ESG report automatic generation method and system based on dynamic layout planning

Abstract

The invention discloses an ESG report automatic generation method and system based on dynamic layout planning. The automatic generation method comprises the following steps: S1, receiving ESG index content generated by an ESG questionnaire platform; s2, constructing a multi-dimensional tagged visual component library, and dynamically distributing appropriate visual presentation components for the generated ESG index content to obtain a preliminary sequence formed by the content and component pairs; s3, inputting the initial sequence into a layout planner, executing automatic layout planning by the planner, and outputting a structured typesetting instruction; and S4, rendering a final PDF report about the ESG according to the typesetting instruction and the visual theme selected by the user, and outputting the final PDF report. According to the invention, through visual rhythm control and assembly dynamic matching, the professionality and design feeling of a report are improved; anti-segmentation rules and logic group optimization are used for keeping reading coherence; the blank area is automatically identified and filled, the page utilization rate is improved, and the report generation time is shortened.

Description

Technical Field

[0001] This invention relates to a method for automatically generating ESG reports, and more particularly to a method for automatically generating ESG reports based on dynamic layout planning. Background Technology

[0002] In the field of automated ESG (Environmental, Social, and Governance) report generation, as companies increasingly demand disclosure of sustainable development information, efficiently and professionally generating reports that conform to visual standards and dynamically change content has become a technical challenge. Currently, there are two main technical approaches in the market.

[0003] The first type is static template-filled paths. This path pre-sets a fixed report layout, filling the generated content into preset "placeholders." Its core flaw lies in its rigid layout, which cannot adapt to dynamic content. Furthermore, when the content length is variable or the component types are diverse, it is prone to producing unsightly blank spaces or layout errors. In addition, all report layouts are monotonous, lacking visual personality and professional design, and are not suitable for current needs.

[0004] The second approach is the direct generation path using a large model: This approach attempts to have all content, including text and typesetting instructions, generated by a large language model in one go. Its core drawback is that the output results are extremely unstable and uncontrollable. At the same time, the model is prone to generating instructions that do not conform to typesetting standards, leading to serious problems such as separation of content and components, image splitting, and improper handling of cross-pages. The generated quality cannot meet the requirements of professional publishing.

[0005] Therefore, the current market lacks an automated method for generating ESG PDF reports that can achieve intelligent, personalized, and high-quality typesetting output while ensuring generation stability. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art by providing a method and system for automatically generating ESG reports based on dynamic layout planning, which can intelligently, personally and with high-quality typesetting output a PDF report on ESG.

[0007] To achieve the above objectives, the technical solution adopted by this invention is: an automated ESG report generation method based on dynamic layout planning, comprising the following steps:

[0008] S1. Receive ESG indicator content generated via the ESG questionnaire platform;

[0009] S2. Construct a multi-dimensional labeled visual component library and dynamically assign appropriate visual presentation components to the generated ESG indicator content to obtain a preliminary sequence composed of content and component pairs;

[0010] S3. Input the preliminary sequence into a layout planner, which will perform automated layout planning and output a structured typesetting instruction.

[0011] S4. Based on the layout instructions and the visual theme selected by the user, render the final PDF report on ESG and output it.

[0012] Furthermore, each component in the visual component library is tagged with a set of multi-dimensional labels, which include component type, layout occupancy, visual weight, and content adaptability.

[0013] Furthermore, in step S2, when matching visual presentation components for ESG indicator content, the execution process is as follows: First, candidate components are selected from the visual component library based on the attributes of the ESG indicator content; then, anti-duplicate rules are applied to avoid the same type as the previously assigned component, thereby determining the final matched component, and finally obtaining a preliminary sequence composed of content and component pairs.

[0014] Furthermore, step S3 includes the following sub-steps:

[0015] S31. For each item in the preliminary sequence, predict its height in the final document based on the length of its text content and predefined PDF font measurement information;

[0016] S32. On a virtual canvas conforming to PDF page specifications, place each content and component pair in sequence according to the predicted height;

[0017] S33. During the simulated placement process, the visual weight sequence of the arranged sequence is calculated in real time; when the visual weight sequence is detected to be inconsistent with the preset visual rhythm rule, the component fine-tuning mechanism is triggered to automatically replace specific components in the sequence in order to optimize the visual rhythm of the layout.

[0018] S34. Generate the structured layout instructions based on the virtual canvas layout optimized for visual rhythm.

[0019] Furthermore, step S3 also includes the following parallel sub-steps:

[0020] S35. During the simulation, strictly implement the anti-segmentation rules to ensure that any component is completely located on the same page. At the same time, detect blank areas at the end of the page caused by insufficient content.

[0021] S36. In response to detecting the blank area, intelligently select a placeholder component with suitable size and style from an industry-specific placeholder component library and insert it into the blank area.

[0022] S37. Identify a logical group consisting of multiple related components, and prioritize ensuring the proximity of components within the logical group on the page during layout planning. When executing the anti-segmentation rules, prioritize ensuring that the logical group is not split by pagination.

[0023] Furthermore, the height prediction step in step S31 adopts the following prediction model: for a text content i, its predicted height H i The calculation formula is: H i =(L i ×F h ×F ls )+M t +M b , where L i : The number of lines in the text; F h Font height (including line height); F ls Line spacing factor; M t M b The top and bottom margins inherent to the component template.

[0024] Furthermore, in step S33, optimizing the visual rhythm of the layout, a visual weight sequence V=[v1,v2,...,v] is defined. n To quantify the rhythm of the page layout; among them, Represents the first in the sequence The visual weights of each component are calculated; the sense of rhythm is evaluated by calculating the standard deviation σ of the local sequence, when σ < threshold θ. low When the rhythm is monotonous, the component fine-tuning mechanism is triggered to adjust σ to the ideal range.

[0025] Furthermore, when a blank area with a height of H is detected... blank At that time, the system selects the optimal placeholder from the placeholder library P. Its selection strategy is as follows: , where H p : The default height of placeholder p; D(I, I p ): The current company's industry I and the placeholder p's label industry I p Semantic differences between them; λ: a weighting coefficient that balances high matching and industry relevance.

[0026] Furthermore, step S4 specifically includes:

[0027] Load the component template library corresponding to the theme selected by the user;

[0028] According to the layout instructions, fill the content of each indicator into the corresponding thematic component template;

[0029] Dynamically inject the corresponding theme color based on the ESG section to which the current content belongs;

[0030] Finally, the PDF report is generated and output.

[0031] The automated ESG report generation system based on dynamic layout planning includes: a content input interface for receiving ESG indicator content generated via an ESG questionnaire platform; a dynamic component matching module for dynamically matching visual components to the ESG indicator content based on a multi-dimensional tagged component library, generating a preliminary sequence consisting of content and component pairs; a layout planning module for automatically planning the layout of the preliminary sequence and generating structured typesetting instructions; and a thematic rendering module for rendering and generating the final PDF report based on the typesetting instructions and the user-selected visual theme, followed by output.

[0032] Due to the application of the above technical solution, the present invention has the following advantages compared with the prior art:

[0033] The present invention provides an automated ESG report generation method and system based on dynamic layout planning. Through a multi-dimensional tagged visual component library and a context-aware matching process, the system can dynamically allocate appropriate visual components for different ESG content. Combined with a visual weight sequence evaluation and component fine-tuning mechanism, it automatically optimizes the layout rhythm, giving the generated report a professional design feel. Based on anti-segmentation rules and a logical group identification mechanism, it ensures that related components are not split by pagination, prioritizes maintaining the proximity relationship of components within logical groups, and improves the readability and structure of the report.

[0034] Secondly, through the synergistic effect of a precise height prediction model, visual rhythm control, and white space optimization mechanism, the system can achieve efficient page space utilization and significantly reduce white space waste; it can also achieve deep integration of report style with corporate brand or industry characteristics by dynamically matching theme colors according to the visual theme and ESG section selected by the user; and the automated process driven by structured typesetting instructions can ensure highly stable and reliable output results, high typesetting success rate, and significantly reduce manual intervention costs and error rates. Attached Figure Description

[0035] The technical solution of the present invention will be further described below with reference to the accompanying drawings:

[0036] Figure 1 This is a flowchart of an embodiment of the present invention;

[0037] Figure 2 This is a detailed flowchart of an embodiment of the present invention;

[0038] Figure 3 This is a schematic diagram illustrating the specific usage process of step S3 in one embodiment of the present invention;

[0039] Figure 4 The enterprise information provided by the ESG questionnaire platform to enterprises;

[0040] Figure 5 This refers to the completion of standard diagnostics in the questionnaires provided by the ESG questionnaire platform for enterprises.

[0041] Figure 6 This refers to the completion of standard diagnostics in the questionnaires provided by the ESG questionnaire platform for enterprises.

[0042] Figure 7 This method generates a PDF report on ESG.

[0043] Figure 8 This is the table of contents for the PDF report on ESG generated using this method. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0045] This invention provides an automated ESG report generation method based on dynamic layout planning to solve the problems of low report quality and lack of visual personality and professional design in the prior art.

[0046] For ease of understanding, the specific processes in the embodiments of this application are described below. Please refer to [link / reference]. Figure 1 and Figure 2 This application provides an automated ESG report generation method based on dynamic layout planning, comprising the following steps: S1, receiving ESG indicator content generated via an ESG questionnaire platform; S2, constructing a multi-dimensional tagged visual component library and dynamically allocating appropriate visual presentation components to the generated ESG indicator content to obtain a preliminary sequence composed of content and component pairs; S3, inputting the preliminary sequence into a layout planner, which performs automated layout planning and outputs a structured typesetting instruction; S4, rendering and generating a final ESG PDF report based on the typesetting instruction and the user-selected visual theme, and then outputting the report.

[0047] Specifically, in step S1, ESG indicator content generated by the ESG questionnaire platform is received, which includes text, data and related information related to environment, society and governance.

[0048] In step S2, after constructing a multi-dimensional tagged visual component library, each component in the visual component library is tagged with a set of multi-dimensional tags, which include component type, layout occupancy, visual weight, and content adaptability.

[0049] In addition, when matching visual presentation components for ESG indicator content in step S2, a context-aware matching process is executed: first, candidate components are selected from the visual component library based on the attributes of the ESG indicator content; then, anti-duplicate rules are applied to avoid the same type as the previously assigned component, thereby determining the final matched component, thus obtaining a preliminary sequence composed of content and component pairs.

[0050] See Figure 3 Specifically, step S3 includes the following sub-steps:

[0051] S31. For each item in the preliminary sequence, predict its height in the final document based on the length of its text content and predefined PDF font measurement information.

[0052] In predicting the occupied height, this embodiment proposes an accurate height prediction model based on font metrics and content structure. The specific prediction method is as follows: For a text content i, its predicted height H... i The calculation formula is:

[0053] H i =(L i ×F h ×F ls )+M t +M b , where L i The number of lines in the text content; the number of lines is determined by the text length, the width of the component content area, the font size, and the character spacing, and is precisely calculated by dividing the text length by the maximum number of characters per line; F h Font height (including line height); F ls Line spacing factor. M t M b The top and bottom margins inherent to the component template.

[0054] Under standard A4 page conditions, through testing on text blocks of various lengths, this height prediction model can significantly improve prediction accuracy compared to traditional estimation methods based on a fixed number of characters or lines, effectively reducing the problems of layout overflow or excessive white space caused by inaccurate height estimation.

[0055] S32. On a virtual canvas conforming to PDF page specifications, place each content and component pair in sequence according to the predicted height.

[0056] In this step, the "visual weight sequence" V=[v1,v2,...,v] is redefined to quantify the "visual rhythm" of the layout. n ], where v j The visual weight value of the j-th component in the sequence is represented by "high" = 3, "medium" = 2, and "low" = 1.

[0057] The goal of rhythm control is to avoid monotony or abruptness; the system assesses rhythm by calculating the standard deviation σ of local sequences. For example, for a sliding window [v k ,v (k+1) ,v (k+2) When its standard deviation σ < θ low When the value is 0.5, it is judged as "monotonous rhythm", triggering the component fine-tuning mechanism to replace the middle component with a component with a large weight difference until σ reaches the ideal range.

[0058] In addition, user surveys and evaluations have shown that the layout optimized by this algorithm has significantly improved its visual rhythm and aesthetics, effectively avoiding the monotony and mechanical feel common in automated typesetting.

[0059] S33. During the simulated placement process, the visual weight sequence of the arranged sequence is calculated in real time; when the visual weight sequence is detected to be inconsistent with the preset visual rhythm rules, the component fine-tuning mechanism is triggered to automatically replace specific components in the sequence in order to optimize the visual rhythm of the layout.

[0060] S34. Generate the structured layout instructions based on the virtual canvas layout optimized for visual rhythm.

[0061] Furthermore, step S3 also includes the following parallel sub-steps:

[0062] S35. During the simulation, strictly implement the anti-segmentation rules to ensure that any component is completely located on the same page. At the same time, detect blank areas at the end of the page caused by insufficient content.

[0063] S36. In response to detecting the blank area, intelligently select a placeholder component with suitable size and style from an industry-specific placeholder component library and insert it into the blank area.

[0064] In this step, when a blank area with a height of H is detected... blank When the system selects the optimal placeholder from the placeholder library P, Its selection strategy can be expressed as:

[0065] ; where H p : The default height of the placeholder p;

[0066] D(I, I p ): The current company's industry I and the placeholder p's label industry I p Semantic differences between them; λ: a weighting coefficient that balances high matching and industry relevance.

[0067] S37. Identify a logical group consisting of multiple related components, and prioritize ensuring the proximity of components within the logical group on the page during layout planning. When executing the anti-segmentation rules, prioritize ensuring that the logical group is not split by pagination.

[0068] Through the global planning and optimization algorithm of the layout planner in step 3, the system outputs a report layout with a professional visual rhythm, reasonable space utilization, and strict adherence to complex layout rules, achieving a layout quality at the level of a human designer.

[0069] The layout planner in step 3 transforms the uncertain AI layout problem into a deterministic layout planning process based on rules and algorithms, and uses structured layout instructions to drive rendering, fundamentally eliminating problems such as content and component separation and layout disorder, and the output results are highly stable and reliable.

[0070] Furthermore, the specific process of step S4 is as follows:

[0071] S41. Load the component template library corresponding to the user-selected theme, and fill the content of each indicator into the corresponding thematic component template according to the layout instructions. The semantic color wheel is the core of the visual system of this invention, which defines a color system composed of the brand primary color and multiple semantic chapter theme colors. As a preferred embodiment, for the ESG report, the chapter theme colors include environmental theme colors, social theme colors, and governance theme colors. All component templates are flexible templates whose height is automatically expanded by the content, and their style rules are bound to the variables in the semantic color wheel.

[0072] S42. Based on the ESG section to which the current content belongs, dynamically inject the corresponding theme color; according to the semantic section identifier marked for each content and component in the layout instructions, dynamically query the corresponding theme color value from the loaded semantic color palette. When the section identifier corresponds to 'Environment,' 'Society,' or 'Governance,' inject its corresponding section theme color; for common elements such as the cover and table of contents, inject the brand primary color. This process is achieved by assigning color values ​​to predefined CSS variables in the component template, realizing precise and automatic binding between color and the semantic structure of the report.

[0073] S43. Finally, after rendering and generating the PDF report, output the report. Specifically, the components filled with content and injected with dynamic colors are precisely drawn onto the PDF page according to the coordinates and dimensions specified in the layout instructions. At this point, the system has completed a beautifully designed report that is professional and rigorous in layout, and visually appealing with a strong overall brand identity and clear chapter distinctions.

[0074] In step S4, through the precise height prediction model and flexible templates, the system can accurately adapt to dynamically generated content of varying lengths, while ensuring the aesthetics of the layout through visual rhythm control and white space optimization mechanisms. More importantly, through the dynamic color injection in the thematic rendering module, the report achieves a professional design sense that transcends fixed templates and is deeply integrated with the semantics of the content.

[0075] Secondly, leveraging the semantic color palette system and dynamic color binding mechanism defined in the thematic rendering module, the system can generate reports for a massive number of enterprises with distinct visual styles and professionally differentiated chapters. Enterprises can choose themes based on their own brand or industry characteristics, completely solving the core pain point of automated tools outputting "uniform" layouts.

[0076] The technical effects of this invention are illustrated below through a comparison: When multiple ESG reports of varying lengths and complexities were typed using static template filling, direct generation of large models, and the solution of this invention, the solution of this invention demonstrated significant advantages in key indicators such as typesetting efficiency, layout reliability, space utilization, visual effects, and adaptability. A detailed comparison is shown in the table below:

[0077] Performance and quality indicators Static template filling Large model directly generated Invention Solution Layout efficiency Generally speaking, but stable. Extremely low and unstable High and stable Layout reliability Low resolution, prone to blank spaces or errors. Extremely low accuracy, often resulting in serious errors such as element splitting and content separation. Highly compliant with layout rules (such as preventing splitting). Space utilization Low, with fixed white space Unstable, sometimes at the expense of content integrity. High, through intelligent planning and space optimization Visual effects Monotonous and lacking in design Uncontrollable, inconsistent quality It features a professional design aesthetic and supports visual rhythm optimization and theme customization. Adaptability Poor, only applicable to fixed content The theory is universal but the practice is uncontrollable. It is powerful and can dynamically adapt to content of different lengths and types.

[0078] As shown in the table above, the present invention, while ensuring stable output, is significantly superior to existing technical solutions in terms of typesetting efficiency, layout reliability, space utilization, and visual professionalism.

[0079] Furthermore, in the traditional model, hiring a professional ESG team to write and design a high-quality ESG report takes 50-70 working days and costs approximately 150,000 to 300,000 RMB. With this solution, the writing and typesetting process is fully automated, reducing the time to less than 3 days and the cost per report to below 20,000 RMB. Based on an average company generating one ESG report per year, the writing and design costs are reduced by over 85%, enabling SMEs to easily obtain reports with world-class content and design standards.

[0080] The following example, using XYZ Technology Company's ESG report generation, illustrates an embodiment of the present invention:

[0081] 1. The structured text content generated by the ESG questionnaire platform for "XYZ Technology Company", such as "carbon emission intensity decreased by 15% year-on-year", is entered into this system.

[0082] 2. Based on the "data-intensive" nature of the content, the dynamic component matching module matches a component (text + color block + data) with a visual weight of "medium" and a type of "data highlighting" from its multi-dimensional tagged component library. The system applies anti-duplicate rules to ensure that this component is different in type from the previous component.

[0083] 3. After the user clicks "Generate Report", the layout planner starts:

[0084] The layout planner predicts the height occupied by each piece of content based on the font information of the "Tech Blue" theme of "XYZ Technology Company".

[0085] The layout planner places components sequentially on a virtual A4 canvas. When it detects that three consecutive components have a visual weight of "medium", it triggers a component fine-tuning mechanism, which actively replaces the middle component with a "plain text" component with a visual weight of "low" to break the monotony.

[0086] The layout planner identified a "metric-case" logical group and calculated that the remaining space on the current page was insufficient to accommodate the case image component. Based on the anti-segmentation rules, it moved the entire logical group (metric text and case) to the top of the next page, ensuring the integrity of the case.

[0087] When a small blank space appears at the end of a page, the layout planner invokes the blank space optimization mechanism, selects a "Data Security Architecture Diagram" placeholder component related to the "technology" industry from the library, and fills the area.

[0088] 4. Thematic rendering and output

[0089] The layout planner outputs the final layout result as a structured JSON instruction.

[0090] The PDF rendering engine receives the instruction and loads the "Tech Blue" theme's flexible component template library.

[0091] The engine populates the template with content and dynamically injects the "environment theme color" from the theme into the color block area of ​​the component according to the current chapter (such as the "Environment" chapter) marked in the JSON instructions.

[0092] Ultimately, the engine generates a beautifully formatted PDF document titled "XYZ Technology Company ESG Report," which conforms to the "tech blue" visual style and whose content perfectly matches "XYZ Technology Company."

[0093] In performing the above processing, the system processed a report containing 85 content components, and the total time from content input to PDF output was only 10 minutes. Throughout the process, the layout planner successfully avoided 6 potential component splits, fine-tuned 3 sequences with monotonous visual rhythms, and intelligently filled 5 industry-related placeholders. It achieved automated layout and avoided a lot of manual layout rework.

[0094] Through precise height prediction and global planning, the system generated a 72-page PDF report in this embodiment. Thanks to the layout optimization mechanism, the report features a compact layout, with minimal wasted page space and significantly improved space utilization efficiency.

[0095] Furthermore, this invention also discloses an automated ESG report generation system based on dynamic layout planning, comprising: a content input interface for receiving ESG indicator content generated via an ESG questionnaire platform; a dynamic component matching module for dynamically matching visual components to the ESG indicator content based on a multi-dimensional tagged component library, generating a preliminary sequence composed of content and component pairs; a layout planning module for automatically planning the layout of the preliminary sequence and generating structured typesetting instructions; and a thematic rendering module for rendering and generating the final PDF report according to the typesetting instructions and the visual theme selected by the user in this embodiment.

[0096] By combining the modules described above and using the generation methods described above, the system can intelligently and efficiently automate the production of a high-quality PDF report on ESG, meeting the actual needs of enterprises.

[0097] Please see Figures 4 to 6 This is the questionnaire provided by the ESG questionnaire platform for enterprises to complete in actual use. Finally, it automatically generates a subsequent PDF report on ESG with one click. The generated report style is as follows: Figure 7 and Figure 8 As shown.

[0098] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for automatically generating ESG reports based on dynamic layout planning, characterized in that, Includes the following steps: S1. Receive ESG indicator content generated via the ESG questionnaire platform; S2. Construct a multi-dimensional labeled visual component library and dynamically assign appropriate visual presentation components to the generated ESG indicator content to obtain a preliminary sequence composed of content and component pairs; S3. Input the preliminary sequence into a layout planner, which will perform automated layout planning and output a structured typesetting instruction. S4. Based on the layout instructions and the visual theme selected by the user, render and generate the final PDF report on ESG and then output it.

2. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 1, characterized in that: Each component in the visual component library is tagged with a set of multi-dimensional labels, which include component type, layout occupancy, visual weight, and content adaptability.

3. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 1, characterized in that, In step S2, when matching suitable visual presentation components for ESG indicator content, the execution process is as follows: First, candidate components are selected from the visual component library based on the attributes of the ESG indicator content; then, anti-duplicate rules are applied to avoid the same type as the previously assigned component, thereby determining the final matched component, and finally obtaining a preliminary sequence composed of content and component pairs.

4. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 1, characterized in that, Step S3 includes the following sub-steps: S31. For each item in the preliminary sequence, predict its height in the final document based on the length of its text content and predefined PDF font measurement information; S32. On a virtual canvas conforming to PDF page specifications, place each content and component pair in sequence according to the predicted height; S33. During the simulated placement process, the visual weight sequence of the arranged sequence is calculated in real time; when the visual weight sequence is detected to be inconsistent with the preset visual rhythm rule, the component fine-tuning mechanism is triggered to automatically replace specific components in the sequence in order to optimize the visual rhythm of the layout. S34. Generate the structured layout instructions based on the virtual canvas layout optimized for visual rhythm.

5. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 4, characterized in that, Step S3 also includes the following parallel sub-steps: S35. During the simulation, strictly implement the anti-segmentation rules to ensure that any component is completely located on the same page. At the same time, detect blank areas at the end of the page caused by insufficient content. S36. In response to detecting the blank area, intelligently select a placeholder component with suitable size and style from an industry-specific placeholder component library and insert it into the blank area. S37. Identify a logical group consisting of multiple related components, and prioritize ensuring the proximity of components within the logical group on the page during layout planning. When executing the anti-segmentation rules, prioritize ensuring that the logical group is not split by pagination.

6. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 4, characterized in that, The height prediction step in step S31 uses the following prediction model: For a text content i, its predicted height H i The calculation formula is: H i =(L i ×F h ×F ls )+M t +M b , where L i : The number of lines in the text; F h Font height (including line height); F ls Line spacing factor; M t M b The top and bottom margins inherent to the component template.

7. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 4, characterized in that, In step S33, optimizing the visual rhythm of the layout, the visual weight sequence V=[v1,v2,...,v] is defined. n To quantify the rhythm of the page layout; among them, Represents the first in the sequence The visual weights of each component are calculated; the sense of rhythm is evaluated by calculating the standard deviation σ of the local sequence, when σ < threshold θ. low When the rhythm is monotonous, the component fine-tuning mechanism is triggered to adjust σ to the ideal range.

8. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 5, characterized in that, When a blank area with a height of H is detected blank At that time, the system selects the optimal placeholder from the placeholder library P. Its selection strategy is as follows: , where H p : The default height of placeholder p; D(I, I p ): The current company's industry I and the placeholder p's label industry I p Semantic differences between them; λ: a weighting coefficient that balances high matching and industry relevance.

9. The method for automatically generating ESG reports based on dynamic layout planning as described in claim 1, characterized in that, The steps in S4 specifically include: Load the component template library corresponding to the theme selected by the user; According to the layout instructions, fill the content of each indicator into the corresponding thematic component template; Dynamically inject the corresponding theme color based on the ESG section to which the current content belongs; Finally, the PDF report is generated and output.

10. An automated ESG report generation system based on dynamic layout planning as described in any one of claims 1-9, characterized in that, include: The content input interface is used to receive ESG indicator content generated by the ESG questionnaire platform; The dynamic component matching module is used to dynamically match visual components for the ESG indicator content based on a multi-dimensional tagged component library, and generate a preliminary sequence consisting of content and component pairs. The layout planning module is used to automatically plan the layout of the preliminary sequence and generate structured typesetting instructions; The theme rendering module is used to render and generate the final PDF report based on the layout instructions and the visual theme selected by the user, and then output it.

Images