Intelligent auditing method and system for engineering service plan

Abstract

An electric power engineering service class plan report is composed of a demand plan of an ERP system and a technical specification book and a quotation table of an ECP system, the invention provides an intelligent auditing method and system for an engineering service class plan, and the system comprises a data preprocessing module which carries out decompression and document format conversion on a plan file in the ECP system; the data analysis module is used for carrying out all-around and multi-dimensional review by utilizing preset criteria from an ERP (Enterprise Resource Planning) demand plan, an ECP (Error Correction Protocol) technical specification book and a quotation table; according to the system, various kinds of information and documents in the plan reporting and auditing process are processed, auditing suggestions are intelligently given, and plan auditing is completed more efficiently.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system informatization, and more specifically, relates to a method and system for intelligent review of engineering service plans. Background Art

[0002] The bidding documents for engineering service projects are the basic guiding documents for procurement activities. The review of engineering service plans is an important part of organizing and implementing procurement. Scientifically and completely reviewing engineering service procurement plans is the prerequisite for completing procurement tasks with high quality and efficiency. Researching the problems, countermeasures and measures in the submission of engineering service procurement plans is of great significance for the correct implementation of procurement and improving procurement efficiency, benefits and effectiveness.

[0003] The engineering service procurement plan submission consists of three parts: the demand plan of the ERP system planning module and the technical specifications and quotation form of the ECP system. At present, the submitted engineering service procurement plan has problems such as non-standard preparation and inconsistent format; the plan review process has problems such as inconsistent standards and inefficient review, which are mainly manifested in the following aspects:

[0004] (1) ERP demand planning is not prepared in a standardized manner.

[0005] 1. The fields of the bidding project section name, start date, bidding amount, estimated amount, bidding scope, project scale and overview, contractor requirements, etc. in the demand plan are not filled in in a standardized manner and fail to be filled in according to the latest requirements.

[0006] 2. The demand plan is split to avoid the situation of budget saving. When using the limited authorized procurement clause, budget saving is avoided by splitting the project and submitting the authorized procurement.

[0007] 3. The demand plan exceeds the procurement scope. For engineering service projects that have been included in the provincial recruitment scope, the authorization procurement is still being submitted.

[0008] (2) The ECP technical specifications and quotation sheets are not prepared in a standardized manner.

[0009] 1. The template was not prepared according to the latest technical specifications.

[0010] 2. The technical specifications and quotation sheets finally uploaded to the ECP system are not the final review versions.

[0011] 3. Some descriptions in the technical specifications are inconsistent with the demand plan, or involve the negative list of business outsourcing.

[0012] (3) Audit standards are not unified and efficiency is low.

[0013] Since the procurement plans submitted in some batches are relatively concentrated and the number of items in the summarized procurement plans is also relatively large, and as the audit requirements continue to increase, the audit experts' audit of the procurement plans is often not standardized, and the preparation of audit opinions is time-consuming, labor-intensive, and inefficient. Summary of the invention

[0014] In order to solve the deficiencies in the prior art, the present invention provides a method and system for intelligent review of engineering service plans, which processes various types of information and documents in the plan submission and review process, intelligently gives review opinions, and completes the plan review more efficiently.

[0015] The present invention adopts the following technical solution.

[0016] In one aspect, the present invention provides a method for intelligent review of engineering service plans, comprising:

[0017] S1: In the ERP system, the plans of each department are summarized and the ERP demand plan summary table is exported and uploaded to the data preprocessing module. The technical specifications and quotation table are uploaded to the ECP system, and the technical specification ID number is generated and filled into the demand plan summary table. According to the ID number, the ECP technical specification and ECP quotation table documents are downloaded and exported in batches to generate a ZIP compressed package;

[0018] S2: Fill the directory where the compressed package obtained in step S1 is located into the data preprocessing module for decompression to obtain corresponding documents, and perform corresponding format conversion on different documents as required;

[0019] S3: After decompressing the format and converting it in step S2, the technical specification and quotation sheet are obtained, and the demand plan summary sheet is imported again for review according to the requirements of different fields. The demand plan, technical specification, quotation sheet and other related information are associated through the technical specification ID number, and a comprehensive and multi-dimensional review is performed using the preset criteria;

[0020] S4: After the data analysis module completes the review of all plans, it will classify and sort these opinions according to the ERP business opinions and ECP technical opinions based on the judgment opinions of the procurement plan, and place them in the review opinion table in order. The data output module will generate the review opinion table and plan summary table.

[0021] Preferably, the demand planning data reported by each department in the planning module of the ERP platform is obtained, including department ID, product category, demand quantity and time period, and the ERP system collects and summarizes the demand planning data by using database access;

[0022] The acquired demand planning data of all departments are synchronized to a unified data storage location through the database data replication tool, and the demand planning data is cleaned. The cleaning operation includes deduplication of data, standardization of data and removal of outliers. A cross-departmental data integration algorithm is used to coordinate the consistency of data reported by different departments.

[0023] Preferably, the specific workflow of the cross-departmental data integration algorithm is:

[0024] According to the demand planning data reported by each department i , and confirm the weight coefficient w based on the comprehensive evaluation of each department i , analyze and obtain the weighted average of the demand planning data of all departments as the target weight mean μ w ;

[0025] The weighted average formula:

[0026]

[0027] Where:

[0028] w i represents the weight of the i-th department;

[0029] X i Represents the demand planning data of the i-th department;

[0030] n represents the total number of departments that have uploaded demand planning data.

[0031] For each department's demand planning data X i , the square root index difference metric is used to evaluate the data and the target weight mean μ w The deviation value d i , the formula is:

[0032]

[0033] Where:

[0034] e is the base of natural logarithms.

[0035] Preferably, if the deviation value d i A large value indicates that the data source of the department is significantly different from the expected weighted average and data adjustment may be required;

[0036] For each department's demand planning data X i , if d i If it is greater than the preset threshold, the stepwise approximation method is used to adjust the data to obtain the adjusted demand plan data. The formula is:

[0037]

[0038] Where:

[0039] represents the demand planning data after adjustment of the demand planning data of the i-th department;

[0040] η represents the adjustment step size parameter.

[0041] Preferably, the pandas library is used to read and operate the ERP demand plan summary table in xlsx format, and data cleaning and format conversion are performed after removing blank fields, and multi-dimensional verification and review are performed according to the requirements of different fields;

[0042] After decompressing the format, the technical specifications and quotation sheets are converted to extract key information from the technical specifications through text parsing, and the pandas library is used to read the quotation sheets for multi-dimensional verification and review based on the requirements of different fields;

[0043] The demand plan, technical specification, quotation and other relevant information are linked through the technical specification ID number, and a comprehensive and multi-dimensional review is conducted using preset criteria.

[0044] Preferably, the Levenshtein distance is used to calculate the similarity to determine whether the bidding project section name in the ERP demand plan summary table is similar to the project name that has been included in the provincial bidding and procurement scope, and then the authorization procurement is submitted, and the opinions that require further verification by experts are given;

[0045] The character string similarity calculation specifically includes:

[0046] The Levenshtein distance algorithm is used to calculate the similarity between the bidding project section name and the project name within the provincial bidding and procurement scope; the Levenshtein distance can measure the minimum number of editing operations between two strings, including insertion, deletion and replacement;

[0047] The recursive relationship of Levenshtein distance is as follows:

[0048]

[0049] in:

[0050] D(i,j) represents the minimum edit distance to convert string A[i] into string B[j];

[0051] A[i] represents the first i characters of string A;

[0052] B[j] represents the first j characters of string B

[0053] A represents the name string A of the bidding project section;

[0054] B represents the project name string B that has been included in the provincial bidding and procurement scope;

[0055] D(i-1,j)+1 represents the deletion operation, that is, deleting a character from string A;

[0056] D(i,j-1)+1 represents an insertion operation, that is, inserting a character into string A;

[0057] cost(A[i],B[j]) indicates whether character A[i] and character B[j] are the same. If they are the same, the cost is 0; if they are different, the cost is 1;

[0058] If the distance is less than the preset threshold, the similarity is judged to be high and an opinion is given that further verification by experts is required.

[0059] Preferably, the bidding scope, project scale and overview of the ERP demand plan summary table are judged, special characters are used to detect whether there are illegal characters and garbled characters in the text, and the fields are checked through pattern matching according to the set template to check whether the fields are omitted or incomplete, and the year information is extracted by regular expression to check whether the wrong year is filled in, and the content in the bidding document is associated with the predefined terms for judgment, and the natural language processing technology is used to identify whether there are inappropriate descriptions in the content of the bidding document; the software directly gives the review opinion, if it involves the content of the negative list of business outsourcing, the specific text content of the negative list of business outsourcing is given, and opinions that require further verification by experts are put forward.

[0060] Preferably, the technical specification is compared with the fixed template built into the system for text content, and an exact match is used to check whether the submitted content is completely consistent with the fixed template to determine whether it is compiled according to the fixed template. Fuzzy matching is used to detect spelling and format differences for non-critical content, and keyword detection is used to determine whether key fields and content are missing, including the content that needs to be replaced and modified in the fixed template has not been modified, and the software directly gives the review opinion. 9. A method for intelligent review of engineering service plans according to claim 5, characterized in that:

[0061] The quotation sheet is judged by using Levenshtein distance fuzzy matching to check whether the filled-in project title is similar to the bidding project section name in the demand plan, and then determine whether they are consistent. The serial number column in the quotation sheet is traversed to check whether it is filled in in ascending order to determine whether the serial number column is filled in correctly.

[0062] The present invention also provides a system for intelligent review of engineering service plans, and the method for intelligent review of engineering service plans includes:

[0063] Data preprocessing module, data analysis module and data output module;

[0064] The data preprocessing module is used to receive the ERP demand plan summary table and fill in the data preprocessing module according to the directory where the compressed package is located to decompress and obtain the corresponding documents, and perform corresponding format conversion on different documents according to the requirements;

[0065] The data analysis module is used to obtain the technical specifications, quotation sheets and demand plan summary sheets after decompression and format conversion by the data preprocessing module, and conduct reviews according to the requirements of different fields, associate the demand plan, technical specifications, quotation sheets and other relevant information through the technical specifications ID number, and conduct a comprehensive and multi-dimensional review using preset criteria;

[0066] The data output module is used to classify and sort out the opinions according to ERP business opinions and ECP technical opinions according to the judgment opinions of the procurement plan after the data analysis module completes the review of all plans, and sort them in the review opinion table.

[0067] Compared with the prior art, the beneficial effects of the present invention include at least:

[0068] 1. The present invention starts from the actual difficulties and pain points of the engineering service plan review, is developed in Python, uses Pandas to process the table data, and conducts a comprehensive and multi-dimensional review of the engineering service procurement plan of each reporting department, starting from the ERP demand plan, ECP technical specification, and quotation sheet. It uses preset criteria to review each plan in a comprehensive and multi-dimensional manner. It only takes a few seconds to review each plan, which is much faster than manual review. The review scale is unified and the accuracy is high. It also judges some simple association logics in multiple fields in ERP and ECP, and automatically gives review opinions. For some more complex logics and descriptions, it can provide experts with reference opinions for further review.

[0069] 2. The cross-departmental data integration algorithm is applicable to data obtained from different departments, different systems or different sources. Its core purpose is to ensure the consistency of data during the integration process and reduce the deviation caused by data differences between departments. In different scenarios such as demand plan aggregation, financial aggregation, inventory integration, etc., this algorithm can help enterprises improve the accuracy and consistency of data, thereby providing reliable support for decision-making;

[0070] 3. By using the Levenshtein distance algorithm to determine the similarity between the names of bidding project sections and the names of projects that have been included in the provincial bidding and procurement scope, combined with the authorized procurement logo, the system can automatically and accurately identify potential duplications or naming errors, reducing the burden of manual review, improving review efficiency, and providing experts with clear review guidance, ensuring the efficiency and accuracy of the entire process. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is a flow chart of a method for intelligent review of engineering service plans provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The described embodiments are only embodiments of a part of the present invention, not all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0073] like Figure 1 As shown, embodiment 1 of the present invention provides a method for intelligent review of engineering service plans, comprising the following steps:

[0074] S1: In the ERP system, the plans of each department are summarized and the ERP demand plan summary table is exported and uploaded to the data preprocessing module. At the same time, the technical specifications and quotation sheets are uploaded to the ECP system, and the technical specification ID number is generated and filled into the demand plan summary table. According to the ID number, the ECP technical specification and ECP quotation sheet documents are downloaded and exported in batches to generate a ZIP compressed package.

[0075] Further preferably, step S1 specifically includes:

[0076] S1.1: The plans reported by each department in the planning module of the ERP platform are summarized using a cross-departmental data integration algorithm to generate a demand plan summary table, which is then imported into the data preprocessing module to complete the import of the demand plan.

[0077] In the embodiment of the present invention, the planning personnel of each department prepare the engineering service plan in the enterprise resource management system (ERP) planning module, and the plan summary review personnel can summarize the plans reported by each department on the ERP platform, export the demand plan summary table, and then import this table into the data preprocessing module to complete the import of the demand plan.

[0078] Further preferably, step S1.1 specifically includes:

[0079] S1.1.1: Obtain the demand planning data reported by each department in the planning module of the ERP platform, including department ID, product category, demand quantity and time period. The ERP system collects and summarizes the demand planning data by using database access;

[0080] In the implementation of the present invention, database access refers to obtaining the plan data reported by each department directly through SQL query based on the known ERP system database, or obtaining it through API interface. Most modern ERP systems support RESTful API or SOAP API, and the data reported by the department can be obtained in real time through API interface.

[0081] S1.1.2: Synchronize the demand planning data of all departments obtained in step S1.1.1 to a unified data storage location through the database data replication tool, and clean the demand planning data. The cleaning operation includes deduplication, standardization and removal of outliers, and use a cross-departmental data integration algorithm to coordinate the consistency of data reported by different departments;

[0082] The specific workflow of the cross-departmental data integration algorithm is as follows:

[0083] According to the demand planning data reported by each department X i , and confirm the weight coefficient w based on the comprehensive evaluation of each department i , analyze and obtain the weighted average of the demand planning data of all departments as the target weight mean μ w ;

[0084] The weighted average formula:

[0085]

[0086] Where:

[0087] w i represents the weight of the i-th department;

[0088] X i Represents the demand planning data of the i-th department;

[0089] n represents the total number of departments that have uploaded demand planning data.

[0090] For each department's demand planning data X i , the square root index difference metric is used to evaluate the data and the target weight mean μ w The deviation value d i , the formula is:

[0091]

[0092] Where:

[0093] e is the base of natural logarithms.

[0094] If the deviation value d i A large value indicates that the data source of the department is significantly different from the expected weighted average and data adjustment may be required;

[0095] For each department's demand planning data X i , if d i If it is greater than the preset threshold, the stepwise approximation method is used to adjust the data to obtain the adjusted demand plan data. The formula is:

[0096]

[0097] Where:

[0098] represents the demand planning data after adjustment of the demand planning data of the i-th department;

[0099] η represents the adjustment step size parameter;

[0100] After the demand planning data of all departments are adjusted, the updated demand planning data will be synchronized back to all departments, and then the deviation value of the demand planning data of all departments will be calculated again to ensure that it does not exceed the preset threshold, and the consistency of the adjusted demand planning data will be confirmed.

[0101] The weight of each department's data may be different in different scenarios. For example, the weight can be set based on the quality of the department's historical data and its business influence. Some key departments such as finance or core business departments may have higher weights. By weighted averaging, the demand data of different departments can be aggregated into a comprehensive data.

[0102] The cross-departmental data integration algorithm is applicable to data obtained from different departments, different systems or different sources. Its core purpose is to ensure the consistency of data during the integration process and reduce the deviation caused by data differences between departments. In different scenarios such as demand planning aggregation, financial aggregation, inventory integration, etc., this algorithm can help enterprises improve the accuracy and consistency of data, thereby providing reliable support for decision-making.

[0103] S1.1.3: Summarize the demand plan adjusted by the data in step S1.1.2, generate an ERP demand plan summary table, export the Excel table data, and upload it to the data preprocessing module.

[0104] S1.2: Each department needs to upload its own technical specifications and quotation sheets to the ECP system, generate a technical specification ID number and fill it in the ERP demand plan summary table in step S1, and import the ERP demand plan summary table with the added ID number field into the data preprocessing module again.

[0105] In the implementation of the present invention, the planners of each department also need to upload their own technical specifications and quotation sheets to the State Grid Electronic Commerce Platform (ECP) system, generate a technical specification ID number and fill it into the ERP demand plan table corresponding to the plan they reported.

[0106] S1.3: The technical specification ID number corresponds to the plan reported in the ERP demand plan summary table. According to the technical specification ID number, the ECP technical specification and ECP quotation table documents are downloaded in batches in the ECP system and packaged and exported as a ZIP compressed package.

[0107] In the embodiment of the present invention, the plan summary auditor can download and export these documents in batches according to the technical specification ID number, and the exported document is a ZIP compressed package. At the same time, the ERP system refers to the enterprise resource management system, and the ECP system refers to the State Grid e-commerce platform. The combination of the ERP system and the ECP system can realize the comprehensive integration of internal enterprise resources and the effective coordination of the external supply chain, thereby enhancing the operational efficiency and supply chain management capabilities of the enterprise. ECP is mainly used for electronic procurement, supplier management, technical document upload and other functions, while the ERP system covers the core business processes of the enterprise such as finance, production, inventory, and sales. The core purpose of the combination of the two is to make the entire process from suppliers to internal operations of the enterprise smoother and more efficient through data sharing and information flow.

[0108] S2: Fill the directory where the compressed package obtained in step S1.3 is located into the data preprocessing module for decompression to obtain the corresponding documents, and perform corresponding format conversion on different documents as required.

[0109] Further preferably, step S2 specifically includes:

[0110] S2.1: Fill the directory where the compressed package is located into the data preprocessing module. The data preprocessing module decompresses the compressed package in multiple folders under the directory, and automatically identifies and extracts the technical specifications and quotation sheets, and matches them with the demand plan summary sheet imported again in step S1.2.

[0111] S2.2: For documents uploaded in WPS and DOC formats, the data preprocessing module converts them into docx format; for tables uploaded in xls format, the data preprocessing module converts them into xlsx format

[0112] S3: Based on the technical specifications and quotation sheets obtained after decompression and format conversion in step S2.2 and the demand plan summary sheet imported again in S1.2, they are reviewed according to the requirements of different fields, and the demand plan, technical specifications, quotation sheets and other relevant information are associated through the technical specification ID number, and a comprehensive and multi-dimensional review is performed using preset criteria.

[0113] Further preferably, step S3 specifically includes:

[0114] S3.1: The pandas library is used to read and operate the ERP demand plan summary table in xlsx format that is re-imported in S1.2. After removing blank fields, data cleaning and format conversion are performed, and multi-dimensional verification and review are performed according to the requirements of different fields.

[0115] Further preferably, step S3.1 specifically includes:

[0116] S3.1.1: Regular expression matching is used to determine whether the bidding project section names of the demand plan meet the naming specification requirements. For example, the bidding project section names must be named in accordance with the requirements and in the format of "company abbreviation + department (optional) + year + procurement content + service";

[0117] Use string suffix matching to determine whether it corresponds to the content description of the purchase scope of the reported entry. For example, the entry for manufacturer technical service needs to end with "manufacturer technical service", the entry for rental service needs to end with "rental service", and the framework project needs to contain the byte "framework project";

[0118] Keywords are used in the bidding project section name to check whether it contains the description of the service master data characteristic value; for the above fields that are not filled in as required, the software will directly give the review opinion;

[0119] The Levenshtein distance was used to calculate the similarity to determine whether the bidding project section name was similar to the project name that had been included in the provincial bidding and procurement scope, and then the authorization for procurement was submitted, and the opinion of "further verification by experts" was given;

[0120] The character string similarity calculation specifically includes:

[0121] The Levenshtein distance algorithm is used to calculate the similarity between the bidding project section name and the project name within the provincial bidding and procurement scope; the Levenshtein distance can measure the minimum number of editing operations between two strings, including insertion, deletion and replacement;

[0122] The recursive relationship of Levenshtein distance is as follows:

[0123]

[0124] in:

[0125] D(i,j) represents the minimum edit distance to convert string A[i] into string B[j];

[0126] A[i] represents the first i characters of string A;

[0127] B[j] represents the first j characters of string B

[0128] A represents the name string A of the bidding project section;

[0129] B represents the project name string B that has been included in the provincial bidding and procurement scope;

[0130] D(i-1,j)+1 represents the deletion operation, that is, deleting a character from string A;

[0131] D(i,j-1)+1 represents an insertion operation, that is, inserting a character into string A;

[0132] cost(A[i],B[j]) indicates whether character A[i] and character B[j] are the same. If they are the same, the cost is 0; if they are different, the cost is 1;

[0133] If the distance is less than the preset threshold, the similarity is judged to be high and the opinion of "further verification by experts" is given.

[0134] In the embodiment of the present invention, it is determined whether the content description of the reported procurement scope corresponds to that of the reported requisition. For example, the manufacturer's technical service entry needs to end with "manufacturer's technical service", the rental service needs to end with "rental service", and the framework project needs to contain the "framework project" byte. If it does not meet the review opinion, it may be "not in compliance with the requirements"; in the process of project submission, you may encounter situations similar to project names that have been included in the provincial procurement scope, especially when the plan involves authorized procurement, the system needs to make more detailed judgments to avoid repeated or unnecessary procurement. In summary, the final review opinion will be given, such as "the section name complies with the specifications, the content is consistent, and the service characteristic value is correct; due to the similarity with the project name within the provincial procurement scope, it is recommended that experts conduct further verification."

[0135] S3.1.2: Use date format conversion to judge the planned start date and duration, including whether it is earlier than the start date of this batch, whether the filled duration is reasonable, and whether the cost project crosses the year, and associate it with other fields. If the duration is less than 365 days, there should be no description of every day in the technical specification, bidding scope, and project overview. Use the string search method to check whether the text contains the word "every day". For fields that are not filled in as required, the software directly gives review opinions.

[0136] The association judgment with other fields described in the examples of the present invention means that in addition to independent judgments on fields such as the planned start date and construction period, it is also necessary to consider the association between these fields and other related fields including the technical specifications, bidding scope, project overview, etc. For example, when the construction period is less than 365 days, it is necessary not only to verify whether the value of the construction period field itself meets expectations, but also to check whether other fields related to the construction period, including the technical specifications, bidding scope and project overview, have a description of "every day". If there is a description of "every day", the system will automatically give an audit opinion: "The construction period is less than 365 days, and there should be no description of "every day" in the technical specifications, bidding scope or project overview. Please modify it."

[0137] S3.1.3: Judge the bidding amount. If the bidding amount of the authorized purchase is greater than the authorized limit, the software will give an audit opinion that the amount exceeds the limit; if the bidding amount is less than 10,000 yuan, the software will give an audit opinion that the amount is too small and needs to be verified to avoid errors in the order of magnitude of the procurement amount. For the estimated amount and the bidding amount, the rate association judgment is made. If the bidding amount divided by the estimated amount is not equal to 1 or 1.03 or 1.06 or 1.09 or 1.13, it will prompt "the ratio of the bidding amount to the estimated amount of this procurement plan is wrong", and the software will directly give an audit opinion and reject it. For procurement plans that have selected the authorized limit terms, when the cumulative procurement amount exceeds the limit, the software will link the historical procurement data and give the cumulative amount exceeds the limit. For the same procurement plan, or plans with a high degree of similarity in the bidding project section name, if the difference in the bidding amount ratio is plus or minus 20%, it will prompt "the bidding amount of this procurement plan is significantly different from similar projects in historical batches", and the opinions of experts need to be further verified.

[0138] S3.1.4: Judge the scope of the tender, the scale of the project and the overview. Use special characters to detect whether there are illegal characters and garbled characters in the text. Check whether the fields are omitted or incomplete through pattern matching according to the set template. Use regular expressions to extract year information to check whether the wrong year is filled in. Associate the content in the tender documents with predefined terms and judge. Use natural language processing technology to identify whether there are inappropriate descriptions in the content of the tender documents. The software directly gives the review opinion. If it involves the negative list of business outsourcing, it will give the specific text content of the negative list of business outsourcing, and put forward opinions that require further verification by experts.

[0139] S3.1.5: Judge the contractor's requirements, including whether the filling meets the requirements of the solidified template, whether the latest contract number is filled in, and whether the contract revision content is inconsistent with the contracting method. For fields that are not filled in as required, the software directly gives review opinions.

[0140] In the embodiment of the present invention, the system checks whether there is a contradiction between the contract revision content and the contracting method. If there is a conflict, the system will generate an audit opinion and prompt that it needs to be modified. For example, for a lump sum contracting project, there should be no requirement for fixed unit price settlement.

[0141] S3.1.6: Check the duplication of the purchase application number. For plans submitted in the general plan mode, remind the experts to verify whether the plan is a project that the system does not support. Check whether the technical specification ID has adopted the technical specification that has been submitted before. Check whether the terms selected by ECP are consistent with ERP. For fields that are not filled in as required, the software directly gives review opinions.

[0142] S3.1.7: Determine the quotation method and contracting method, including whether the quotation method is consistent with the quotation table description in the ECP, whether the contracting method is consistent with the technical specification description in the ECP, and whether the basic project information fields are filled in according to the specification requirements. If it does not meet the requirements, the software will directly give an audit opinion.

[0143] The basic project information fields described in the embodiment of the present invention include project approval / feasibility study approval number, preliminary design approval number, design unit, supervision unit, construction location, submitter contact information, reason for selecting procurement method, proposed inviting unit, proposed inviting unit contact person and telephone number and other fields.

[0144] S3.2: After decompressing and converting the format in step S2.2, the technical specification and quotation sheet are obtained, and the key information of the technical specification is extracted through text parsing. The pandas library is used to read the quotation sheet and perform multi-dimensional verification and review according to the requirements of different fields.

[0145] Further preferably, step S3.2 specifically includes:

[0146] S3.2.1: For the solidified technical specifications, the technical specification fields in the demand plan table are judged. For example, if the procurement plan starting with G00L is prepared with a solidified template, it is judged to be a plan prepared with a solidified template. Then the text content of the reported technical specifications is compared with the solidified template built into the system. Exact matching is used to check whether the submitted content is completely in line with the solidified template, and whether it is prepared according to the solidified template. Fuzzy matching is used to detect the differences in spelling and format for some non-critical content, and the comparison results of the two are given. They are presented in two columns in the form of a web page, and the differences between the two are marked with colors, so that experts can quickly find the differences and improve the accuracy of the review. For some obvious errors, keyword detection is used to determine whether some key fields or content are missing. For example, if the content that needs to be replaced and modified in the solidified template has not been modified and is still "XX", the software will directly give the review opinion.

[0147] S3.2.2: For self-compiled technical specifications, determine whether the following errors exist in the written content, including that the laws and regulations in the document are not currently valid, that the template numbers of certain technical standards and contracts have not been changed to the latest versions, that descriptions that are inconsistent with the enterprise qualifications, personnel and performance requirements solidified in the terms of the procurement strategy, and that descriptions that are inconsistent with the bidding project section names, term feature values, construction periods, and contracting methods in the demand plan. If there are errors in the written content, the software will directly give a review opinion and reject it.

[0148] In the embodiment of the present invention, the judgment of whether the laws and regulations in the document are currently valid is based on the current "Civil Code of the People's Republic of China" rather than the "Contract Law". If the template numbers of certain technical standards and contracts are not changed to the latest version, it may cause disputes in the later contract signing and performance process.

[0149] S3.2.3: For some complex criteria, including whether some descriptions in the technical specifications involve the contents of the negative list of business outsourcing and whether the contents of the technical specifications of the authorized procurement plan involve the scope of provincial recruitment, if they do not meet the requirements, further verification by experts is required.

[0150] S3.2.4: Judge the quotation sheet, use Levenshtein distance fuzzy matching to check whether the filled-in project title is similar to the bidding project section name in the demand plan, and then judge whether it is consistent. Traverse the serial number column in the quotation sheet to check whether it is filled in in ascending order to determine whether the serial number column is filled in correctly, whether the quantity column is filled in with numbers, and whether the unit price and total price columns are empty. For the above fields that are not filled in as required, the software directly gives an audit opinion and rejects them. The measurement units used in the quotation sheet need to be further verified by experts, and remind experts to judge whether they are described in detail in the technical specifications.

[0151] The embodiments of the present invention use relatively rough measurement units including set, book, volume, item, etc., which need to be described in detail in the technical specification book.

[0152] S3.3: The demand plan, technical specification, quotation and other relevant information are linked through the technical specification ID number, and a comprehensive and multi-dimensional review is conducted using preset criteria.

[0153] The basis for judgment is mainly the key points of the provincial company's engineering service procurement plan review and the key points of the authorized procurement plan review, as well as the engineering service plan bidding documents issued by various professional departments. According to the requirements of these documents, we convert these judgment logics into codes that can be judged by the program, conduct pre-examination at the stage of plan preparation, conduct intelligent review at the summary stage of local municipal companies, and conduct intelligent batch review during the provincial company's expert review meeting. The key points of each stage of review will be richer than the previous stage. The data analysis module will review the plan according to different logical criteria at each stage and give the review opinions to the experts. For example, when judging the name of the bidding project section, when judging the demand plan, it is determined that its format meets the requirements, whether there are historical similar procurement projects, and whether there are repeated reporting plans in this batch. At the same time, the technical specification book ID in the demand plan is also associated with the bidding section name filled in the technical specification book and the quotation form file. That is, whether the title in the technical specification book is consistent, whether the title in the quotation form is consistent, and whether the file names of the two documents meet the specifications.

[0154] S4: After the data analysis module completes the review of all plans, it classifies and sorts these opinions according to the ERP business opinions and ECP technical opinions based on the judgment opinions of the procurement plans in steps S3.1 and S3.2, and places them in the review opinion table in order. The data output module generates the review opinion table and plan summary table.

[0155] After the data analysis module completes the review of all plans, the data output module will generate an audit opinion table. The content of the table mainly consists of two parts. One part is the business opinion for the business experts to review, which corresponds to the content that needs to be changed in the demand plan table. The other part is the technical opinion for the technical experts to review, which mainly corresponds to the content that needs to be modified in the technical specifications and quotation tables. The audit opinion table will also have shortcuts to open the technical specifications and quotation tables for the convenience of experts. For plans reported by templates, template comparison results will also be provided. At the same time, in the plan review table, a one-click link is set for experts to view the technical specifications, quotation tables, and historical procurement data tables, so that the review experts can use their time and energy on the "cutting edge" and complete the plan review more efficiently.

[0156] Embodiment 2 of the present invention provides a system for intelligent review of engineering service plans and a method for intelligent review of engineering service plans described in embodiment 1, including: a data preprocessing module, a data analysis module and a data output module;

[0157] Data preprocessing module, data analysis module and data output module;

[0158] The data preprocessing module is used to receive the ERP demand plan summary table and fill in the data preprocessing module according to the directory where the compressed package is located to decompress and obtain the corresponding documents, and perform corresponding format conversion on different documents according to the requirements;

[0159] The data analysis module is used to obtain the technical specifications, quotation sheets and demand plan summary sheets after decompression and format conversion by the data preprocessing module, and conduct reviews according to the requirements of different fields, associate the demand plan, technical specifications, quotation sheets and other relevant information through the technical specifications ID number, and conduct a comprehensive and multi-dimensional review using preset criteria;

[0160] The data output module is used to classify and sort out the opinions according to ERP business opinions and ECP technical opinions according to the judgment opinions of the procurement plan after the data analysis module completes the review of all plans, and sort them in the review opinion table.

[0161] Compared with the prior art, the beneficial effects of the present invention include at least:

[0162] 1. The present invention starts from the actual difficulties and pain points of the engineering service plan review, is developed in Python, uses Pandas to process the table data, and conducts a comprehensive and multi-dimensional review of the engineering service procurement plan of each reporting department, starting from the ERP demand plan, ECP technical specification, and quotation sheet. It uses preset criteria to review each plan in a comprehensive and multi-dimensional manner. It only takes a few seconds to review each plan, which is much faster than manual review. The review scale is unified and the accuracy is high. It also judges some simple association logics in multiple fields in ERP and ECP, and automatically gives review opinions. For some more complex logics and descriptions, it can provide experts with reference opinions for further review.

[0163] 2. The cross-departmental data integration algorithm is applicable to data obtained from different departments, different systems or different sources. Its core purpose is to ensure the consistency of data during the integration process and reduce the deviation caused by data differences between departments. In different scenarios such as demand plan aggregation, financial aggregation, inventory integration, etc., this algorithm can help enterprises improve the accuracy and consistency of data, thereby providing reliable support for decision-making;

[0164] 3. By using the Levenshtein distance algorithm to determine the similarity between the names of bidding project sections and the names of projects that have been included in the provincial bidding and procurement scope, combined with the authorized procurement logo, the system can automatically and accurately identify potential duplications or naming errors, reducing the burden of manual review, improving review efficiency, and providing experts with clear review guidance, ensuring the efficiency and accuracy of the entire process.

[0165] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for intelligent review of engineering service plans, characterized in that: include: S1: In the ERP system, the plans of each department are summarized and the ERP demand plan summary table is exported and uploaded to the data preprocessing module. The technical specifications and quotation table are uploaded to the ECP system, and the technical specification ID number is generated and filled into the demand plan summary table. According to the ID number, the ECP technical specification and ECP quotation table documents are downloaded and exported in batches to generate a ZIP compressed package; S2: Fill the directory where the compressed package obtained in step S1 is located into the data preprocessing module for decompression to obtain corresponding documents, and perform corresponding format conversion on different documents as required; S3: After decompressing the format and converting it in step S2, the technical specification and quotation sheet are obtained, and the demand plan summary sheet is imported again for review according to the requirements of different fields. The demand plan, technical specification, quotation sheet and other related information are associated through the technical specification ID number, and a comprehensive and multi-dimensional review is performed using the preset criteria; S4: After the data analysis module completes the review of all plans, it will classify and sort these opinions according to the ERP business opinions and ECP technical opinions based on the judgment opinions of the procurement plan, and place them in the review opinion table in order. The data output module will generate the review opinion table and plan summary table.

2. According to claim 1, a method for intelligent review of engineering service plans is characterized by: Obtain the demand planning data reported by each department in the planning module of the ERP platform, including department ID, product category, demand quantity and time period. The ERP system collects and summarizes the demand planning data by using database access; The acquired demand planning data of all departments are synchronized to a unified data storage location through the database data replication tool, and the demand planning data is cleaned. The cleaning operation includes deduplication of data, standardization of data and removal of outliers. A cross-departmental data integration algorithm is used to coordinate the consistency of data reported by different departments.

3. The method for intelligent review of engineering service plans according to claim 2, characterized in that: The specific workflow of the cross-departmental data integration algorithm is as follows: According to the demand planning data reported by each department i , and confirm the weight coefficient w based on the comprehensive evaluation of each department i , analyze and obtain the weighted average of the demand planning data of all departments as the target weight mean μ w ; The weighted average formula: Where: w i represents the weight of the i-th department; X i Represents the demand planning data of the i-th department; n represents the total number of departments that have uploaded demand planning data; For each department's demand planning data X i , the square root index difference metric is used to evaluate the data and the target weight mean μ w The deviation value d i , the formula is: Where: e is the base of natural logarithms.

4. The method for intelligent review of engineering service plans according to claim 3 is characterized in that: If the deviation value d i A large value indicates that the data source of the department is significantly different from the expected weighted average and data adjustment may be required; For each department's demand planning data X i , if d i If it is greater than the preset threshold, the stepwise approximation method is used to adjust the data to obtain the adjusted demand plan data. The formula is: Where: represents the demand planning data after adjustment of the demand planning data of the i-th department; η represents the adjustment step size parameter.

5. The method for intelligent review of engineering service plans according to claim 1, characterized in that: The pandas library is used to read and operate the ERP demand plan summary table in xlsx format, and data cleaning and format conversion are performed after removing blank fields. Multi-dimensional verification and review are performed according to the requirements of different fields; After decompressing the format, the technical specifications and quotation sheets are converted to extract key information from the technical specifications through text parsing, and the pandas library is used to read the quotation sheets for multi-dimensional verification and review based on the requirements of different fields; The demand plan, technical specification, quotation and other relevant information are linked through the technical specification ID number, and a comprehensive and multi-dimensional review is conducted using preset criteria.

6. The method for intelligent review of engineering service plans according to claim 5, characterized in that: The Levenshtein distance is used to calculate the similarity to determine whether the names of the bidding projects in the ERP demand plan summary table are similar to the names of the projects that have been included in the provincial bidding and procurement scope, and then the authorization for procurement is submitted, and the opinions that further verification by experts are given; The character string similarity calculation specifically includes: The Levenshtein distance algorithm is used to calculate the similarity between the bidding project section name and the project name within the provincial bidding and procurement scope; the Levenshtein distance can measure the minimum number of editing operations between two strings, including insertion, deletion and replacement; The recursive relationship of Levenshtein distance is as follows: in: D(i,j) represents the minimum edit distance to convert string A[i] into string B[j]; A[i] represents the first i characters of string A; B[j] represents the first j characters of string B A represents the name string A of the bidding project section; B represents the project name string B that has been included in the provincial bidding and procurement scope; D(i-1,j)+1 represents the deletion operation, that is, deleting a character from string A; D(i,j-1)+1 represents an insertion operation, that is, inserting a character into string A; cost(A[i],B[j]) indicates whether character A[i] and character B[j] are the same. If they are the same, the cost is 0; if they are different, the cost is 1; If the distance is less than the preset threshold, the similarity is judged to be high and an opinion is given that further verification by experts is required.

7. The method for intelligent review of engineering service plans according to claim 5, characterized in that: The bidding scope, project scale and overview of the ERP demand plan summary table are judged, special characters are used to detect whether there are illegal characters and garbled characters in the text, and pattern matching is used to check whether the fields are omitted or incomplete according to the set template. Regular expressions are used to extract year information to check whether the wrong year is filled in. The content in the bidding documents is associated with predefined terms and judged, and natural language processing technology is used to identify whether there are inappropriate descriptions in the content of the bidding documents. The software directly gives review opinions. If it involves the negative list of business outsourcing, the specific text content of the negative list of business outsourcing is given, and opinions that require further verification by experts are put forward.

8. The method for intelligent review of engineering service plans according to claim 5, characterized in that: Compare the text content of the technical specification with the built-in fixed template of the system, use exact matching to check whether the submitted content is in full compliance with the fixed template, and determine whether it is compiled according to the fixed template. Use fuzzy matching to detect spelling and format differences for non-critical content, and use keyword detection to determine whether key fields and content are missing, including content that needs to be replaced and modified in the fixed template but has not been modified. The software directly gives review opinions.

9. The method for intelligent review of engineering service plans according to claim 5, characterized in that: The quotation sheet is judged by using Levenshtein distance fuzzy matching to check whether the filled-in project title is similar to the bidding project section name in the demand plan, and then determine whether they are consistent. The serial number column in the quotation sheet is traversed to check whether it is filled in in ascending order to determine whether the serial number column is filled in correctly.

10. A system for intelligent review of engineering service plans, running a method for intelligent review of engineering service plans as described in any one of claims 1 to 9, characterized in that: include: Data preprocessing module, data analysis module and data output module; The data preprocessing module is used to receive the ERP demand plan summary table and fill in the data preprocessing module according to the directory where the compressed package is located to decompress and obtain the corresponding documents, and perform corresponding format conversion on different documents according to the requirements; The data analysis module is used to obtain the technical specifications, quotation sheets and demand plan summary sheets after decompression and format conversion by the data preprocessing module, and conduct reviews according to the requirements of different fields, associate the demand plan, technical specifications, quotation sheets and other relevant information through the technical specifications ID number, and conduct a comprehensive and multi-dimensional review using preset criteria; The data output module is used to classify and sort out the opinions according to ERP business opinions and ECP technical opinions according to the judgment opinions of the procurement plan after the data analysis module completes the review of all plans, and sort them and place them in the review opinion table. The data output module generates the review opinion table and the plan summary table.

Images