Intelligent file management system based on RPA

Abstract

The present invention relates to the field of archive management technology, and in particular to an intelligent archive management system based on RPA, the system comprising: a storage module identifies an invoice structure and performs distributed storage; a frequency setting module collects first node performance data to determine an initial detection frequency; a three-type node analysis module detects whether the three-type storage node is abnormal, adjusts the initial detection frequency according to the abnormal judgment result, or generates a second-type storage node abnormal judgment instruction; a second-type node analysis module determines whether the second-type storage node is abnormal based on the node abnormality ratio or node abnormality ratio change trend of a number of three-type storage nodes, so as to formulate a second-type node repair strategy according to the comparison result of the number of second-type abnormalities and the preset number of abnormalities, or generates a first-type storage node abnormal judgment instruction; a first-type node analysis module analyzes the abnormal distribution of a number of abnormal second-type storage nodes, so as to determine the first-type node repair strategy based on the degree of abnormality. The present invention improves the efficiency and security of archive management.

Description

Technical Field

[0001] The present invention relates to the field of file management technology, and in particular to an intelligent file management system based on RPA.

[0002] Background technology

[0003] RPA (Robotic Process Automation) is a software automation technology that simulates the interaction between humans and computers to perform repetitive tasks based on clear rules. It usually runs on office workers' computers in the form of "software" and can automatically operate various software in the computer, such as opening websites, clicking mice, keyboard input, etc., thereby realizing the automation of business processes.

[0004] The patent document with Chinese patent application publication number CN115511588A discloses an electronic accounting file archiving management system, including an electronic invoice management module for uploading and identifying electronic invoices; an invoice reimbursement module for filling out corresponding reimbursement forms based on the identified electronic invoices and generating corresponding original vouchers; a financial module for performing account entry processing based on the corresponding original vouchers and generating corresponding accounting vouchers; and an archiving management module for establishing an index relationship between the generated accounting vouchers and the corresponding original vouchers, and archiving and storing the accounting vouchers and the original vouchers associated therewith.

[0005] Although the existing electronic accounting archive management system can handle the uploading and identification of electronic invoices, it lacks real-time monitoring and analysis of the security of storage nodes and cannot effectively deal with abnormal situations that occur in storage nodes, resulting in low security of archive management data. Summary of the invention

[0006] To this end, the present invention provides an intelligent archive management system based on RPA, which can solve the problem of low security of archive management data by organizing the storage of invoice data, monitoring the performance of storage nodes, judging abnormalities of storage nodes at different levels, and formulating repair strategies.

[0007] To achieve the above object, the present invention provides an intelligent file management system based on RPA, which includes:

[0008] A storage module is used to identify the invoice structure corresponding to the invoices to be stored, identify the first keywords and the second keywords based on the invoice structure, and perform distributed storage on the invoices to be stored based on the invoice structure, the first keywords and the second keywords to obtain a number of first-class storage nodes, a number of second-class storage nodes and a number of third-class storage nodes;

[0009] A frequency setting module, connected to the storage module, for collecting first node performance data corresponding to a plurality of first-class storage nodes, and determining an initial detection frequency based on the first storage node performance data;

[0010] The three-type node analysis module is connected to the frequency setting module, and is used to detect the node safety anomaly corresponding to any of the three types of storage nodes in real time based on the initial detection frequency, determine whether the three types of storage nodes are abnormal based on the node safety anomaly, and adjust the initial detection frequency according to the abnormal judgment result of the three types of storage nodes, or generate a second type of storage node abnormal judgment instruction;

[0011] A second-category node analysis module is connected to the third-category node analysis module, and is used to receive a second-category storage node abnormality judgment instruction, and judge whether the second-category storage node is abnormal based on the node abnormality ratio or node abnormality ratio change trend of a number of third-category storage nodes corresponding to any of the second-category storage nodes, so as to formulate a second-category node repair strategy according to a comparison result of the number of second-category abnormalities of a number of abnormal second-category storage nodes with a preset number of abnormalities, or generate a first-category storage node abnormality judgment instruction;

[0012] A first-class node analysis module is connected to the second-class node analysis module, and is used to receive a first-class storage node abnormality judgment instruction, analyze the abnormal distribution of a number of abnormal second-class storage nodes, so as to judge the abnormality degree of a number of first-class storage nodes, and determine the first-class node repair strategy according to the abnormality degree judgment result;

[0013] Among them, any of the first-category storage nodes includes at least one second-category storage node, and any of the second-category storage nodes includes at least one third-category storage node.

[0014] Furthermore, the storage module includes:

[0015] A structure division unit, for identifying a plurality of edge contour regions corresponding to any of the invoices to be stored based on an edge detection algorithm, determining a plurality of actual area areas corresponding to the plurality of edge contour regions, and determining actual position distributions of the plurality of edge contour regions in the invoices to be stored, and using the similarities between the plurality of actual position distributions and the plurality of actual area areas corresponding to the plurality of invoices to be stored as a primary division standard to perform division, and obtain a primary division result;

[0016] A first division unit, connected to the structure division unit, is used to determine the invoice header area based on the actual location distribution corresponding to any of the invoices to be stored, identify the first keyword in the invoice header area, and use the similarity between the first keywords corresponding to the invoices to be stored as a secondary division standard to perform division, and obtain a secondary division result;

[0017] A second division unit, connected to the first division unit, is used to determine the invoice text area based on the actual location distribution corresponding to any of the invoices to be stored, identify a plurality of second keywords in the invoice text area, and divide the invoices using the similarity between the plurality of second keywords corresponding to the plurality of invoices to be stored as a third division standard to obtain a third division result;

[0018] The storage unit is connected to the structure division unit, the first division unit and the second division unit, and is used for distributing and storing a plurality of invoices to be stored based on the first division result, the second division result and the third division result respectively.

[0019] Furthermore, the frequency setting module includes:

[0020] A performance collection unit, used to collect the real-time CPU usage and real-time memory usage corresponding to any one of the storage nodes;

[0021] A frequency calculation unit is connected to the performance collection unit, and is used to calculate the real-time performance index of any one type of storage node based on the real-time CPU usage and the real-time memory usage, and determine the initial detection frequency based on the real-time performance index.

[0022] Furthermore, the frequency calculation unit includes:

[0023] An index calculation subunit, used for calculating a weighted average of the real-time CPU usage and the real-time memory usage to generate the real-time performance index;

[0024] The frequency determination subunit is connected to the index calculation subunit and is used to compare any of the real-time performance indexes with the preset performance index. When the real-time performance index is greater than the preset performance index, the preset detection frequency is shortened to obtain the initial detection frequency. When the real-time performance index is less than or equal to the preset performance index, the preset detection frequency is used as the initial detection frequency.

[0025] Furthermore, the three types of node analysis modules include:

[0026] A data monitoring unit, used for detecting the actual completeness and actual accuracy of a number of stored invoices in any of the three types of storage nodes in real time based on the initial detection frequency;

[0027] a safety determination unit, connected to the data monitoring unit, for determining an actual node safety degree based on the actual integrity and the actual accuracy;

[0028] The third type of node judgment unit is connected to the safety determination unit, and is used to compare the actual node safety with the preset node safety, and adjust the initial detection frequency based on the comparison result, or further judge the abnormal situation of the second type of storage node.

[0029] Furthermore, the data monitoring unit includes:

[0030] A completeness determination subunit is used to compare the actual memory proportions of the stored invoices with the preset memory proportions to obtain a proportion comparison result, and determine the actual completeness based on the proportion comparison result;

[0031] The accuracy determination subunit is used to compare the stored text information on any of the stored invoices with the corresponding actual text information to obtain an information comparison result, and determine the actual accuracy based on a number of information comparison results.

[0032] Furthermore, the second type of node analysis module includes:

[0033] A proportion calculation unit, used to count the total number of nodes and the number of three types of abnormalities of a number of third-type storage nodes corresponding to any of the second-type storage nodes, and calculate the node abnormality proportion based on the number of three types of abnormalities and the total number of nodes;

[0034] A proportion determination unit, connected to the proportion calculation unit, for determining the abnormal situation of the second type of storage nodes based on the comparison result between the node abnormality proportion and the preset abnormality proportion, or analyzing the change trend of the node abnormality proportion;

[0035] A trend judgment unit, connected to the proportion judgment unit, is used to analyze the change trend of the abnormal proportion of several nodes within the preset collection frequency, so as to determine the abnormal situation of the second type of storage node according to the change trend judgment result;

[0036] The first strategy determination unit is used to formulate a second-category node repair strategy according to a comparison result between the second-category abnormality quantity of a plurality of abnormal second-category storage nodes and a preset abnormality quantity, or to generate a first-category storage node abnormality judgment instruction.

[0037] Furthermore, the first strategy determination unit includes:

[0038] an abnormality locating subunit, used to identify a number of abnormal category II storage nodes, and locate a number of abnormal category III storage nodes corresponding to any abnormal category II storage node;

[0039] A strategy formulation unit is connected to the abnormal location subunit, and is used to calculate the safety difference between the actual node safety and the preset node safety corresponding to any of the three abnormal storage nodes, compare the safety difference with the preset difference, and determine the second-category node repair strategy based on the difference comparison result.

[0040] Furthermore, the first type of node analysis module includes:

[0041] A degree determination unit, used to identify a number of abnormal second-class storage nodes, determine abnormal distribution density and abnormal correlation corresponding to the number of abnormal second-class storage nodes, and determine the abnormal degree based on the abnormal distribution density and abnormal correlation;

[0042] The second strategy determination unit is connected to the degree determination unit and is used to determine a type of node repair strategy according to a comparison result between the abnormality degree and a preset degree.

[0043] Furthermore, the degree determination unit includes:

[0044] A density calculation subunit, used to calculate the distribution density of a number of abnormal type II storage nodes in any abnormal type I storage node;

[0045] A correlation analysis subunit is used to extract a number of second keywords corresponding to a number of abnormal third-category storage nodes in a number of abnormal second-category storage nodes in any of the abnormal first-category storage nodes as a number of abnormal keywords, and determine the abnormal correlation between the number of abnormal keywords;

[0046] The degree determination subunit is connected to the density calculation subunit and the correlation analysis subunit, and is used to determine the abnormality degree based on the distribution density and the abnormality correlation.

[0047] Compared with the prior art, the beneficial effect of the present invention is that, by setting the storage module to identify the invoice structure corresponding to the invoice to be stored, and further identifying the first keyword and the second keyword based on the invoice structure, a clear framework is provided for subsequent storage, query and management, and the efficiency of data management is improved. Based on the distributed storage method, the invoice is reasonably distributed to different storage nodes according to different characteristics, so that the data storage is more balanced, and the problem of excessive storage pressure caused by a large amount of data concentrated in a certain place is avoided. By setting the frequency setting module, the monitoring frequency is dynamically adjusted according to the actual performance status of the storage node. When the node performance tends to be tense, the monitoring frequency is timely increased so as to pay closer attention to the status of the storage node, discover the problems in advance, and ensure the security and stability of data storage. By setting three types of node analysis modules to detect the node security anomalies corresponding to the three types of storage nodes, it is helpful to quickly locate the storage node level where the problem is located, and provide a clear direction for subsequent further investigation and processing. By adjusting the initial detection frequency, the subsequent situation of the relevant storage node can be paid closer attention after the anomaly is found. By generating the second type of storage node abnormality judgment instruction, the impact range of the abnormal situation is further transmitted and analyzed to the upper node, so as to achieve The system realizes a gradual investigation and response mechanism from local to upper level, which ensures the stable operation of the entire archive management system. By setting up a second-class node analysis module to comprehensively consider the judgment method of multiple third-class storage nodes, it can more comprehensively and accurately evaluate the overall status of the second-class storage nodes. According to the comparison result of the second-class anomaly number of several abnormal second-class storage nodes with the preset anomaly number, a second-class node repair strategy is formulated, which can take different response measures for different degrees of abnormal situations, realizes the refined management of the abnormal situation of the second-class storage nodes, improves the maintainability and data security of the system, and generates a first-class storage node abnormality judgment instruction. The abnormal conditions of the Class II storage nodes are further transmitted to the Class I storage node level for analysis, so that the entire system can check and handle the abnormal conditions layer by layer from bottom to top, ensuring that any potential problems can receive adequate attention and resolution. By setting up a Class I node analysis module to analyze the abnormal distribution of several abnormal Class II storage nodes, the abnormal degree of several Class I storage nodes can be judged. Based on the evaluation method of the abnormal distribution of the lower-level nodes, a more comprehensive understanding of the scope and degree of impact that a Class I storage node may be affected can be achieved, which provides an important basis for the subsequent formulation of a reasonable Class I node repair strategy and improves the efficiency of archive management and data security.

[0048] In particular, by setting up a structural division unit, by determining the actual area corresponding to the edge contour area and the actual position distribution in the invoice to be stored, the important features of the invoice in terms of layout and space occupancy are extracted, reflecting the similarities and differences between different invoices in terms of format, typesetting, etc., providing a rich basis for subsequent division based on similarity, and facilitating a more detailed and accurate classification of invoices. By setting up a first division unit to accurately determine the invoice header area and identify the first keyword for division, the classification of invoices is more in line with their business attributes and important features, improving the pertinence and accuracy of the classification. By setting up a second division unit to determine the invoice body area and identify several second keywords for division, invoices with specific business content can be classified. Similar invoices are further subdivided, making the classification of invoices more refined and improving the flexibility and practicality of data management. By setting up storage units to distribute the storage of several invoices to be stored, the characteristics and classification results of different levels of invoices extracted by the previous division units are fully utilized, and invoices with different similarities and business attributes are reasonably allocated to different storage nodes, avoiding the problems of excessive storage pressure and low query efficiency caused by the centralized storage of a large number of invoices. When querying a specific type of invoice, the relevant data can be quickly located according to the storage node category to which it belongs, improving the speed and accuracy of data retrieval, and also facilitating the targeted management and maintenance of storage nodes of different categories, thereby improving the operation efficiency and data management efficiency of the entire archive management system.

[0049] In particular, by setting up a data monitoring unit to detect the actual completeness and actual accuracy of several stored invoices in any of the three types of storage nodes in real time, it is helpful to promptly discover problems in the storage nodes and prevent data loss or damage. By setting up a security determination unit to determine the actual node security based on the actual completeness and actual accuracy, it is helpful to comprehensively quantify the current security status of the storage node, making the judgment on the node security more accurate and reliable, and taking corresponding measures as needed to ensure the security and stability of data storage. By setting up three types of node judgment units to adjust the initial detection frequency based on the comparison result of the actual node security and the preset node security, the monitoring frequency is dynamically adjusted according to the actual situation of the storage node, thereby making more efficient use of resources. It is also used to judge the abnormal situation of the second type of storage node, which helps to further transmit and analyze the impact range of the abnormal situation to the upper-level node, realizing a gradual investigation and response mechanism from the local to the upper level, and ensuring the stable operation of the entire archive management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A structural diagram of an RPA-based intelligent archive management system provided in an embodiment of the present invention;

[0051] Figure 2A structural block diagram of a storage module in an RPA-based intelligent archive management system provided in an embodiment of the present invention;

[0052] Figure 3 A structural diagram of three types of node analysis modules in the RPA-based intelligent archive management system provided in an embodiment of the present invention;

[0053] Figure 4 A structural block diagram of the second type of node analysis module in the RPA-based intelligent archive management system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0056] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be understood as a limitation on the present invention.

[0057] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] See also Figure 1 As shown, an embodiment of the present invention provides an intelligent file management system based on RPA, which includes:

[0059] The storage module 10 is used to identify the invoice structure corresponding to the invoices to be stored, identify the first keywords and the second keywords based on the invoice structure, and perform distributed storage on the invoices to be stored based on the invoice structure, the first keywords and the second keywords to obtain a first-class storage node, a second-class storage node and a third-class storage node;

[0060] A frequency setting module 20, connected to the storage module 10, for collecting first node performance data corresponding to a plurality of first-class storage nodes, and determining an initial detection frequency based on the first storage node performance data;

[0061] The three-type node analysis module 30 is connected to the frequency setting module 20, and is used to detect the node safety anomaly corresponding to any of the three types of storage nodes in real time based on the initial detection frequency, determine whether the three types of storage nodes are abnormal based on the node safety anomaly, adjust the initial detection frequency according to the abnormal judgment result of the three types of storage nodes, or generate a second type of storage node abnormal judgment instruction;

[0062] The second-class node analysis module 40 is connected to the third-class node analysis module 30, and is used to receive the second-class storage node abnormality judgment instruction, and judge whether the second-class storage node is abnormal based on the node abnormality ratio or node abnormality ratio change trend of a number of third-class storage nodes corresponding to any of the second-class storage nodes, so as to formulate a second-class node repair strategy according to the comparison result of the second-class abnormality number of the number of abnormal second-class storage nodes and the preset abnormality number, or generate a first-class storage node abnormality judgment instruction;

[0063] The first-class node analysis module 50 is connected to the second-class node analysis module 40, and is used to receive the first-class storage node abnormality judgment instruction, analyze the abnormal distribution of a number of abnormal second-class storage nodes, so as to judge the abnormality degree of a number of first-class storage nodes, and determine the first-class node repair strategy according to the abnormality degree judgment result;

[0064] Among them, any of the first-category storage nodes includes at least one second-category storage node, and any of the second-category storage nodes includes at least one third-category storage node.

[0065] Specifically, the embodiment of the present invention sets the storage module to identify the invoice structure corresponding to the invoice to be stored, and further identifies the first keyword and the second keyword based on the invoice structure, thereby providing a clear framework for subsequent storage, query and management, improving the efficiency of data management, and reasonably allocating the invoice to different storage nodes according to different characteristics based on the distributed storage method, so that the data storage is more balanced, avoiding the problem of excessive storage pressure caused by a large amount of data concentrated in a certain place, and dynamically adjusting the monitoring frequency according to the actual performance of the storage node by setting the frequency setting module. When the node performance tends to be tense, the monitoring frequency is timely increased to pay more attention to the status of the storage node, discover problems in advance, and ensure the security and stability of data storage. By setting three types of node analysis modules to detect node security anomalies corresponding to three types of storage nodes, it is helpful to quickly locate the storage node level where the problem is located, and provide a clear direction for subsequent further investigation and processing. By adjusting the initial detection frequency, the subsequent situation of the relevant storage node can be paid more attention after the anomaly is found. By generating the second type of storage node abnormality judgment instruction, the impact range of the abnormal situation is further transmitted and analyzed to the upper node, realizing the local The gradual investigation and response mechanism from the top to the bottom ensures the stable operation of the entire archive management system. By setting up a second-class node analysis module to comprehensively consider the judgment method of multiple third-class storage nodes, it can more comprehensively and accurately evaluate the overall status of the second-class storage nodes. According to the comparison results of the second-class anomaly number of several abnormal second-class storage nodes with the preset anomaly number, a second-class node repair strategy is formulated. Different response measures can be taken for different degrees of abnormal situations, realizing the refined management of the abnormal situation of the second-class storage nodes, improving the maintainability and data security of the system, generating a first-class storage node abnormality judgment instruction, and converting the second-class storage nodes into a second-class node repair strategy. The abnormal situation of the storage node is further transmitted to the first-class storage node level for analysis, so that the entire system can check and handle the abnormal situation layer by layer from bottom to top, ensuring that any potential problems can get sufficient attention and resolution. By setting up a first-class node analysis module to analyze the abnormal distribution of several abnormal second-class storage nodes to determine the abnormal degree of several first-class storage nodes, based on the evaluation method of the abnormal distribution of lower-level nodes, a more comprehensive understanding of the scope and degree of impact that a first-class storage node may be affected is achieved, which provides an important basis for the subsequent formulation of a reasonable first-class node repair strategy and improves the efficiency of archive management and data security.

[0066] It is understandable that the invoice structure of the embodiment of the present invention mainly includes the following parts:

[0067] ‌Heading text‌: The beginning part of the invoice, usually including the invoice name, invoice code, number and other information.

[0068] ‌Article‌: The main body of the invoice lists the specific contents of the invoice in detail, such as the track number, copy number and purpose, invoice date, opening bank and account number, product name or business item, unit of measurement, amount, etc.

[0069] ‌Ending‌: The ending part of the invoice, usually containing the signature or seal of the invoicing unit and the invoice issuer. ‌

[0070] Specifically, the storage module 10 includes:

[0071] The structure division unit 11 is used to identify a plurality of edge contour regions corresponding to any of the invoices to be stored based on an edge detection algorithm, determine a plurality of actual area areas corresponding to the plurality of edge contour regions, and determine the actual position distribution of the plurality of edge contour regions in the invoices to be stored, and use the similarity between the plurality of actual position distributions and the plurality of actual area areas corresponding to the plurality of invoices to be stored as a primary division standard to perform division and obtain a primary division result;

[0072] The first division unit 12 is connected to the structure division unit 11, and is used to determine the invoice header area based on the actual location distribution corresponding to any of the invoices to be stored, identify the first keyword in the invoice header area, and use the similarity between the first keywords corresponding to the invoices to be stored as a secondary division standard to perform division, and obtain a secondary division result;

[0073] The second division unit 13 is connected to the first division unit 12, and is used to determine the invoice text area based on the actual location distribution corresponding to any of the invoices to be stored, identify a plurality of second keywords in the invoice text area, and divide the invoices using the similarity between the plurality of second keywords corresponding to the plurality of invoices to be stored as a third division standard to obtain a third division result;

[0074] The storage unit 14 is connected to the structure division unit 11, the first division unit 12 and the second division unit 13, and is used to perform distributed storage on the plurality of invoices to be stored based on the first division result, the second division result and the third division result, respectively, to obtain a plurality of first-class storage nodes, a plurality of second-class storage nodes and a plurality of third-class storage nodes.

[0075] Specifically, the embodiment of the present invention extracts important features of the invoice in terms of layout and space occupancy by setting a structural division unit to determine the actual area corresponding to the edge contour area and the actual position distribution in the invoice to be stored, reflecting the similarities and differences between different invoices in terms of format, typesetting, etc., providing a rich basis for subsequent division based on similarity, and facilitating more detailed and accurate classification of invoices. By setting a first division unit to accurately determine the invoice header area and identify the first keyword for division, the classification of invoices is more in line with its business attributes and important features, thereby improving the pertinence and accuracy of the classification. By setting a second division unit to determine the invoice body area and identify several second keywords for division, the specific business content can be classified into Invoices with similarities in different aspects are further subdivided, making the classification of invoices more refined and improving the flexibility and practicality of data management. By setting up storage units to distribute the invoices to be stored, the characteristics and classification results of the invoices at different levels extracted by the previous division units are fully utilized, and the invoices with different similarities and business attributes are reasonably allocated to different storage nodes, avoiding the problems of excessive storage pressure and low query efficiency caused by the centralized storage of a large number of invoices. When querying a specific type of invoice, the relevant data can be quickly located according to the storage node category to which it belongs, which improves the speed and accuracy of data retrieval. At the same time, it is also convenient for targeted management and maintenance of storage nodes of different categories, thereby improving the operation efficiency and data management efficiency of the entire archive management system.

[0076] It can be understood that any one of the first-category storage nodes in the embodiments of the present invention includes at least one second-category storage node, and any one of the second-category storage nodes includes at least one third-category storage node.

[0077] It can be understood that in the embodiment of the present invention, the invoice header area, invoice body area and invoice tail area are determined based on the actual position distribution corresponding to any of the invoices to be stored, on the premise that the invoice needs to be determined as a positive detection, which can be determined by determining the direction of any text in the invoice to be stored, wherein the invoice header area is the beginning part of the invoice to be stored, generally located at the upper end of the invoice, the invoice tail area is the end of the invoice, generally located at the lower end of the invoice, and the rest is used as the invoice body area.

[0078] It can be understood that in the embodiment of the present invention, the first keyword is determined based on the characteristic information of the header area of ​​the invoice, and the characteristic information is usually the fixed header information of the invoice, such as the invoice number, the invoice date, etc.;

[0079] The second keyword is determined based on the detailed information in the text area of ​​the invoice, which usually includes the name, quantity, unit price, etc. of the goods or services.

[0080] Embodiment 1: After receiving a batch of invoices to be stored, the structure division unit first starts the edge detection algorithm for each invoice. For example, for an invoice for purchasing office supplies, the edge detection algorithm can clearly identify the edge contour areas of each part of the invoice, such as the edge contours formed by the invoice header, the body, the amount column, the invoice number area, etc., and then determine the actual area corresponding to these edge contour areas. For example, the invoice header may occupy one-tenth of the entire invoice paper area, and the body occupies one-half. At the same time, the actual position distribution of each edge contour area in the invoice to be stored is clarified, such as the invoice header is located in the center of the top of the invoice, the amount column is in the lower right corner of the invoice, etc. Then, the actual position distribution corresponding to these invoices to be stored and the similarity between the actual area are used as the primary division standard for division. Assuming that we have ten invoices, the position and area similarity of the header area of ​​five invoices are relatively high, and the layout and area of ​​the body part are also relatively similar, then these five invoices will be divided into one group, forming part of the primary division result.

[0081] For each group of invoices in the above-mentioned primary division results, taking one group as an example, based on the actual location distribution corresponding to each invoice in the group, the invoice header area can be determined, that is, the area that usually contains key information such as the invoice issuing unit and the invoice number. In the invoice header area, the first keyword is identified. For example, for this group of invoices for purchasing office supplies, the first keyword in the header area may be "office supply purchase", "invoice number", "issuing unit name", etc. The similarity between several first keywords corresponding to several invoices to be stored in this group is used as the secondary division standard for division. If the first keywords of three of the invoices have extremely high similarity, all revolve around a specific office supply purchase business and the invoice number and the issuing unit also have a high match, then these three invoices will be further divided out to form part of the secondary division results.

[0082] For each group of invoices in the secondary division result, take a certain group as an example, determine the invoice text area based on the actual location distribution corresponding to each invoice in the group, that is, the area containing detailed information such as specific purchase details and amount. In the invoice text area, identify several second keywords. For invoices for purchasing office supplies, the second keywords in the text area may include specific office supply names such as "laptops", "printer cartridges", "office desks and chairs" and corresponding amount numbers. The similarity between several second keywords corresponding to several invoices to be stored in this group is used as the third division standard for division. Assuming that the second keywords of two of the invoices are very similar in terms of specific office supply purchase details and amounts, then these two invoices will be divided out again to form part of the third division results.

[0083] Specifically, the frequency setting module includes:

[0084] A performance collection unit, used to collect the real-time CPU usage and real-time memory usage corresponding to any one of the storage nodes;

[0085] A frequency calculation unit is connected to the performance collection unit, and is used to calculate the real-time performance index of any one type of storage node based on the real-time CPU usage and the real-time memory usage, and determine the initial detection frequency based on the real-time performance index.

[0086] Specifically, the frequency calculation unit includes:

[0087] An index calculation subunit, used for calculating a weighted average of the real-time CPU usage and the real-time memory usage to generate the real-time performance index;

[0088] The frequency determination subunit is connected to the index calculation subunit and is used to compare any of the real-time performance indexes with the preset performance index. When the real-time performance index is greater than the preset performance index, the preset detection frequency is shortened to obtain the initial detection frequency. When the real-time performance index is less than or equal to the preset performance index, the preset detection frequency is used as the initial detection frequency.

[0089] It can be understood that an example of the frequency calculation unit described in the embodiment of the present invention is that at a certain moment, the real-time CPU usage of the current class of storage nodes is 60%, and the real-time memory usage is 40% obtained through the system monitoring tool. At the same time, the weight of the CPU usage is set to 0.6, and the weight of the memory usage is set to 0.4, then the corresponding real-time performance index is 0.52. Assuming that the preset performance index is set to 0.4, the frequency determination subunit increases the preset detection frequency, and the increase percentage can be calculated according to the difference between the real-time performance index and the preset performance index divided by the preset performance index, so as to increase the preset detection frequency according to the increase percentage, wherein the preset weight can be determined according to the actual situation, based on the evaluation of the importance of CPU and memory.

[0090] It can be understood that the preset performance index in the embodiment of the present invention is determined by the average of several historical performance indexes corresponding to several storage nodes of the first category;

[0091] The preset detection frequency is a standard detection frequency preset by the system under normal load conditions. For example, the preset detection frequency may be a check once a week.

[0092] See also Figure 3 As shown, the three-type node analysis module 30 includes:

[0093] A data monitoring unit 31, used for detecting the actual completeness and actual accuracy of a number of stored invoices in any of the three types of storage nodes in real time based on the initial detection frequency;

[0094] A safety determination unit 32, connected to the data monitoring unit 31, for determining an actual node safety degree based on the actual integrity and the actual accuracy;

[0095] The third-category node judgment unit 33 is connected to the safety determination unit 32, and is used to compare the actual node safety with the preset node safety, and adjust the initial detection frequency based on the comparison result, or further judge the abnormal situation of the second-category storage node.

[0096] Specifically, the embodiment of the present invention helps to discover problems in storage nodes in a timely manner and prevent data loss or damage by setting a data monitoring unit to detect the actual completeness and actual accuracy of several stored invoices in any of the three types of storage nodes in real time. By setting a security determination unit to determine the actual node security based on the actual completeness and actual accuracy, it helps to comprehensively quantify the current security status of the storage node, making the judgment on the node security more accurate and reliable, and taking corresponding measures as needed to ensure the security and stability of data storage. By setting three types of node judgment units to adjust the initial detection frequency according to the comparison result of the actual node security and the preset node security, the monitoring frequency is dynamically adjusted according to the actual situation of the storage node, thereby more effectively utilizing resources. It is also used to judge the abnormal situation of the second type of storage node, which helps to further transmit and analyze the impact range of the abnormal situation to the upper-level node, realizing a gradual investigation and response mechanism from the local to the upper level, and ensuring the stable operation of the entire archive management system.

[0097] It can be understood that the three-category node judgment unit described in the embodiment of the present invention is used to determine that the three-category storage node is non-abnormal when the actual node safety is greater than the preset node safety and less than 100%, and adjust the initial detection frequency; when the actual node safety is less than or equal to the preset node safety, it is used to determine that the three-category storage node is abnormal, and then judge the abnormal situation of the two-category storage node.

[0098] It can be understood that the actual node security described in the embodiment of the present invention may be actual integrity×actual accuracy.

[0099] It can be understood that the preset node safety level is 90% in the embodiment of the present invention. Setting the preset node safety level to 90% is a practical and feasible goal, which can not only ensure safety, but also detect potential safety issues early, thereby taking preventive measures to prevent the problem from worsening.

[0100] Specifically, the data monitoring unit includes:

[0101] A completeness determination subunit is used to compare the actual memory proportions of the stored invoices with the preset memory proportions to obtain a proportion comparison result, and determine the actual completeness based on the proportion comparison result;

[0102] The accuracy determination subunit is used to compare the stored text information on any of the stored invoices with the corresponding actual text information to obtain an information comparison result, and determine the actual accuracy based on a number of information comparison results.

[0103] It can be understood that the actual completeness described in the embodiment of the present invention can be calculated by multiplying the difference between the actual memory ratio and the preset memory ratio by 100%, wherein the preset ratio is the historical ratio corresponding to the complete storage of the invoice history. If the difference between the actual memory ratio and the preset memory ratio is 0, the corresponding actual completeness is 100%.

[0104] It can be understood that the actual accuracy described in the embodiment of the present invention can be obtained by comparing each text item (such as each line of text, each field) on the stored invoice with the corresponding text item on the original invoice, counting the number of matching items found in the comparison, determining the total number of items compared, and calculating the matching rate based on the formula: Matching rate (%) = (number of matches / total number of items) × 100%. For example, if there are 100 text items on the original invoice, and during the comparison process it is found that 90 text items on the stored invoice completely match the original invoice, then the matching rate is (90 / 100) × 100% = 90%.

[0105] See also Figure 4 As shown, the second type node analysis module 40 includes:

[0106] A proportion calculation unit 41 is used to count the total number of nodes and the number of three types of abnormalities of a number of third-type storage nodes corresponding to any of the second-type storage nodes, and calculate the node abnormality proportion based on the number of three types of abnormalities and the total number of nodes;

[0107] A proportion determination unit 42, connected to the proportion calculation unit 41, is used to determine the abnormal situation of the second type of storage node based on the comparison result between the node abnormality proportion and the preset abnormality proportion, or to analyze the change trend of the node abnormality proportion;

[0108] A trend determination unit 43, connected to the proportion determination unit 42, is used to analyze the change trend of the abnormal proportion of several nodes within the preset collection frequency, so as to determine the abnormal situation of the second type of storage node according to the change trend determination result;

[0109] The first strategy determination unit 44 is used to formulate a second-category node repair strategy according to a comparison result between the second-category abnormality quantity of a plurality of abnormal second-category storage nodes and a preset abnormality quantity, or to generate a first-category storage node abnormality determination instruction.

[0110] Specifically, the embodiment of the present invention calculates the node abnormality ratio by setting a ratio calculation unit, which helps to quantify the health status of the second-class storage nodes and provide a basis for determining their abnormal conditions. By setting a ratio determination unit to determine the abnormal conditions of the second-class storage nodes based on the comparison result of the node abnormality ratio and the preset abnormality ratio, it helps to find problems in time and take corresponding measures to ensure the security of data storage. By setting a trend judgment unit to analyze the changing trend of the node abnormality ratio within the preset collection frequency to determine the abnormal conditions of the second-class storage nodes, it helps to judge the problem trend and take preventive measures in advance. By setting a first strategy determination unit to formulate a second-class node repair strategy according to the comparison result of the second-class abnormality number of several abnormal second-class storage nodes with the preset abnormality number, different response measures can be taken for abnormal conditions of different degrees, and refined management of the abnormal conditions of the second-class storage nodes is realized, which improves the maintainability and data security of the system, generates a first-class storage node abnormality judgment instruction, and further transmits the abnormal conditions of the second-class storage nodes to the first-class storage node level for analysis, so that the entire system can check and handle the abnormal conditions from bottom to top layer by layer, ensuring that any potential problems can be fully concerned and resolved, thereby improving the security of data storage.

[0111] It can be understood that the preset abnormality ratio in the embodiment of the present invention is 1 / 5.

[0112] It can be understood that the proportion determination unit described in the embodiment of the present invention is used to determine that the second type of storage node is abnormal when the node abnormality proportion is greater than or equal to the preset abnormality proportion, and to analyze the changing trend of the node abnormality proportion when the node abnormality proportion is less than the preset abnormality proportion.

[0113] It can be understood that the preset collection frequency in the embodiment of the present invention is ten times.

[0114] It can be understood that the trend judgment unit described in the embodiment of the present invention is used to determine that the second type of storage node is abnormal when the node abnormality ratio change trend is gradually increasing, and to determine that the second type of storage node is normal when the node abnormality ratio change trend is not gradually increasing.

[0115] Specifically, the first strategy determination unit includes:

[0116] an abnormality locating subunit, used to identify a number of abnormal category II storage nodes, and locate a number of abnormal category III storage nodes corresponding to any abnormal category II storage node;

[0117] A strategy formulation unit is connected to the abnormal location subunit, and is used to calculate the safety difference between the actual node safety and the preset node safety corresponding to any of the three abnormal storage nodes, compare the safety difference with the preset difference, and determine the second-category node repair strategy based on the difference comparison result.

[0118] It can be understood that the first strategy determination unit in the embodiment of the present invention is used to generate a type of storage node abnormality judgment instruction when the number of the second type of abnormalities is greater than or equal to the preset number of abnormalities, and to formulate a repair strategy when the number of the second type of abnormalities is less than the preset number of abnormalities.

[0119] It can be understood that the preset number of exceptions in the embodiment of the present invention is 1 / 5 of the total number of the second-class storage nodes.

[0120] It can be understood that the Class II node repair strategy described in the embodiment of the present invention can be: when the safety difference is greater than or equal to the preset difference, the abnormal Class III storage node is deleted, the Class III storage node is re-established, and the corresponding data on the abnormal Class III storage node is migrated to the new Class III storage node; when the safety difference is less than the preset difference, the data in the abnormal Class III storage node is backed up and restored, wherein the preset difference is set to 10%.

[0121] Specifically, the first type of node analysis module includes:

[0122] A degree determination unit, used to identify a number of abnormal second-class storage nodes, determine abnormal distribution density and abnormal correlation corresponding to the number of abnormal second-class storage nodes, and determine the abnormal degree based on the abnormal distribution density and abnormal correlation;

[0123] The second strategy determination unit is connected to the degree determination unit and is used to determine a type of node repair strategy according to a comparison result between the abnormality degree and a preset degree.

[0124] The degree determination unit comprises:

[0125] A density calculation subunit, used to calculate the distribution density of a number of abnormal type II storage nodes in any abnormal type I storage node;

[0126] A correlation analysis subunit is used to extract a number of second keywords corresponding to a number of abnormal third-category storage nodes in a number of abnormal second-category storage nodes in any of the abnormal first-category storage nodes as a number of abnormal keywords, and determine the abnormal correlation between the number of abnormal keywords;

[0127] The degree determination subunit is connected to the density calculation subunit and the correlation analysis subunit, and is used to determine the abnormality degree based on the distribution density and the abnormality correlation.

[0128] It can be understood that the correlation analysis subunit described in the embodiment of the present invention can determine the abnormal correlation between several abnormal keywords by comparing the actual frequency of the several abnormal keywords in the invoice of the same purchase order with the preset frequency.

[0129] It can be understood that the degree determination subunit described in the embodiment of the present invention is used to determine that the abnormality degree is high when the distribution density is greater than the preset density and the actual frequency is greater than the preset frequency, and to determine that the abnormality degree is low when the distribution density is less than the preset density and the actual frequency is less than the preset frequency, and in other cases, to determine that the abnormality degree is medium.

[0130] It is understood that the preset density in the embodiment of the present invention is 0.3;

[0131] The preset frequency is 0.5.

[0132] Embodiment 2, for example, assume that a class I storage node covers a total of 100 logical storage areas, and through system detection, it is found that there are abnormalities in the class II storage nodes in 20 logical storage areas. Then, the density calculation subunit can calculate the distribution density as 20 / 100=0.2 according to the formula: distribution density = number of areas where the abnormal class II storage nodes are located / total number of areas. In the invoice data stored in the abnormal class II storage nodes, for some invoices involving procurement business, the second keyword in the text area of ​​the invoice stored in the abnormal class III storage nodes may have "computer equipment" and "printer cartridges". Assume that the frequency of "computer equipment" and "printer cartridges" appearing in the invoices of the same purchase order is 0.3, which is less than the preset frequency, then the correlation is low, the distribution density 0.2 is less than the preset density 0.3 and the correlation is low, then it is determined that the degree of abnormality this time is low.

[0133] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0134] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent file management system based on RPA, characterized in that: include: A storage module is used to identify the invoice structure corresponding to the invoices to be stored, identify the first keywords and the second keywords based on the invoice structure, and perform distributed storage on the invoices to be stored based on the invoice structure, the first keywords and the second keywords to obtain a number of first-class storage nodes, a number of second-class storage nodes and a number of third-class storage nodes; A frequency setting module, connected to the storage module, for collecting first node performance data corresponding to a plurality of first-class storage nodes, and determining an initial detection frequency based on the first storage node performance data; The three-type node analysis module is connected to the frequency setting module, and is used to detect the node safety anomaly corresponding to any of the three types of storage nodes in real time based on the initial detection frequency, determine whether the three types of storage nodes are abnormal based on the node safety anomaly, and adjust the initial detection frequency according to the abnormal judgment result of the three types of storage nodes, or generate a second type of storage node abnormal judgment instruction; A second-category node analysis module is connected to the third-category node analysis module, and is used to receive a second-category storage node abnormality judgment instruction, and judge whether the second-category storage node is abnormal based on the node abnormality ratio or node abnormality ratio change trend of a number of third-category storage nodes corresponding to any of the second-category storage nodes, so as to formulate a second-category node repair strategy according to a comparison result of the number of second-category abnormalities of a number of abnormal second-category storage nodes with a preset number of abnormalities, or generate a first-category storage node abnormality judgment instruction; A first-class node analysis module is connected to the second-class node analysis module, and is used to receive a first-class storage node abnormality judgment instruction, analyze the abnormal distribution of a number of abnormal second-class storage nodes, so as to judge the abnormality degree of a number of first-class storage nodes, and determine the first-class node repair strategy according to the abnormality degree judgment result; Among them, any of the first-category storage nodes includes at least one second-category storage node, and any of the second-category storage nodes includes at least one third-category storage node.

2. The RPA-based intelligent archive management system according to claim 1 is characterized in that: The storage module comprises: A structure division unit, for identifying a plurality of edge contour regions corresponding to any of the invoices to be stored based on an edge detection algorithm, determining a plurality of actual area areas corresponding to the plurality of edge contour regions, and determining actual position distributions of the plurality of edge contour regions in the invoices to be stored, and using the similarities between the plurality of actual position distributions and the plurality of actual area areas corresponding to the plurality of invoices to be stored as a primary division standard to perform division, and obtain a primary division result; A first division unit, connected to the structure division unit, is used to determine the invoice header area based on the actual location distribution corresponding to any of the invoices to be stored, identify the first keyword in the invoice header area, and use the similarity between the first keywords corresponding to the invoices to be stored as a secondary division standard to perform division, and obtain a secondary division result; A second division unit, connected to the first division unit, is used to determine the invoice text area based on the actual location distribution corresponding to any of the invoices to be stored, identify a plurality of second keywords in the invoice text area, and divide the invoices using the similarity between the plurality of second keywords corresponding to the plurality of invoices to be stored as a third division standard to obtain a third division result; The storage unit is connected to the structure division unit, the first division unit and the second division unit, and is used for distributing and storing a plurality of invoices to be stored based on the first division result, the second division result and the third division result respectively.

3. The RPA-based intelligent archive management system according to claim 2 is characterized in that: The frequency setting module comprises: A performance collection unit, used to collect the real-time CPU usage and real-time memory usage corresponding to any one of the storage nodes; A frequency calculation unit is connected to the performance collection unit, and is used to calculate the real-time performance index of any one type of storage node based on the real-time CPU usage and the real-time memory usage, and determine the initial detection frequency based on the real-time performance index.

4. The RPA-based intelligent archive management system according to claim 3 is characterized in that: The frequency calculation unit comprises: An index calculation subunit, used for calculating a weighted average of the real-time CPU usage and the real-time memory usage to generate the real-time performance index; The frequency determination subunit is connected to the index calculation subunit and is used to compare any of the real-time performance indexes with the preset performance index. When the real-time performance index is greater than the preset performance index, the preset detection frequency is shortened to obtain the initial detection frequency. When the real-time performance index is less than or equal to the preset performance index, the preset detection frequency is used as the initial detection frequency.

5. The RPA-based intelligent archive management system according to claim 4 is characterized in that: The three types of node analysis modules include: A data monitoring unit, used for detecting the actual completeness and actual accuracy of a number of stored invoices in any of the three types of storage nodes in real time based on the initial detection frequency; a safety determination unit, connected to the data monitoring unit, for determining an actual node safety degree based on the actual integrity and the actual accuracy; The third type of node judgment unit is connected to the safety determination unit, and is used to compare the actual node safety with the preset node safety, and adjust the initial detection frequency based on the comparison result, or further judge the abnormal situation of the second type of storage node.

6. The RPA-based intelligent archive management system according to claim 5 is characterized in that: The data monitoring unit comprises: A completeness determination subunit is used to compare the actual memory proportions of the stored invoices with the preset memory proportions to obtain a proportion comparison result, and determine the actual completeness based on the proportion comparison result; The accuracy determination subunit is used to compare the stored text information on any of the stored invoices with the corresponding actual text information to obtain an information comparison result, and determine the actual accuracy based on a number of information comparison results.

7. The RPA-based intelligent archive management system according to claim 6 is characterized in that: The second type of node analysis module includes: A proportion calculation unit, used to count the total number of nodes and the number of three types of abnormalities of a number of third-type storage nodes corresponding to any of the second-type storage nodes, and calculate the node abnormality proportion based on the number of three types of abnormalities and the total number of nodes; A proportion determination unit, connected to the proportion calculation unit, for determining the abnormal situation of the second type of storage nodes based on the comparison result between the node abnormality proportion and the preset abnormality proportion, or analyzing the change trend of the node abnormality proportion; A trend judgment unit, connected to the proportion judgment unit, is used to analyze the change trend of the abnormal proportion of several nodes within the preset collection frequency, so as to determine the abnormal situation of the second type of storage node according to the change trend judgment result; The first strategy determination unit is used to formulate a second-category node repair strategy according to a comparison result between the second-category abnormality quantity of a plurality of abnormal second-category storage nodes and a preset abnormality quantity, or to generate a first-category storage node abnormality judgment instruction.

8. The RPA-based intelligent archive management system according to claim 7 is characterized in that: The first strategy determination unit includes: an abnormality locating subunit, used to identify a number of abnormal category II storage nodes, and locate a number of abnormal category III storage nodes corresponding to any abnormal category II storage node; A strategy formulation unit is connected to the abnormal location subunit, and is used to calculate the safety difference between the actual node safety and the preset node safety corresponding to any of the three abnormal storage nodes, compare the safety difference with the preset difference, and determine the second-category node repair strategy based on the difference comparison result.

9. The RPA-based intelligent archive management system according to claim 8, characterized in that: The first type of node analysis module includes: A degree determination unit, used to identify a number of abnormal second-class storage nodes, determine abnormal distribution density and abnormal correlation corresponding to the number of abnormal second-class storage nodes, and determine the abnormal degree based on the abnormal distribution density and abnormal correlation; The second strategy determination unit is connected to the degree determination unit and is used to determine a type of node repair strategy according to a comparison result between the abnormality degree and a preset degree.

10. The RPA-based intelligent archive management system according to claim 9, characterized in that: The degree determination unit comprises: A density calculation subunit, used to calculate the distribution density of a number of abnormal type II storage nodes in any abnormal type I storage node; A correlation analysis subunit is used to extract a number of second keywords corresponding to a number of abnormal third-category storage nodes in a number of abnormal second-category storage nodes in any of the abnormal first-category storage nodes as a number of abnormal keywords, and determine the abnormal correlation between the number of abnormal keywords; The degree determination subunit is connected to the density calculation subunit and the correlation analysis subunit, and is used to determine the abnormality degree based on the distribution density and the abnormality correlation.

Images