Method for extracting, positioning and analyzing content of tender information of bidding document based on OCR (Optical Character Recognition) engine

Abstract

The invention provides an OCR (optical character recognition) engine-based bidding file flow bid information content extraction, positioning and analysis method. The method comprises the following steps of collection of bidding files: collecting document files participating in a bidding project; file analysis: aiming at different types of document files, adopting corresponding information extraction processing methods to obtain key information; bid invitation file flow bid item detection: bid invitation factor standard information in the bid invitation file is extracted, and bid invitation file flow bid item factors are positioned and extracted; the structure and the text of the bidding document are analyzed, the directory structure of the bidding document is extracted, and the flow bidding item of the bidding document is positioned; stream bid item analysis: analyzing stream bid items in the bid invitation file; and bidding document flow bidding item prompting: according to the comparison result of the flow bidding items, carrying out inconformity item prompting. The precision and accuracy of identifying the content of the scanned PDF document are improved, the workload of manual proofreading and correction is reduced, and the working efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bidding information processing, and in particular to a method for extracting, locating and analyzing bid failure information content in bidding documents based on an OCR engine. Background Art

[0002] During the bidding process, extracting and analyzing bid documents for failed bids (i.e., bids that were not awarded or did not meet the bidding requirements) is crucial for optimizing companies' bidding strategies and improving their bid success rates. Traditional manual review methods are inefficient and prone to errors. With the advancement of information technology, the use of optical character recognition (OCR) technology to automate the processing of bid documents has become a trend. However, existing OCR technology often suffers from low recognition accuracy and incomplete information extraction when processing complex bid documents. Furthermore, existing technologies lack effective solutions for extracting and analyzing specific content from failed bid information. Summary of the Invention

[0003] The present invention aims to provide an efficient and accurate method for extracting, locating and analyzing the content of bid failure information in bidding documents based on an OCR engine, aiming to achieve efficient and accurate extraction of text content in bidding documents, and to be able to reversely locate the specific position in the bidding document based on the extracted content and effectively analyze the extracted content, thereby realizing the extraction, positioning and analysis of the content of bid failure information.

[0004] In order to achieve the above technical features, the purpose of the present invention is to achieve the following: a method for extracting, locating and analyzing the content of bid failure information in bidding documents based on an OCR engine, comprising the following steps: Step 1: Procurement document collection: Collect documents and files for bidding projects; Step 2, file parsing: For different types of document files, corresponding information extraction and processing methods are used to obtain key information; Step 3: Check if the bidding documents have failed: Extract standard information on bidding factors from bidding documents, and locate and extract factors that lead to unsuccessful bidding in bidding documents; Step 4: Positioning of failed bidding items in the bidding documents: Analyze the structure and text of the bidding documents, extract the directory structure of the bidding documents, and locate the bidding items in the bidding documents; Step 5: Flow item analysis: Analyze the failed bid items in the bidding documents; Step 6, prompt for bidding document failure: Prompt for non-compliant items based on the comparison results of the unsuccessful bid items.

[0005] Preferably, the document files collected in step 1 for the bidding project specifically include: Collect document files involved in the bidding project from the procurement system, including: bidding document procurement, bid documents, bidding announcement documents and bidding projects.

[0006] Preferably, in step 2, different types of document files are divided into two categories: scanned PDF files and non-scanned PDF files. Different information extraction processing methods are performed based on the bids of different types of document files to obtain key information; For scanned PDF files, each page will be identified as an image. The image will be processed first, and different information extraction and processing methods will be called according to the different content of each page. At the same time, the content of the bid includes tables, drawings and technical documents, and different processing will be performed for different contents.

[0007] Preferably, the method for extracting information from a table in a scanned PDF file specifically includes: For scanned pages whose image content is a table, a table OCR tool is used for processing; first, the image is corrected, then the image is segmented to obtain the table title and table range, and then the table content in the table range is extracted; during the extraction process: first, the table structure in the image is identified, and the horizontal and vertical lines of the table are obtained based on the recognition results. According to these lines, the original image is divided into multiple independent rectangles, and each rectangle is a cell of the table; then each identified cell is processed, first determining whether the center point of each identified character is in the cell, and if so, the character is added to the character list of the corresponding cell; after the traversal is completed, the data list of each cell is integrated into a table data structure for storage, which is used for subsequent bid processing and analysis.

[0008] Preferably, the method for extracting information from drawings in a scanned PDF file specifically includes: The attached drawings are divided into license type and non-license type. The scanned copies of license types are compared with the pre-text to verify the authenticity and relevant qualifications of the bidder. At the same time, the bidder name, license name or permit name on the license are extracted, and the official seal on the license is located and extracted; Among them, name extraction uses OCR recognition to obtain all the text on the license image, and then common watermark text is removed through regular expressions; the text is matched with common license keywords for similarity to obtain the name in the license text; the official seal in the image is extracted using Textmind, and the extracted license name, bidder name and official seal are stored together for subsequent bidder qualification verification and technical factor review.

[0009] Preferably, the method for extracting information from a scanned PDF document specifically includes: First, split the long text and remove some stop words in the sentences. Split the large text into several short sentences, using full stops, semicolons, and commas as boundaries. For the short sentences obtained from the split, remove the punctuation marks such as quotation marks, book titles, and parentheses, and remove the stop words such as "in order to" and "thus" that have little impact on the original sentence meaning; then, perform similarity matching between the processed short sentences of the technical document and the tender factor keywords respectively, and take the short sentences with a similarity greater than 85% as auxiliary positioning points; display these positioning points in association with the tender factors, serving as an auxiliary bid evaluation system to improve the efficiency of expert review.

[0010] Preferably, it also includes identifying the seal signature. Use seal detection and recognition technology to detect the seals on the page and recognize the seal content, obtaining the seal position and text on each page.

[0011] Preferably, the specific steps in step 3 include: Extract the tender factor standard information in the tender document using similarity matching technology; map the tender document bid failure items to common directory headings. After mapping the abstract tender document bid failure items and standards to the directly related headings or upper-level headings, directly match these headings in the directory to improve the accuracy of similarity matching; then use similarity matching technology to locate and extract the bid failure item factors.

[0012] Preferably, the specific steps in step 4 include: For the tender documents with a directory structure, directly split them; For the tender documents without a directory structure, quickly extract the directory structure from the full text through specification judgment or font size judgment for the splitting of the tender documents and the positioning of tender factors; By locating the financial statements in the qualification review materials, during the positioning process, extract all the pictures in the qualification review materials and save them in page number order to prevent disorder. Then perform OCR recognition on each picture, and match the keywords related to the financial statements in the recognized text. Starting from the first keywords that appear: "audit report", "financial statements" as the starting positions, there will be two financial statements in the qualification review materials, and there will be two recognition results. The positioning interval of the first financial statement is from the first positioning result to the second positioning result, and the end position of the second statement directly adopts the position starting from the next heading.

[0013] Preferably, the specific steps in step 5 include: Analyze the unsuccessful bids in the bidding documents, including the bidder's name, signature and seal of the bid letter, unique quotation, business license, production safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, reputation requirements, bid content, construction period, project quality, bid validity period, bid bond, rights and obligations, and priced bill of quantities. For common failed bid items, we use the positioning text content in the bidding documents and the failed bid item standards in the tender documents to construct prompt words, call the large model to analyze the failed bid items, and conduct specific analysis on some failed bid items with clear requirements; For financial requirements, we use optimized table recognition technology to identify and extract the balance sheet, income statement, and cash flow statement of the financial statements in the bidding documents; For skewed or partially distorted tables, the commonly used extraction model is modified, and the report image is input. The extraction result is output after passing through three modules: text detection, detection regularization, and text recognition. The text detection first locates the location of the text and segments it, then divides the text block into square detection areas through regularization and performs data correspondence. Finally, text recognition is performed on each detection area, and the report result is output.

[0014] Preferably, the step 6 specifically includes: Non-conformity prompts will be given based on the comparison results of the bidder's name, signature and seal of the bid letter, unique quotation, business license, production safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, credibility requirements, bid content, construction period, project quality, bid validity period, bid deposit, rights and obligations, and unsuccessful items in the priced bill of quantities.

[0015] The present invention has the following beneficial effects: 1. The present invention improves the precision and accuracy of content recognition of scanned PDF documents, reduces the workload of manual proofreading and correction, and improves work efficiency.

[0016] 2. The present invention realizes accurate reverse positioning of text content in the original document, which facilitates users to quickly find the required information and improves the efficiency and accuracy of information retrieval.

[0017] 3. The present invention performs structured analysis on text content, making the extracted text easier to understand and process, and providing users with a better usage experience.

[0018] 4. The present invention designs a set of efficient processing procedures to achieve rapid and accurate extraction, positioning and analysis of bidding document failure information content.

[0019] 5. The present invention has high practical value and economic benefits and can be applied in the field of bidding and procurement. In addition, it can also be extended to the fields of information processing, document management, digital library, office automation, etc., providing strong support for the processing and utilization of digital information. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be further described below with reference to the accompanying drawings and examples.

[0021] Figure 1 This is a module diagram of the method for extracting, locating and analyzing the content of bid failure information in bidding documents based on the OCR engine of the present invention.

[0022] Figure 2 This is a flow chart of the method for extracting, locating and analyzing the content of bid failure information in bidding documents based on the OCR engine of the present invention.

[0023] Figure 3 This is an example of a table in a bid document during the file parsing process of the present invention.

[0024] Figure 4 This is a structural diagram of the model used in the present invention to extract skewed and partially distorted tables during the analysis of unsuccessful bid items.

[0025] Figure 5 This is a diagram of the regularized detection process of text detection in the process of flow item analysis of the present invention. DETAILED DESCRIPTION

[0026] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0027] Example 1: See also Figure 1-2 The present invention provides a method for extracting, locating, and analyzing the content of bid failure information in a bidding document based on an OCR engine, comprising the following steps: Step 1: Procurement document collection: Collect information data of documents involved in bidding projects from the procurement system. Document information data includes but is not limited to: bidding documents, tender documents, bidding announcement documents, bidding projects, etc.

[0028] Step 2, file parsing: Bidding documents are divided into two categories: scanned PDF files and non-scanned PDF files. Due to the difference in file content, these two types of bids must undergo different information extraction processes to obtain key information.

[0029] Next, each page of the scanned PDF file is recognized as an image, so the image must be processed first. Different processing methods are required depending on the content of each page. Bids often include tables, drawings, and technical documentation, which require different processing methods.

[0030] See also Figure 3 For an example of a table in a bidding document, next, we process the table.

[0031] For scanned pages with table contents (such as Figure 3 As shown), it is processed using a table OCR tool. First, some image correction is required, and then the image is segmented. The table page in the bid generally contains the table title and table content. By segmenting the title and table range, the table content can be extracted more accurately. Table extraction will first identify the table structure in the image, and obtain the horizontal and vertical lines of the table based on the recognition results. The original image is divided into rectangles based on these lines, and each rectangle is a cell of the table. Each recognized cell is then processed. First, it is determined whether the center point of each recognized character is in the cell. If so, the character is added to the character list of the corresponding cell. After the traversal is completed, the data list of each cell is integrated into a table data structure for storage for subsequent bid processing and analysis.

[0032] Next, the drawings are processed.

[0033] The accompanying drawings can be divided into license type and non-license type. Among them, scanned copies of license type usually need to be compared with the pre-text to verify the authenticity and relevant qualifications of the bidder. The license needs to extract the bidder's name, patent name or license name, etc., and the official seal on the license needs to be located and extracted. Name extraction uses OCR recognition to obtain all the text on the license image, and then uses regular expressions to remove common watermark text. The text is matched with the keywords of common licenses for similarity to obtain the name in the license text. The official seal in the image is extracted using Textmind, and the extracted license name, bidder name, etc. are stored together with the official seal for subsequent bidder qualification verification, technical factor review, etc.

[0034] Next, the technical documentation is analyzed.

[0035] Tenders usually contain medium and long technical documents to describe the technical implementation details of the tender project. Technical documents are the key in the bid evaluation factors. By locating the key information in the technical documents, it can assist the evaluation experts to complete the tender evaluation more quickly. Since the long documents involve a large amount of text, the accuracy of directly performing text summarization or keyword extraction is relatively low and it is difficult to meet the actual needs. In order to further improve the accuracy of location, first, the long text is segmented, and some stop words in the sentences are removed, and the large paragraph of text is segmented into several short sentences. The text segmentation uses full stops, semicolons and commas as boundaries. For the short sentences obtained by segmentation, punctuation marks such as quotation marks, book titles and brackets are removed, and stop words such as "in order to" and "thus" that have little impact on the original sentence meaning are removed. The processed short sentences of the technical documents are respectively matched with the keywords of the tender factors, and the short sentences with a similarity greater than 85% are taken as auxiliary location points. These location points are associated and displayed with the tender factors as an auxiliary bid evaluation system to improve the efficiency of expert evaluation.

[0036] Next, identify the seal signatures.

[0037] Use seal detection and recognition technology to detect the seals on the page and recognize the seal content, obtaining the seal positions and texts of each page. Such as the seal of the bidding company on the bidding document, Shandong xxx Co., Ltd.; the seal of the auditing company in the financial statements, Shandong xxx Firm; the seal of the Party A company in the performance contract, Shandong xxx, etc.

[0038] Step 3, Detection of the non-winning items in the tender documents: Adopt similarity matching technology to extract the standard information of the tender factors in the tender documents. Map the non-winning items in the tender documents to common catalog titles. As shown in Table 1, after mapping the abstract non-winning items and standards in the tender documents to the directly related titles or superior titles, these titles can be directly matched in the catalog, which can greatly improve the accuracy of similarity matching.

[0039] Table 1 Mapping Table of Non-winning Items in the Tender Documents of a Certain Tender Project

[0040] Next, use similarity matching technology to locate and extract the non-winning item factors. Such as each page needs to have the seal of the bidder, the financial statements need to have the seal of the firm, the quotation list needs to have a seal, and the details of the quotation list in the bidding document are consistent with the quotation list in the tender document, etc.

[0041] Step 4, Location of the non-winning items in the bidding documents: Most bid documents include a table of contents structure, which can be directly used for bid segmentation. However, some scanned PDFs lack an outline structure, and some non-scanned PDFs may also lack an outline. Analyzing the structure and text of bid documents reveals that all bids contain a heading, which typically follows a fixed pattern or display style. For example, first-level section headings typically begin with a sequence number like "1.", while second-level subheadings are typically displayed in a larger font and bolder than the main text. Using rule-based or font size analysis, the table of contents structure can be quickly extracted from the full text, enabling bid segmentation and bid factor identification. For example, financial statements provided in bid documents are typically compiled by year. Bids often require the submission of the last two years' financial statements. Therefore, it is necessary to search the entire bid document for the last two years' financial statements, locate them, and extract them based on the current year. Most bid documents insert a heading at the beginning of the financial statements, indicating that the next section will contain the financial statements for a specific year. However, some bids do not include this heading, necessitating special location processing for these statements.

[0042] Next, the directory structure of the entire bid document is identified. Since the bid document does not include a title for the financial statements, no relevant information about the financial statements exists within the directory structure. Analysis of multiple bids reveals that financial statements appear in the qualification review documents, and each bid document has a directory titled "Qualification Review Documents." Therefore, the financial statements can be located within the qualification review documents. During the location process, all images within the qualification review documents must be extracted and stored in page order to prevent misalignment. Each image is then subjected to OCR recognition, and the recognized text is matched for keywords related to financial statements, such as "audit report" and "financial statements," starting with the first keyword that appears. Typically, two financial statements appear in the qualification review documents, resulting in two recognition results. The location interval for the first financial statement is from the first to the second location result. The end position for the second report can be directly taken from the beginning of the next title. Because images within the financial statements undergo separate recognition processing, even if some irrelevant images are included, they will be eliminated during recognition.

[0043] Step 5: Flow item analysis: We analyze failed bid items in the bidding documents, including bidder name, bid signature and seal, unique bid price, business license, production safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, reputation requirements, bid content, construction period, project quality, bid validity period, bid bond, rights and obligations, and priced bill of quantities. For common failed bid items, we construct prompts based on the positioning text in the bid document and the failed bid item standards in the bidding document, and then use the large model to analyze failed bid items. We also conduct specific analysis on specific failed bid items, such as requiring the bidder's seal on every page and the firm's seal on financial statements.

[0044] For financial requirements, we optimized the table recognition technology to identify and extract the financial data of the balance sheet, income statement and cash flow statement in the bidding documents. The accuracy of ordinary table OCR tools is acceptable when recognizing non-rotated table images, but there are certain problems when recognizing skewed or partially distorted tables, and it cannot strictly meet the data correspondence. The commonly used extraction model was modified, and the extraction model structure is as follows: Figure 4 As shown in the figure, the report image is input and the extraction result is output after the three modules of text detection, detection regularization and text recognition are completed. Figure 5 ,Text detection first locates the location of the text and segments it, then divides the text block into square detection areas through regularization and performs data correspondence, and finally performs text recognition on each detection area and outputs the report results.

[0045] Step 6, prompt for bidding document failure: Based on the comparison results of unsuccessful bids, such as bidder name, signature and seal of bid letter, unique quotation, business license, production safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, reputation requirements, bid content, construction period, project quality, bid validity period, bid bond, rights and obligations, and priced bill of quantities, non-conformity notices will be displayed. For example, the financials of bid document A must meet the requirements.

[0046] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for extracting, locating and analyzing bidding document failure information based on an OCR engine, characterized in that: The following steps are involved: Step 1: Procurement document collection: Collect documents and files for bidding projects; Step 2, file parsing: For different types of document files, corresponding information extraction and processing methods are used to obtain key information; Step 3: Check if the bidding documents have failed: Extract standard information on bidding factors from bidding documents, and locate and extract factors that lead to unsuccessful bidding in bidding documents; Step 4: Positioning of failed bidding items in the bidding documents: Analyze the structure and text of the bidding documents, extract the directory structure of the bidding documents, and locate the bidding items in the bidding documents; Step 5: Flow item analysis: Analyze the failed bid items in the bidding documents; Step 6, prompt for bidding document failure: Prompt for non-compliant items based on the comparison results of the unsuccessful bid items.

2. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: The document files collected in step 1 for the bidding project specifically include: Collect document files involved in the bidding project from the procurement system, including: bidding document procurement, bid documents, bidding announcement documents and bidding projects.

3. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: In step 2, different types of document files are divided into two categories: scanned PDF files and non-scanned PDF files. Different information extraction processing methods are performed based on the bids of different types of document files to obtain key information; For scanned PDF files, each page will be identified as an image. The image will be processed first, and different information extraction and processing methods will be called according to the different content of each page. At the same time, the content of the bid includes tables, drawings and technical documents, and different processing will be performed for different contents.

4. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 3, characterized in that: The method for extracting information from tables in scanned PDF files specifically includes: For scanned pages whose image content is a table, a table OCR tool is used for processing; first, the image is corrected, then the image is segmented to obtain the table title and table range, and then the table content in the table range is extracted; during the extraction process: first, the table structure in the image is identified, and the horizontal and vertical lines of the table are obtained based on the recognition results. According to these lines, the original image is divided into multiple independent rectangles, and each rectangle is a cell of the table; then each identified cell is processed, first determining whether the center point of each identified character is in the cell, and if so, the character is added to the character list of the corresponding cell; after the traversal is completed, the data list of each cell is integrated into a table data structure for storage, which is used for subsequent bid processing and analysis.

5. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 4, characterized in that: The method for extracting information from attached drawings in a scanned PDF file specifically includes: The attached drawings are divided into license type and non-license type. The scanned copies of license types are compared with the pre-text to verify the authenticity and relevant qualifications of the bidder. At the same time, the bidder name, license name or permit name on the license are extracted, and the official seal on the license is located and extracted; Among them, name extraction obtains all the text on the license picture through OCR recognition, and then removes common watermark text through regular expressions; the text is matched with keywords of common licenses to obtain the name in the license text; the official seal in the picture is extracted using Textmind, and the extracted license name, bidder name, and official seal are stored together for subsequent bidder qualification verification and technical factor review.

6. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 5, characterized in that: The information extraction processing method for scanned PDF file technical documents specifically includes: First, the long text is segmented, and some stop words in the sentences are removed. The large paragraph of text is segmented into several short sentences. The text segmentation uses full stops, semicolons, and commas as boundaries. For the segmented short sentences, punctuation marks such as quotation marks, book titles, and parentheses are removed, and stop words such as "in order to" and "thus" that have little impact on the original sentence meaning are removed; the processed short sentences of the technical document are respectively matched with the tender factor keywords, and the short sentences with a similarity greater than 85% are taken as auxiliary positioning points; these positioning points are associated and displayed with the tender factors as an auxiliary bid evaluation system to improve the efficiency of expert review.

7. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 6, characterized in that: It also includes the recognition of seals and signatures. The seal detection and recognition technology is used to detect the seals on the page and recognize the seal content to obtain the seal position and text of each page.

8. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: Specifically included in step 3 is: The extraction of tender factor standard information in the tender document is carried out using the similarity matching technology; the tender document non-winning bid items are mapped to common directory headings. After mapping the abstract tender document non-winning bid items and standards to the directly related headings or superior headings, these headings can be directly matched in the directory, thereby improving the accuracy of similarity matching; then the similarity matching technology is used to locate and extract the non-winning bid item factors.

9. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: Specifically included in step 4 is: For the tender documents with a directory structure, they are directly split; For the tender documents without a directory structure, the directory structure is quickly extracted from the full text through specification judgment or font size judgment for the splitting of the tender documents and the positioning of tender factors; By locating the financial statements in the qualification review materials, during the positioning process, all the pictures in the qualification review materials are extracted and saved in the page number order to prevent disorder of the order. Then, OCR recognition is performed on each picture, and keywords related to the financial statements are matched in the recognized text. Starting from the first keywords that appear: "audit report" and "financial statements" as the starting positions, there will be two financial statements in the qualification review materials, and there will be two recognition results. The positioning interval of the first financial statement is from the first positioning result to the second positioning result, and the end position of the second statement directly uses the position where the next heading starts.

10. The method for extracting, locating, and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: Specifically included in step 5 is: Analyze the non-winning bid items in the tender documents from the aspects of bidder name, signature and seal of the bid letter, uniqueness of quotation, business license, work safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, reputation requirements, bid content, construction period, project quality, bid validity period, bid bond, rights and obligations, and priced bill of quantities. For common failed bid items, we use the positioning text content in the bidding documents and the failed bid item standards in the tender documents to construct prompt words, call the large model to analyze the failed bid items, and conduct specific analysis on some failed bid items with clear requirements; For financial requirements, we use optimized table recognition technology to identify and extract the balance sheet, income statement, and cash flow statement of the financial statements in the bidding documents; For skewed or partially distorted tables, the commonly used extraction model is modified, and the report image is input. The extraction result is output after passing through three modules: text detection, detection regularization, and text recognition. The text detection first locates the location of the text and segments it, then divides the text block into square detection areas through regularization and performs data correspondence. Finally, text recognition is performed on each detection area, and the report result is output.

11. The method for extracting, locating and analyzing bidding document failure information based on an OCR engine according to claim 1, characterized in that: The step 6 specifically includes: Non-conformity prompts will be given based on the comparison results of the bidder's name, signature and seal of the bid letter, unique quotation, business license, production safety license, qualification requirements, financial requirements, performance requirements, personnel requirements, credibility requirements, bid content, construction period, project quality, bid validity period, bid deposit, rights and obligations, and unsuccessful items in the priced bill of quantities.

Images