channel pixel values ​​into a single grayscale value using a weighted average algorithm to generate the corresponding grayscale image data. This process preserves the shape information of all lines, text, and arrows in the original image while eliminating interference caused by color differences. For example, for waveforms that originally used different colors to represent clock and data signals, the transformation can still clearly show their transition edges and logic state changes, meeting the requirements of subsequent binarization and edge detection.

[0055] S302, perform threshold segmentation on the grayscale image data to obtain binary image data.

[0056] Threshold segmentation refers to the operation of dividing a grayscale image into foreground and background parts according to a set brightness threshold; binary image data refers to an image that contains only black and white pixels (usually 0 and 255), which is convenient for subsequent edge and contour extraction.

[0057] As an example, a fixed global threshold is used to binarize the grayscale image, setting areas above the threshold as background (white) and areas below the threshold as foreground (black), thereby highlighting key graphic elements such as wavy lines, arrows, and bounding boxes. This processing effectively removes gradient shadows and slight noise in the image, enhancing structural clarity. For example, setting the threshold to 128 ensures that all dark lines in the original image are fully preserved, while light gray backgrounds or scan marks are removed, generating a clean black and white image for subsequent analysis.

[0058] S303 performs edge detection processing based on binary image data to obtain edge image data.

[0059] Edge detection processing refers to methods for identifying locations of abrupt changes in brightness in an image, used to locate object boundaries. Edge image data refers to image results that retain only edge pixels, reflecting the outline and skeleton of the original graphic.

[0060] As an example, the Canny edge detection algorithm is executed on a binary image. Through steps such as gradient calculation, non-maximum suppression, and double threshold connection, continuous and accurate edge lines are extracted. This process can effectively identify the contour paths of the outer frame, signal transition edges, and arrow boundaries in the waveform image. For example, after processing, the rectangular timing frame, horizontal data line, and triangular arrow in the original image are all restored to clear edge segments, forming a complete graphic structure representation.

[0061] S304, Extract a set of candidate contours from edge image data.

[0062] The candidate contour set refers to the set of closed or multi-segment curved paths formed by edge connections, with each contour corresponding to a potential independent graphic object in the image; the extraction process aims to identify all possible graphic region boundaries.

[0063] As an example, all non-zero pixels in the edge image data are traversed, and a contour tracking algorithm is used to aggregate connected edge points into continuous paths, and the sequence of boundary points of the region enclosed by each path is recorded. The results include the original contour data of various graphics such as time series map outlines, internal grid lines, text borders and arrow contours. For example, multiple rectangular structures are identified, including the large rectangle of the main waveform area, as well as small icons or annotation boxes, all of which are uniformly included in the candidate contour set for filtering.

[0064] S305: Filter the candidate contour set according to the preset area conditions to obtain the target contour set.

[0065] Among them, the area condition refers to the quantitative standard used to filter invalid contours, which is usually set as the lower limit of pixel area; the target contour set refers to the effective contour group that meets the area requirement and is most likely to represent the main structure of the time series diagram.

[0066] As an example, the pixel area enclosed by each candidate contour is calculated, and contours with an area smaller than a set threshold are discarded. The threshold is set to 8000 pixels², which is statistically derived from a large number of measured samples and is sufficient to exclude text boxes, small icons, and noise interference, while retaining the complete waveform region contour. Among the remaining contours, the one with the largest area or the one in the middle position is selected as the target contour representing the main body of the time series map. For example, when a contour with an area of ​​15000 pixels² is detected and located in the central region of the image, it is determined to be the outer frame of the main time series map and added to the target contour set; while small rectangles with an area of ​​only a few hundred pixels are regarded as annotation boxes and discarded. This filtering mechanism effectively improves the robustness and consistency of region localization.

[0067] S306, The positional information corresponding to the target contour set is determined as the segmentation information for dividing the temporal region.

[0068] Among them, positional information refers to the spatial distribution parameters of the target contour in the image coordinate system, which is usually represented by the coordinates of the upper left and lower right corners of its circumscribed rectangle.

[0069] As an example, for each contour retained in the target contour set, its minimum bounding rectangle is calculated, and the boundary coordinates of the rectangle are obtained. This bounding rectangle is then used as the effective analysis area for the time series diagram. This area covers the entire waveform display range, excluding margins and irrelevant annotations. For example, in a screenshot, the bounding rectangle range of the main contour is determined to be (180, 210, 960, 650), and this range is used as the operation window for subsequent arrow recognition and segmentation, ensuring that all analysis is focused on the core area.

[0070] Step S400: Divide the target time series graph into segments based on the segmentation information to obtain time series segment association information.

[0071] Among them, the time sequence segment association information refers to the structured data generated after dividing the complete time sequence diagram into multiple logical cycles, which includes the correspondence between the position, content and functional semantics of each segment.

[0072] Exemplarily, the image region defined by the segmented information is used as the operating range. Graphical markers representing the start and end of time within this region are identified and located, and paired according to their spatial arrangement. This divides the horizontal time axis into several continuous logical periodic segments. Each segment corresponds to a specific operation stage, such as instruction sending, address latching, or data output. For example, after detecting multiple pairs of left and right arrows, they are used as periodic boundary markers to sequentially segment time segments such as "CS# valid interval", "address transmission segment", and "data reading segment". The coordinate position and internal text content of each segment are recorded to form an analysis unit with associated attributes.

[0073] In one alternative embodiment, such as Figure 3 As shown, step S400 includes the following sub-steps: S401, extract candidate labeled objects representing temporal boundaries based on segmentation information, and obtain the position information of each candidate labeled object.

[0074] Among them, candidate marker objects refer to specific graphic symbols in the image used to identify the start and end positions of the timing logic, usually represented by left or right arrows, used to indicate the start or end of a certain operation cycle; position information refers to the center point or bounding box coordinates of the symbol in the image coordinate system.

[0075] Demonstratively, a dedicated image recognition operation is performed within the target area defined by segmentation information. A trained object detection model is used to scan and recognize arrow-like symbols in the image. The model can distinguish between left and right arrows and output the category label and the coordinates of the bounding rectangle of each detected candidate object. The center x-coordinate of each arrow is further calculated as a reference for its relative position on the time axis. For example, in an SPI timing diagram, four arrows are identified: two left arrows are located in the beginning and middle of the waveform, and two right arrows are distributed accordingly. All positions are accurately recorded.

[0076] S402, sort all candidate marker objects according to their location information, and pair adjacent candidate marker objects based on the sorting results to obtain multiple sets of temporal boundary pairs.

[0077] Here, sorting refers to the process of arranging candidate marker objects from smallest to largest according to their horizontal coordinate values; a time boundary pair is a combination consisting of a left-hand arrow and a right-hand arrow, used to define a complete logical cycle interval.

[0078] As an example, all identified candidate marker objects are sorted in ascending order according to the size of their center x-coordinates; then the sorted sequence is traversed, and the first left arrow is paired with the first right arrow to form the first time boundary pair, then the second left arrow is paired with the second right arrow, and so on; if the number of left and right arrows is inconsistent, the one with the fewer arrows is used for truncation matching.

[0079] S403, determine the corresponding time segment based on the position of each time boundary pair in the target time sequence diagram, and generate time segment association information for each time segment.

[0080] Among them, the time segment refers to the time interval enclosed by a set of time boundary pairs, which is represented as a horizontal strip region on the image.

[0081] As an example, using the left and right horizontal coordinates of each time series boundary pair as boundaries, the horizontal interval of the image between the two is extracted as the current time series segment. For parts without complete boundaries at the beginning and end, they are extended and completed by either the starting point of the previous segment or the ending point of the next segment. Then, the text content contained in the segment, such as annotations like "CLK", "DIN", and "tSU", is searched within the segment and bound to the segment position. Finally, a structured record containing "segment coordinate range + internal text + sequential relationship" is generated. For example, the text "Data In" is extracted between the second pair of arrows, the system classifies it into the segment, and infers that this segment is a data input cycle, forming a complete time series segment association information.

[0082] Step S500: Input the temporal segment association information into the temporal analysis model for semantic reasoning analysis to generate a structured temporal parsing result corresponding to the target temporal diagram.

[0083] Among them, the time series analysis model refers to a semantic analysis system with natural language understanding and contextual reasoning capabilities, used to identify the functional meaning of each time series segment and establish parameter associations; the structured time series parsing result refers to the output data organized in a predetermined format, including the functional definition of each period, key time parameters and signal behavior description.

[0084] As an example, multiple timing segments are sequentially input into a timing analysis model according to their order on the timeline. Based on preset professional knowledge rules and contextual understanding capabilities, the model comprehensively judges the signal state changes, control instruction encoding, and clock dependencies in each segment, derives its corresponding operational stage function, such as "address latch" and "data output enable," and extracts or calculates relevant timing parameters. Finally, all analysis conclusions are integrated to form a unified structured output. For example, when processing SPI read operation timing, the model identifies the synchronous transition phenomenon of "CLK" and "DO" in the third segment, and combined with the "CMD=03h" identifier of the previous segment, determines that this segment is a valid data output cycle, starting from the fourth clock rising edge.

[0085] In an optional embodiment, step S500 includes the following sub-steps: S501, extract time-series text data and identification information of each time-series segment based on the time-series segment association information to form time-series analysis data.

[0086] Among them, time-series text data refers to the text annotation content existing in each time-series segment, including signal name, clock number, operation code or time parameter symbol; identification information refers to the unique number or location label used to distinguish different segments; time-series analysis data refers to the data set to be processed by combining the above two types of information.

[0087] Exemplary approach: traverse the association information of each time series segment, extract all the text content contained therein, and arrange them according to their horizontal order of appearance; simultaneously assign each segment an incrementing sequence number as its identifier to indicate its relative position in the overall time series; then bind the text sequence with the sequence number to form a complete time series analysis data entry; for example, for the second time series segment, the extracted text content is “tDH”, “DIN”, “CLK=1”, and it is assigned the identifier “Segment_2”, ultimately forming an analysis data entry with the structure {“id”:“Segment_2”,“texts”:[“tDH”,“DIN”,“CLK=1”]}; this data retains the original semantic and spatial order features, providing a contextual basis for subsequent model understanding.

[0088] S502 combines time series analysis data with preset analysis constraint information to construct time series analysis request data.

[0089] Among them, analysis constraint information refers to the professional domain restrictions used to guide the model to correctly parse, which usually include role settings, output format requirements and rules to prohibit speculation; time series analysis request data refers to the encapsulated complete input request.

[0090] As an example, the prepared time-series analysis data is concatenated with a set of fixed prompts to construct a composite request body containing task instructions, input content, and format requirements. The constraints explicitly instruct the model to perform analysis as a "professional data manual analysis expert," deriving only from the provided content and not introducing any unmentioned technical assumptions. The output must also conform to the JSON format specification, including fields such as "functional description," "starting period," and "derivation process." For example, the constructed request data includes the following logic: "Based on the text content of the following sections, please outline the complete time-series logic flow and determine from which CLK period the DataOut signal becomes effective." This request structure ensures the consistency and controllability of the model output.

[0091] S503: Input the time series analysis request data into the time series analysis model to obtain the time series analysis result data.

[0092] The time series analysis results data refer to the natural language or semi-structured responses returned by the model, which include functional judgments and parameter derivation conclusions for each time series segment.

[0093] As an example, the constructed timing analysis request data is sent to the remotely deployed timing analysis model via the API interface, and the model waits for it to complete the inference and return a response. The model relies on its built-in knowledge system and prompting engineering mechanism to accurately identify the common patterns in the chip communication protocol and complete logical judgments such as "the first rising edge of CLK after CS# is pulled low to start the transmission". The returned results include detailed derivation explanations and final conclusions.

[0094] S504: Based on the time series analysis results, the data is organized according to a predetermined data structure to obtain a structured time series parsing result corresponding to the target time series diagram.

[0095] The data structure refers to a predefined standard output format, usually JSON or other machine-readable format, which includes fields such as function categories, key parameters, and periodic indexes.

[0096] For example, the system receives the response content from the time series analysis model, performs syntax validation and field extraction, and converts the unstructured natural language description into a standardized key-value pair format; for example, it extracts {"data_start_cycle":3,"reason":"DO and CLK synchronization was detected between the second arrow pair"} from "DataOut starts outputting from the 3rd CLK"; all parsing results are categorized and summarized by section to form a complete structured record.

[0097] In an optional embodiment, the chip data processing method further includes the following steps: The structured timing analysis results are stored to form a timing analysis record corresponding to the target chip device.

[0098] The timing analysis record refers to the data entry that persistently saves the result and is associated with a specific chip device. For example, the generated structured timing analysis result is written to the database management system, creating a new timing analysis record. This record is associated with at least the target chip's name, manufacturer information, signal type, and the corresponding datasheet source, and the analysis content is stored in JSON or similar structured fields. The database table structure is designed to support fast retrieval by chip model and signal type, ensuring that the automated testing process can accurately obtain the required parameters. For instance, after completing the timing analysis of a certain SPI Flash chip, the system stores the output result along with the chip model "W25Q128JV", manufacturer "Winbond", and signal type "READ_TIMING" into the database, forming a complete and traceable record. When subsequent test tasks process the same chip again, this record can be directly read, avoiding repeated analysis and improving overall testing efficiency.

[0099] Figure 4A schematic diagram of a chip data processing apparatus according to an embodiment of this application is shown. Exemplarily, the apparatus 100 includes: The data parsing module 110 is used to acquire a chip device data set associated with the target chip device, and to parse the chip device data set to obtain parsing result data; The data processing module 120 is used to determine the target region information related to the target time series map based on the parsing result data, and generate the corresponding time series map image data; Image processing module 130 is used to perform image analysis processing on the time series image data to determine segmentation information for dividing the time series regions; The region segmentation module 140 is used to segment the target time sequence map based on the segmentation information to obtain time sequence segment association information; The analysis and processing module 150 is used to input the time series segment association information into the time series analysis model for time series analysis and processing, so as to generate a structured time series parsing result corresponding to the target time series diagram.

[0100] It is understood that the apparatus of this embodiment corresponds to the method of the above embodiments, and the options in the above embodiments are also applicable to this embodiment, so they will not be described again here.

[0101] This application also provides a terminal device, exemplary of which includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the terminal device to perform the functions of the various modules in the above-described method or apparatus.

[0102] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including at least one of a Central Processing Unit (CPU), Graphics Processing Unit (GPU), Network Processor (NP), Digital Signal Processor (DSP), Application-Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application.

[0103] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory is used to store computer programs, and the processor can execute the computer programs accordingly after receiving execution instructions.

[0104] This application also provides a computer-readable storage medium for storing the computer program used in the aforementioned terminal device. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0105] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0106] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0107] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application.

[0108] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A chip data processing method, characterized in that, The method includes: Obtain a set of chip device data associated with the target chip device, and parse the set of chip device data to obtain parsing result data; Based on the parsing results, target region information related to the target time series map is determined, and corresponding time series map image data is generated; The time series image data is subjected to image analysis processing to determine the segmentation information used to divide the time series regions; Based on the segmentation information, the target time series graph is divided into segments to obtain time series segment association information; The associated information of the time series segments is input into the time series analysis model for semantic reasoning analysis to generate a structured time series parsing result corresponding to the target time series diagram.

2. The chip data processing method according to claim 1, characterized in that, The method further includes: The structured timing analysis results are stored to form a timing analysis record corresponding to the target chip device.

3. The chip data processing method according to claim 1, characterized in that, The process of acquiring a chip device data set associated with the target chip device, parsing the chip device data set to obtain parsing result data includes: Read the chip device data set associated with the target chip device to form the data to be parsed; Perform text parsing processing on each page in the data to be parsed to obtain page text data; The spatial coordinates of each text object are determined based on the page text data, and the page text data and the spatial coordinates are combined to form the parsing result data.

4. The chip data processing method according to claim 1, characterized in that, The step of determining the target region information related to the target time series map based on the parsing result data and generating the corresponding time series map image data includes: Based on preset text matching rules, target text objects related to the target time sequence diagram are identified in the parsed result data to obtain a set of target text objects; The corresponding page identifier is determined based on the target text object set, and the initial region information where the target time sequence diagram is located is determined by combining the spatial coordinate position in the parsing result data; Rendering is performed based on the initial region information to form the target region information, and time-series image data corresponding to the target region information is generated.

5. The chip data processing method according to claim 1, characterized in that, The step of performing image analysis processing on the time series image data to determine segmentation information for dividing the time series regions includes: Convert the time series image data into grayscale image data; Threshold segmentation is performed on the grayscale image data to obtain binary image data; Edge detection processing is performed based on the binary image data to obtain edge image data; Extract a set of candidate contours from the edge image data; The candidate contour set is filtered according to preset area conditions to obtain the target contour set; The location information corresponding to the target contour set is determined as segmentation information for dividing the temporal region.

6. The chip data processing method according to claim 1, characterized in that, The step of dividing the target time series graph into segments based on the segmentation information to obtain time series segment association information includes: Based on the segmentation information, candidate marker objects representing temporal boundaries are extracted, and the position information of each candidate marker object is obtained; All candidate marker objects are sorted according to the location information, and adjacent candidate marker objects are paired based on the sorting results to obtain multiple sets of temporal boundary pairs; Based on the position of each of the time-series boundaries in the target time-series diagram, the corresponding time-series segments are determined, and time-series segment association information is generated for each of the time-series segments.

7. The chip data processing method according to claim 1, characterized in that, The step of inputting the temporal segment association information into a temporal analysis model for semantic reasoning analysis to generate a structured temporal parsing result corresponding to the target temporal graph includes: Based on the time-series segment association information, time-series text data and the identification information of each time-series segment are extracted to form time-series analysis data; The time series analysis data is combined with preset analysis constraint information to construct time series analysis request data; The time series analysis request data is input into the time series analysis model to obtain the time series analysis result data; Based on the time series analysis results, the data is organized according to a predetermined data structure to obtain a structured time series parsing result corresponding to the target time series diagram.

8. A chip data processing device, characterized in that, include: The data parsing module is used to acquire a set of chip device data associated with the target chip device, and to parse the set of chip device data to obtain parsing result data. The data processing module is used to determine the target region information related to the target time series map based on the parsing result data, and generate the corresponding time series map image data; The image processing module is used to perform image analysis and processing on the time series image data to determine segmentation information for dividing the time series regions; The region segmentation module is used to segment the target time series map based on the segmentation information to obtain time series segment association information; The analysis and processing module is used to input the associated information of the time series segments into the time series analysis model for time series analysis and processing, so as to generate a structured time series parsing result corresponding to the target time series diagram.

9. A terminal device, characterized in that, The terminal device includes a processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the chip data processing method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed on a processor, implements the chip data processing method according to any one of claims 1-7.

Images