Poster making method, apparatus, and storage medium

Abstract

The application discloses a poster making method and device and a storage medium, relates to the technical field of data processing, and comprises the following steps: analyzing a poster generation instruction of a user to obtain poster generation requirements, calling an external data interface to obtain relevant information data according to the poster generation requirements, and making the generation of poster text content based on real and traceable external data sources to ensure the authenticity and timeliness of the poster content; determining a poster title and a poster theme color according to the relevant information data to realize the automatic matching of the title and the theme color; and rendering a first poster according to the poster title, the poster theme color, the relevant information data, fund information and the preset poster template to form an automatic generation process from user demand input to poster image output. The application can guarantee the timeliness of hot topics on the poster materials and improve the traffic conversion effect.

Description

Technical Field

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

[0002] Currently, the production of posters featuring trending news in the financial industry mainly relies on manual operation. The poster production process typically includes: operations staff determining the topic, designers creating visual designs, compliance personnel reviewing the content, and multiple rounds of cross-departmental communication and revisions.

[0003] In the traditional poster production model mentioned above, the generation of poster text content relies on manual writing and manual review. However, the window period from the occurrence of hot news to its dissemination is extremely short, usually only a few hours to a day. It usually takes several days from topic selection to poster launch, which means that the hot news has already passed by the time the poster material is launched, affecting the traffic conversion effect. Summary of the Invention

[0004] The main purpose of this application is to provide a poster production method, device and storage medium, which aims to solve the technical problem that the hot topic has expired when the poster material is launched by the time the traditional poster production mode is launched, which affects the traffic conversion effect.

[0005] To achieve the above objectives, this application proposes a poster production method, which includes: In response to the poster generation command, the poster generation command is parsed to obtain the poster generation requirements; According to the poster generation requirements, an external data interface is called to obtain relevant information data through the external data interface; Based on the poster generation requirements and the relevant information data, determine the poster title and poster theme color; A first poster is generated based on the poster title, poster theme color, relevant information data, and preset poster template.

[0006] In one embodiment, after the step of generating the first poster based on the poster title, poster theme color, relevant information data, and preset poster template, the method further includes: In response to the editing instruction for the first poster, the editing instruction is parsed to determine the content to be edited and the area to be edited; The content to be edited is applied to the area to be edited to replace the original content in the area to be edited; Based on the application result of the content to be edited, update the layout parameters of at least one first element in the first poster that has a positional association with the area to be edited, and generate a second poster.

[0007] In one embodiment, the content to be edited includes text to be edited, the type of the text to be edited includes global text, the area to be edited includes a target text area, and the step of applying the content to be edited to the area to replace the original content in the area to be edited includes: When the type of the text to be edited is global text, the text to be edited is applied as a whole according to the first style attribute corresponding to the target text area in the preset text style table, so as to replace the original text in the target text area with the text to be edited as a whole. The text style table is used to record the styles of different text areas.

[0008] In one embodiment, the type of the text to be edited includes rich text, and the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited further includes: When the type of the text to be edited is rich text, determine the target substring in the text to be edited that needs to have the style applied and the second style attribute corresponding to the target substring; Locate at least one matching substring in the text to be edited that semantically matches the target substring, and embed the second style attribute as a style tag into at least one of the matching substrings; Based on the basic style attributes corresponding to the target text region in the preset text style table, the text to be edited is applied as a whole, and the applied text is stored as a sequence of fragments with style tags; Replace the original text in the target text region with the fragment sequence; After the step of replacing the original content in the target text region with the fragment sequence, the method further includes: Based on the style information of the replaced text, update the style record in the preset text style table corresponding to the target text area.

[0009] In one embodiment, the text to be edited includes fund information. Before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the method further includes: The fund information is parsed to obtain the fund name, fund code, and fund performance data; Based on the preset display format corresponding to the target text area, the fund name, the fund code, and the fund performance data are formatted and combined to generate structured text, which is then applied to the target text area.

[0010] In one embodiment, the content to be edited further includes an image to be edited, the area to be edited includes a target image area, and the step of applying the content to be edited to the area to replace the original content in the area to be edited includes: Obtain the image access link of the image to be edited; Based on the target image region, match the coordinates of the predefined image placeholder in the preset poster template; Obtain the actual size parameters of the image to be edited based on the image access link, and perform size adaptation processing on the image to be edited based on the actual size parameters; The adapted image is inserted into the position corresponding to the coordinates of the image placeholder to replace the original image in the target image area.

[0011] In one embodiment, the step of inserting the adapted image into the position corresponding to the coordinates of the image placeholder to replace the original image in the target image area includes: Obtain the first size parameter of the original image and the second size parameter of the image to be edited; Determine whether the first dimension parameter is consistent with the second dimension parameter; When the first size parameter is inconsistent with the second size parameter, the position offset of at least one second element in the preset poster template that is associated with the target image area is calculated based on the second size parameter. The coordinates of at least one of the second elements are updated according to the position offset, and the height of the overall poster canvas is updated according to the second size parameter, thereby completing the replacement of the original image in the target image area.

[0012] In one embodiment, before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the method further includes: The content to be edited is subjected to compliance checks based on a preset compliance thesaurus. When the content to be edited matches any word in the compliance thesaurus, it is determined that the content to be edited has failed the compliance check, the application operation of the content to be edited is terminated, and a prompt message indicating that the review has not been approved is returned; When the content to be edited does not match any word in the compliant thesaurus, it is determined that the content to be edited passes the compliance check, and the step of applying the content to be edited to the editing area is allowed.

[0013] In addition, to achieve the above objectives, this application also proposes a poster making device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the poster making method as described above.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the poster production method described above.

[0015] One or more technical solutions proposed in this application have at least the following technical effects: This application responds to user poster generation commands by parsing the commands to obtain poster generation requirements. Based on these requirements, it calls external data interfaces to retrieve relevant information data, ensuring that the poster text content is generated based on authentic and traceable external data sources, thus guaranteeing the authenticity and timeliness of the poster content. By determining the poster title and theme color based on the relevant information data, it achieves automatic matching of the title and theme color. Based on the poster title, theme color, relevant information data, fund information, and a preset poster template, a first poster can be rendered, forming an automated generation process from user input to poster image output. This ensures the timeliness of trending topics on poster materials and improves traffic conversion rates. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the poster production method of this application, as shown in Embodiment 1. Figure 2 The system architecture diagram provided for Embodiment 1 of the poster production method of this application; Figure 3 The logical architecture diagram provided for Embodiment 1 of the poster production method of this application; Figure 4 This is a logical architecture diagram provided for Embodiment 2 of the poster production method of this application; Figure 5 This is a logical architecture diagram provided for Embodiment 3 of the poster production method of this application; Figure 6 A first scene diagram provided for an embodiment of the poster production method of this application; Figure 7 This is a schematic diagram of a second scene provided for an embodiment of the poster production method of this application; Figure 8 A schematic diagram of a third scene provided for an embodiment of the poster production method of this application; Figure 9 A schematic diagram of the fourth scene provided for an embodiment of the poster production method of this application; Figure 10 A schematic diagram of the fifth scene provided for an embodiment of the poster production method of this application; Figure 11 A schematic diagram of the sixth scene provided for an embodiment of the poster production method of this application; Figure 12 A schematic diagram of the seventh scene provided for an embodiment of the poster production method of this application; Figure 13 This is a schematic diagram of the hardware operating environment involved in the poster production method in this application embodiment.

[0019] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0021] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0022] Based on this, the embodiments of this application provide a poster production method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the poster production method of this application.

[0023] In this embodiment, the poster production method includes steps S10 to S40: Step S10: In response to the poster generation instruction, the poster generation instruction is parsed to obtain the poster generation requirements; Step S20: According to the poster generation requirements, call the external data interface to obtain relevant information data through the external data interface; Step S30: Determine the poster title and poster theme color based on the poster generation requirements and the relevant information data; Step S40: Generate the first poster based on the poster title, poster theme color, relevant information data, and preset poster template.

[0024] It should be noted that the poster creation method described above can be applied to an agent-based poster creation system (hereinafter referred to as the "system"), whereby the agent-based poster creation system is subordinate to the poster creation equipment. (See reference...) Figure 2 The poster creation system based on intelligent agents includes an interaction layer, an agent core layer, a capability layer, and a data layer.

[0025] Optionally, the poster generation instruction is a command input by the user through an interactive interface, which is a natural language requirement text. The interaction layer responds to the user-input poster generation instruction by parsing and processing it to extract structured poster generation requirements from the unstructured natural language text.

[0026] Reference Figure 3 Since users' needs are typically expressed in colloquial, unformatted natural language, they cannot be directly recognized and processed by subsequent information retrieval interfaces and content generation modules. Therefore, by performing semantic parsing and intent recognition on the poster generation instructions, the user's vague and generalized natural language needs are transformed into clear and structured data fields. This enables information retrieval and content generation based on accurate structured information, avoiding information mismatch or generated results that do not meet user expectations due to misunderstandings of instructions.

[0027] Optionally, the poster generation requirements include key elements such as trending topics, market themes, and fund product information that the user expects the poster to cover.

[0028] Optionally, the system provides a preset format for inputting requirements at the interaction layer. This template includes at least three fields: trending topics, market information, and product information. The interaction layer guides users to input their requirements according to this template. The trending topics field receives a description of a trending topic from the user; the market information field receives a description of a market analysis topic from the user; and the product information field receives financial information such as the fund code, fund name, and risk level from the user.

[0029] For example, a user could input "Hot Topic Material: Gold Futures Price Fluctuations; Market Material: The Impact of the Global Pandemic on Gold Jewelry and Attention to the Gold Sector and Defensive Assets; Product: XXX". The interaction layer would then use the user's input in the above format as the poster generation instruction and pass it to the Agent core layer.

[0030] After receiving the poster generation instruction, the Agent core layer calls the large language model interface to perform semantic parsing on the instruction. Specifically, the large language model uses preset prompts to perform intent recognition and slot filling on the input text, extracting the following fields from the poster generation instruction: Hot Topic Theme Field, Market Interpretation Theme Field, Fund Code Field, Fund Name Field, and Fund Risk Level Field.

[0031] Optionally, the Agent core layer uses the poster generation requirements as query conditions to call an external data interface to obtain relevant information data. Since the real-time nature and authority of financial information data directly affect the credibility and user reference value of the poster content, relying on the parameters of the large language model itself to generate information content cannot guarantee the real-time nature, accuracy, and traceability of the information, and may lead to the illusion that the generated content does not match the facts. By calling an external data interface to obtain relevant information data, the subsequent generation of poster text is based on real-time, authentic external data sources, ensuring the authenticity and timeliness of the poster content.

[0032] Optionally, after completing the structured parsing of the poster generation requirements, the Agent core layer can identify the type of tool to be invoked for this task based on the parsed topic fields. For example, if it identifies that an information retrieval tool needs to be invoked to obtain real-time information data, the information retrieval tool can be invoked through the capability layer. This information retrieval tool encapsulates the data communication protocol with the external financial information service platform.

[0033] Specifically, the information acquisition tool is invoked. The information acquisition tool takes the hot topic in the poster generation requirements as the first query parameter, the market interpretation topic in the poster generation requirements as the second query parameter, and the fund code and fund name as the third query parameter. It generates a data request according to a preset interface call format and sends the data request to an external financial information service platform via Hypertext Transfer Protocol (HTTP).

[0034] Upon receiving the data request, the external financial information service platform retrieves information text matching the query parameters from its information database and returns the search results via a response message. The information acquisition tool receives the response message, extracts frontline news and market analysis text, and returns it to the Agent core layer.

[0035] The relevant information data obtained is in structured text format and includes at least the main text of the information, the timestamp of the information publication, and the source identifier of the information.

[0036] Optionally, the poster title and poster theme color are determined based on the poster generation requirements and the relevant information data. Since the relevant information data is objective text returned by an external interface, its content is usually quite detailed but lengthy, and cannot be directly used as the poster title. It needs to be summarized and refined to form a concise title that meets the poster display requirements.

[0037] Furthermore, different information content has different emotional tendencies and semantic atmospheres. A unified poster color scheme cannot adapt to diverse information themes. By combining the semantic features of the relevant information data to determine the poster title and theme color, the generated title can accurately convey the core information of the information. At the same time, the overall visual style of the poster matches the semantic atmosphere of the information content, improving the readability of the poster content and the harmony of the visual presentation.

[0038] Optionally, the Agent core layer takes relevant information data as input and calls the large language model interface to perform a title generation task. Specifically, guided by preset title generation prompts, the large language model performs semantic understanding and extracts core information from both trending news texts and market analysis texts, generating corresponding main and subtitle texts according to preset character limits (e.g., main title no more than 10 characters, subtitle no more than 15 characters). The generated title texts must accurately reflect the core content of the information, be concise and attractive.

[0039] Optionally, the Agent core layer performs sentiment analysis and semantic atmosphere recognition on the relevant information data to determine the poster's theme color. For example, the Agent core layer calls a sentiment analysis model interface to determine the sentiment polarity (e.g., positive, neutral, or negative) of hot-button news text and market analysis text, and determines the sentiment tag based on the determination result. Based on the sentiment tag, a corresponding theme color scheme is matched from a preset color scheme library.

[0040] For example, when the emotional polarity is determined to be positive, warm colors (such as red and orange) are matched as the poster theme color; when the emotional polarity is determined to be neutral, neutral colors (such as blue and gray) are matched as the poster theme color; and when the emotional polarity is determined to be negative, cool colors (such as green and dark blue) are matched as the poster theme color.

[0041] Optionally, the Agent core layer generates a first poster based on the poster title, the poster theme color, the relevant information data, and a preset poster template. The preset poster template is pre-stored in the data layer; it includes a poster style definition file with a fixed layout structure and fillable content areas, where the fillable content areas include at least a title fill area, a body text fill area, an image fill area, and a fund information fill area.

[0042] Optionally, by filling the poster title, the poster theme color, and the relevant information data into the corresponding areas of the preset poster template according to a preset mapping relationship, and performing overall rendering of the theme color, the automatic conversion from structured data to a visual poster image is achieved. This process eliminates the need for manual layout and design operations in poster generation, realizing end-to-end automated generation of poster materials, shortening the response time from demand input to finished poster, and solving the problem of cumbersome and time-consuming production processes for trending news posters.

[0043] Optionally, the data layer pre-stores multiple preset poster templates, each with a different layout style and design, and each template is stored in a structured data format (e.g., JSON). The Agent core layer selects at least one candidate template that matches the poster theme color from the multiple preset poster templates in the data layer, and then determines the target template from the candidate templates based on the matching degree of the fund information in the poster generation requirements.

[0044] Optionally, the target template defines multiple populated content areas and their corresponding coordinate positions and style attributes, including at least the title populated area and its coordinates, the body text populated area and its coordinates, and the fund information populated area and its coordinates.

[0045] Optionally, after the Agent core layer determines the target template, it performs text filling operations: filling the main title and subtitle into the title filling area of ​​the target template respectively; filling the hot topic frontline information text into the hot topic frontline body text filling area of ​​the target template; filling the market interpretation information text into the market interpretation body text filling area of ​​the target template; and combining the fund code, fund name, and risk level information into a fund information string and filling it into the fund information filling area of ​​the target template. During filling, each text area is rendered according to the font, font size, and base color defined in the target template.

[0046] Optionally, the Agent core layer applies the poster theme color to the overall visual rendering of the target template, including switching the background color blocks, decorative lines, and title text colors of the template to the color values ​​corresponding to the poster theme color. After completing all the above filling and rendering operations, the Agent core layer calls the image rendering engine to synthesize the filled template data into a bitmap format poster image file, which is then used as the first poster output.

[0047] Optionally, the generated first poster can be stored in an object storage service and an accessible link can be returned to the interaction layer for users to preview and download.

[0048] Optionally, after step S40, the following may also be performed: In response to the editing instruction for the first poster, the editing instruction is parsed to determine the content to be edited and the area to be edited; the content to be edited is applied to the area to be edited to replace the original content in the area to be edited; based on the application result of the content to be edited, the layout parameters of at least one first element in the first poster that has a positional association with the area to be edited are updated to generate the second poster.

[0049] Optionally, the system responds to an editing instruction entered by the user after the first poster is generated, and parses the editing instruction to determine the content to be edited and the area to be edited.

[0050] After previewing the first poster, users may need to modify some of the text content or image materials. In traditional poster production, users need to communicate their modification requests to designers or operations personnel through cross-departmental communication, and then the relevant personnel manually complete the modifications in the design software, involving multiple rounds of information transmission and repeated confirmation. This embodiment receives and parses the natural language editing instructions directly input by the user, enabling the system to directly understand the user's modification intentions and automatically identify the content and area to be edited. It transforms the modification requests that originally required cross-departmental communication into instruction data that can be directly parsed and processed by the system, thereby eliminating the information transmission links in multi-role collaboration and reducing the time and labor costs associated with cross-departmental communication.

[0051] Optionally, after the first poster is generated, the interaction layer provides the user with a preview interface and a continuous dialogue interface. The user inputs modification requests for the first poster through the dialogue interface, such as "change 'rise' to 'climb' in the text" or "replace the image of the hotspot frontline area with this image." The interaction layer then transmits the user's input modification request text as the editing instruction to the Agent core layer.

[0052] Optionally, after receiving the editing instruction, the Agent core layer calls the large language model interface to perform semantic parsing and intent recognition on the editing instruction. Specifically, guided by preset editing parsing prompts, the large language model identifies the editing operation type (e.g., text replacement, style modification, or image replacement), the specific value of the content to be edited (e.g., new text content or image file), and the location information of the area to be edited (e.g., "hot topic frontline text area" or "title area") from the editing instruction.

[0053] After the large language model completes parsing, it outputs the recognition results in a structured data format, which includes at least the content to be edited and the area to be edited. Based on the area to be edited, the Agent core layer retrieves and determines the corresponding target filling area and its coordinate position on the canvas from the template layout data of the first poster, thereby completing the localization of the area to be edited.

[0054] Optionally, the Agent core layer performs different types of content replacement operations based on the data type of the content to be edited.

[0055] Optionally, before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the following may also be done: According to a preset compliance terminology database, the content to be edited is subjected to compliance checks. If the content to be edited matches any word in the compliance terminology database, it is determined that the content to be edited has failed the compliance check, the application operation of the content to be edited is terminated, and a prompt message indicating that the review has not been approved is returned. If the content to be edited does not match any word in the compliance terminology database, it is determined that the content to be edited has passed the compliance check, and the step of applying the content to be edited to the area to be edited is allowed.

[0056] Optionally, before applying the content to be edited to the area to be edited, the system performs a compliance check on the content to be edited based on a preset compliance thesaurus, and decides whether to allow subsequent content replacement operations based on the check results.

[0057] Because marketing posters in the financial industry are strictly regulated information dissemination mediums, their content must comply with relevant regulations. The presence of prohibited words in a poster can lead to serious compliance risks and financial losses. In the traditional poster production model, compliance review is only conducted manually by compliance personnel after the copywriting is completed. This is a post-event sampling inspection model; by the time problems are discovered, the content has already been generated, requiring revisions and a re-approval process, making the process cumbersome.

[0058] This embodiment uses a pre-set compliance thesaurus to check the content to be edited for compliance. When the content contains prohibited words, the system immediately terminates the replacement operation and provides an immediate feedback message, preventing prohibited content from entering the final poster. This eliminates the risk of prohibited content going live due to human oversight and avoids the increased iteration rounds and cross-departmental communication costs associated with rework after a failed review.

[0059] Optionally, the data layer pre-stores the preset compliance terminology. The preset compliance terminology is stored in a structured data format and contains at least multiple entries for prohibited words. Each entry includes at least the word text and a violation type tag. The word text is a specific sensitive or prohibited word string; the violation type tag identifies the violation category of the word, such as prohibited words for advertising, words promising revenue, or words ignoring risks. The preset compliance terminology can be dynamically updated and maintained by the system administrator.

[0060] Optionally, the Agent core layer extracts the text to be detected from the content to be edited. When the content to be edited is text data, it is directly extracted as the text to be detected; when the content to be edited is image data, the Agent core layer calls the Optical Character Recognition (OCR) interface to extract text from the image to obtain the text information contained in the image as the text to be detected.

[0061] Optionally, after obtaining the text to be detected, the Agent core layer calls the text review interface of the capability layer to perform compliance detection. The text review interface reads the preset compliance thesaurus from the data layer and matches the text to be detected with each non-compliant word entry in the preset compliance thesaurus.

[0062] For example, the text review interface uses a multi-pattern matching algorithm (e.g., the Aho-Corasick (string matching) automaton algorithm) to perform keyword matching on the text to be inspected, in order to determine whether there is a substring in the text that completely matches or contains a matching substring of any violating word entry. When the matching algorithm locates any violating word entry in the text to be inspected, it is determined that the content to be edited matches the compliance terminology, that is, the content to be edited has failed the compliance check; when the matching algorithm traverses all violating word entries and finds no match, it is determined that the content to be edited does not match the compliance terminology, that is, the content to be edited has passed the compliance check.

[0063] Optionally, when the content to be edited fails the compliance check, the text review interface returns the detected violation words and their location information within the text to the Agent core layer. The Agent core layer generates a notification message indicating that the review failed, based on the location information. This notification message includes at least the detected violation words and their corresponding violation type tags. The notification message is then returned to the user interface via the interaction layer to inform the user of suggested modifications. Simultaneously, the Agent core layer terminates the current editing process, and the first poster remains unchanged.

[0064] Optionally, when the content to be edited passes the compliance check, the text review interface returns a notification message indicating that the check has passed to the Agent core layer. After receiving the notification message, the Agent core layer determines that it is permissible to execute the step of applying the content to be edited to the area to be edited.

[0065] This embodiment extracts structured poster generation requirements by semantically parsing in response to user poster generation commands. Based on these requirements, it calls external data interfaces to obtain real-time, authoritative relevant information data, ensuring that the poster text content is generated on a foundation of authentic and traceable external data sources, guaranteeing the authenticity and timeliness of the poster content. By determining the poster title and theme color based on the semantic features of the relevant information data, it achieves automatic title summarization and automatic matching of visual style with the semantic atmosphere of the information. Finally, by filling the corresponding area of ​​the preset poster template with the poster title, theme color, relevant information data, and fund information, and rendering it to generate the first poster, an automated generation process from user input to poster image output is constructed.

[0066] Based on the first embodiment of this application, in the second embodiment of this application, the same or similar content as the above embodiment can be referred to the above description, and will not be repeated hereafter. In addition, the content to be edited includes text to be edited, the type of the text to be edited includes global text, and the area to be edited includes a target text area. Global text refers to the editing scope of the text to be edited being the entire content of the entire target text area.

[0067] In one feasible implementation, refer to Figure 4 The method of applying the content to be edited to the area to be edited, thereby replacing the original content in the area to be edited, may be: When the type of the text to be edited is global text, the text to be edited is applied as a whole according to the first style attribute corresponding to the target text area in the preset text style table, so as to replace the original text in the target text area with the text to be edited as a whole. The text style table is used to record the styles of different text areas.

[0068] Optionally, a preset text style sheet is pre-stored in the data layer; the preset text style sheet is used to record the style attributes corresponding to each text area in the poster, and the style attributes include at least the font name, font size, font color, bold status, italic status, and alignment.

[0069] Optionally, the preset text style sheet is stored in the data layer in a key-value pair data format, where the key is a unique identifier for the text area (e.g., "title", "hotspot_body", "market_body", etc.), and the value is the style attribute object corresponding to the text area. The style attribute object contains at least the following fields: font name field (font_family), font size field (font_size), font color field (color), whether it is bold field (is_bold), whether it is italic field (is_italic), and horizontal alignment field (alignment).

[0070] Optionally, after determining that the content to be edited is text to be edited and passes the compliance check, the Agent core layer uses a large language model to identify the editing type of the editing instruction and determine the editing type of the text to be edited; when the type of the text to be edited is global text, the Agent core layer performs the following operations: Based on the region identifier of the target text region, a style query request is sent to the data layer, the query request carrying the region identifier. The data layer performs a matching search in the preset text style table based on the region identifier, locates the style record entry corresponding to the region identifier, and reads the values ​​of all style attribute fields stored in the entry, returning them as the first style attribute to the Agent core layer.

[0071] For example, when the target text area is the "Hotspot Frontline Text Area", the first style attribute returned by the data layer may include: font name "Source Han Sans", font size "14pt", font color "#333333", no bold, no italics, left aligned.

[0072] After receiving the first style attribute, the Agent core layer combines the overall text content of the text to be edited with the style parameters in the first style attribute to generate a structured text node. Specifically, the Agent core layer writes the content of the text to be edited into the content field of the text node, and writes the font name, font size, color, bold state, italic state, and alignment method from the first style attribute into the corresponding style field of the text node.

[0073] The Agent core layer writes the structured text node into the template layout data of the first poster at the text node position corresponding to the target text area, thereby overwriting all the original content and style data of that node. This achieves the replacement of the original text in the target text area with the text to be edited as a whole, and uniformly applies the first style attribute.

[0074] Since the preset text style table records the style attributes of each text area in the poster, and the global text editing command initiated by the user through dialogue usually only contains the text content to be edited without explicitly specifying style parameters, directly rendering with the system default style may result in the style of the text area being inconsistent with the overall design style of the poster. By reading the first style attribute corresponding to the target text area from the preset text style table and applying the first style attribute to the entire text to be edited, the replaced text automatically retains the original style setting of that area, without requiring the user to specify additional style parameters or the designer to manually adjust the style.

[0075] Furthermore, since global text editing refers to replacing all content within the entire target text area, the first style attribute is uniformly applied to all characters in the text to be edited, ensuring the consistency and uniformity of the text style within that area. This eliminates the need for designers to manually locate and modify the text; users can drive the system to automatically complete the entire process from content replacement to style application using natural language commands, thereby further reducing communication and design manpower costs associated with style adjustments during multiple iterations of modifications.

[0076] In one feasible implementation, the type of the text to be edited includes rich text, which means that the text to be edited contains at least one local substring that needs to apply an independent style attribute. That is, the text to be edited does not apply a single uniform style as a whole, but is a composite text composed of multiple text fragments with different styles combined in the original order.

[0077] In one feasible implementation, the method of applying the content to be edited to the area to be edited to replace the original content in the area to be edited can be: When the type of the text to be edited is rich text, determine the target substring in the text to be edited that needs to have a style applied and the second style attribute corresponding to the target substring; locate at least one matching substring in the text to be edited that semantically matches the target substring, and embed the second style attribute as a style tag into at least one matching substring; apply the style to the text to be edited as a whole according to the basic style attribute corresponding to the target text area in the preset text style table, and store the applied text as a fragment sequence with style tags; replace the original text in the target text area with the fragment sequence.

[0078] Optionally, the editing instruction can be type-identified using a large language model. When the large language model determines that the editing instruction contains style modification requirements for local text (e.g., the instruction contains expressions limited to local substrings such as "bold," "change," or "highlight"), the type of the text to be edited is determined to be rich text.

[0079] Optionally, the Agent core layer calls the large language model interface to perform fine-grained parsing of the editing instructions. Specifically, guided by preset rich text editing parsing prompts, the large language model identifies and extracts at least three types of information from the editing instructions: First, the overall content of the text to be edited; second, the text content of the local substrings within the text to be edited that require individual style application, i.e., the target substring; third, the style modification requirements corresponding to the target substring. The style modification requirements include at least color modification requirements or bolding requirements. The large language model combines the extracted text content of the target substring and the corresponding style modification requirements into structured data, which is output as the second style attribute.

[0080] Optionally, the system locates at least one matching substring in the text to be edited that semantically matches the target substring, and embeds the second style attribute as a style tag into at least one of the matching substrings.

[0081] Optionally, after determining the target substring and the second style attribute, the Agent core layer performs a matching substring localization operation on the text to be edited. Specifically, the Agent core layer first obtains the full text content of the text to be edited. When there is only one substring in the text to be edited that completely matches the target substring character-wise, the Agent core layer directly identifies that completely matching substring as the matching substring. When there is no substring in the text to be edited that completely matches the target substring character-wise, the Agent core layer calls a semantic similarity matching algorithm to perform semantic localization.

[0082] Optionally, the Agent core layer converts the target substring and each candidate substring (each phrase divided according to punctuation marks or word boundaries) in the text to be edited into corresponding semantic feature vectors by calling the vector embedding model interface. The Agent core layer calculates the cosine similarity between the semantic feature vector of the target substring and the semantic feature vector of each candidate substring, and determines the candidate substring with the highest cosine similarity as the matching substring that semantically matches the target substring. When there are multiple candidate substrings with similarity exceeding a preset threshold, the Agent core layer determines all candidate substrings that meet the conditions as matching substrings, and can choose to embed part or all of the matching substrings according to the limiting conditions in the instruction (such as "first place", "all", etc.).

[0083] Optionally, after the Agent core layer locates the matching substring, it embeds the second style attribute into the matching substring in a preset style tag format, and the style tag is stored in the fragment data of the text to be edited in a structured markup language format.

[0084] Since the target substring specified by the user through natural language instructions may differ in wording from the original expression in the text to be edited (for example, the user expresses it as "valuation bottom" while the corresponding expression in the original text is "historical low"), if the positioning is based solely on precise character matching, the matching may fail due to the inconsistency in expression, and the style modification cannot be completed.

[0085] This embodiment uses a semantic similarity matching algorithm to locate matching substrings that semantically match the target substring. This allows the system to accurately identify the target text location indicated by the user's intent even when the user's expression differs from the original wording, improving the fault tolerance and robustness of natural language commands. Furthermore, by embedding the second style attribute as a style tag into the matching substring, the substring can exhibit independent style effects during subsequent rendering. This enables fine-grained editing, where only the target text undergoes local style modification while other text retains its original style.

[0086] Users often have detailed editing needs for poster modification, such as highlighting specific keywords or phrases in the text (e.g., bolding or highlighting certain words in red). However, the global text editing method only supports applying the same style to the entire text area, which cannot meet the needs for setting such localized differentiated styles.

[0087] This embodiment enables the system to accurately capture the user's intention to modify the style of local text by identifying and extracting the target substring and its corresponding second style attribute from the editing instructions. This allows for refined editing control of the poster text and avoids the efficiency loss caused by being forced to regenerate or replace the entire text due to local modification needs.

[0088] Optionally, the Agent core layer sends a style query request to the data layer based on the region identifier of the target text region. The data layer locates the corresponding style record entry in the preset text style table, reads the values ​​of all style attribute fields stored in the entry (font name, font size, font color, bold status, italic status, alignment, etc.), and returns them to the Agent core layer as the basic style attributes.

[0089] The Agent core layer uses the style parameters in the basic style attributes as default rendering parameters to perform overall style annotation on the entire text content to be edited. It compares the embedded style tags in the text to be edited with the basic style attributes: for ordinary text fragments without independent style tags, the basic style attributes are retained as their style tags; for matching substrings that already carry the second style tag, the second style tag overrides the corresponding style field in the basic style attributes (i.e., the style tag of this substring has higher priority than the basic style attributes), and other style fields not specified in the second style attributes (such as font name, font size, etc.) still use the corresponding values ​​in the basic style attributes.

[0090] Among them, the basic style attributes refer to the set of default style attributes corresponding to the target text area in the preset text style table.

[0091] Optionally, the Agent core layer segments the text to be edited into multiple consecutive text fragments based on the distribution boundaries of style tags, and combines these fragments into a fragment sequence according to their order in the original text. The fragment sequence is stored in a structured data format, containing at least a fragment index field, a fragment text content field, and a fragment style tag field. This ensures that each fragment in the text to be edited carries a clear style attribute, and that the styles of each fragment are independent and do not interfere with each other.

[0092] In one feasible implementation, after the step of replacing the original content in the target text region with the fragment sequence, the following can also be done: Based on the style information of the replaced text, update the style record in the preset text style table corresponding to the target text area.

[0093] Optionally, the Agent core layer performs statistical analysis on the style tags of each segment in the segment sequence to generate a style information summary of the replaced text. The style information summary includes at least the following information: the style attribute with the highest proportion among all segments in the target text region is used as the main style of that region, and a flag indicating whether the target text region currently contains independent style tags.

[0094] Optionally, the Agent core layer sends a style update request to the data layer, the update request carrying a region identifier for the target text region and a style information digest. The data layer locates the corresponding style record entry in the preset text style table based on the region identifier, and updates the basic style attributes recorded in that entry to the main style attribute value in the style information digest.

[0095] Specifically, when the style tags of all fragments in the target text area are consistent, the data layer updates the style attribute of the entry to the unified style attribute and updates the flag containing the independent style tag to no; when the target text area contains multiple different style tags, the data layer updates the basic style attribute of the entry to the style attribute with the highest proportion and updates the flag containing the independent style tag to yes, so as to indicate that the text area is currently in a rich text state.

[0096] This embodiment updates the style information of the replaced text to the preset text style table in a synchronized manner, so that the style record in the preset text style table corresponding to the target text area is always synchronized with the actual content of the target text area, thus ensuring the continuity of subsequent editing operations and data consistency.

[0097] In one feasible implementation, the text to be edited includes fund information. Before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the following steps can be taken: parsing the fund information to obtain the fund name, fund code, and fund performance data; and formatting and combining the fund name, fund code, and fund performance data according to the preset display format corresponding to the target text area to generate structured text, so as to apply the structured text to the target text area.

[0098] Optionally, the Agent core layer extracts the fund information from the editing instructions. The fund information is the raw text string entered by the user in the dialogue. The Agent core layer calls the large language model interface to perform structured parsing of the fund information.

[0099] Optionally, guided by preset fund information parsing prompts, the large language model performs intent recognition and slot filling on the original text string of the fund information, extracting the fund code field, fund name field, and risk level field. When the fund information contains multiple fund data entries, the large language model extracts multiple structured fund data records according to the same rules, with each record containing at least the fund code, fund name, and risk level.

[0100] Optionally, the large language model combines the extracted fund code, fund name, and risk level into structured data and returns it to the Agent core layer as fund performance data. The fund performance data uses the risk level as the primary performance indicator field. The Agent core layer can also use the parsed fund code to call external interfaces to obtain the fund's real-time net asset value, one-month return, and one-year return, as supplementary information to the fund performance data.

[0101] Optionally, the system can format and combine the fund name, fund code, and fund performance data according to the preset display format corresponding to the target text area to generate structured text.

[0102] Optionally, each preset poster template stored in the data layer contains a display format definition corresponding to each fillable content area. The Agent core layer queries the preset poster template for the display format definition corresponding to the area identifier of the target text area to obtain the preset display format.

[0103] Optionally, the preset display format is stored as a string template, where preset placeholders are used to indicate the insertion position of each data field. For example, when the preset display format is "{code} {name} {risk}", it means that the fund code, fund name, and risk level are arranged in sequence, with each field separated by a space; when the preset display format is "{name} (code: {code}, risk level: {risk})", it means that the data is combined according to the format "name (code: xxx, risk level: xxx)".

[0104] Optionally, after obtaining the preset display format, the Agent core layer fills the fund code, fund name, and risk level into the corresponding placeholder positions in the preset display format to generate the structured text. When the fund information contains multiple fund data entries, the Agent core layer generates a corresponding structured text substring for each fund data entry in the manner described above, and then combines the substrings into a complete structured text according to the multiple data delimiters (such as newlines or semicolons) defined in the preset display format. The generated structured text is then used as the content to be edited, enters the compliance detection process and text replacement process, and is finally applied to the target text area.

[0105] Since the fund information provided by users in the editing instructions is usually presented in natural language, the fund code, fund name and risk level are mixed in the same string, the format is inconsistent and lacks structured identification. If the fund information is directly filled into the target text area as a whole, it may result in chaotic layout and non-compliant display order.

[0106] By performing structured parsing on the fund information and extracting various data fields, different types of fund data can be identified and processed independently, thereby ensuring that the fund information meets the information disclosure requirements of the financial industry when it is displayed.

[0107] In this embodiment, the system responds to the user's editing command for the first poster by automatically parsing the content and area to be edited, eliminating the need for cross-departmental communication to convey modification requests and removing information transmission links in multi-role collaboration. Before content application, the system performs pre-compliance checks using a preset compliance thesaurus, automatically intercepting and providing immediate feedback on non-compliant content, upgrading traditional post-event sampling to pre-event automatic interception, effectively avoiding compliance risks caused by oversights in manual review.

[0108] Based on the first and second embodiments of this application, the same or similar content as the above embodiments in the third embodiment of this application can be referred to the above description and will not be repeated hereafter. In addition, the content to be edited also includes an image to be edited, and the area to be edited includes a target image area.

[0109] In one feasible implementation, refer to Figure 5 The method of applying the content to be edited to the area to be edited, thereby replacing the original content in the area to be edited, may be: Obtain the image access link of the image to be edited; match the predefined image placeholder coordinates in the preset poster template according to the target image area; obtain the actual size parameters of the image to be edited according to the image access link, and perform size adaptation processing on the image to be edited according to the actual size parameters; insert the adapted image into the position corresponding to the image placeholder coordinates to replace the original image in the target image area.

[0110] Optionally, the user uploads the image to be edited through the dialog interface of the interaction layer. The interaction layer supports image formats including at least JPG, PNG, and WebP. The interaction layer performs size verification on the uploaded image. When the image file size exceeds a preset threshold (e.g., 10MB), the interaction layer prompts the user to re-upload an image file that meets the requirements.

[0111] Optionally, after successful verification, the interaction layer transmits the binary data of the image to be edited to the Agent core layer. The Agent core layer calls the object storage interface of the capability layer to upload the binary data of the image to be edited to the object storage service. The object storage service is a distributed file storage system that supports highly available and highly persistent data storage.

[0112] Optionally, after successfully receiving and storing the image to be edited, the object storage service generates a globally unique file identifier for the image file, and generates a corresponding image access link based on the file identifier and the bucket address. The image access link is a Uniform Resource Locator (URL) using the HTTP or HTTPS protocol. The object storage service then returns the image access link to the Agent core layer.

[0113] Optionally, after receiving the image access link, the Agent core layer will also temporarily store the metadata of the image to be edited (including the image file name, file size, upload timestamp and the image access link) in the context cache of the current session, so as to quickly retrieve and reuse the image data in multiple rounds of dialogue.

[0114] Optionally, the preset poster templates are stored in the data layer in a structured data format (e.g., JSON). Each preset poster template's definition file contains a list of image placeholders. Each placeholder entry in the image placeholder list contains at least the following fields: a placeholder identifier field (used to uniquely identify the placeholder, such as "hotspot_image" or "market_image"), a placeholder name field (used to semantically describe the area name corresponding to the placeholder, such as "hotspot frontline image area"), a placeholder x-coordinate field, a placeholder y-coordinate field, a placeholder width field, and a placeholder height field.

[0115] Optionally, the Agent core layer sends a placeholder matching query request to the data layer based on the semantic description of the target image region, the query request carrying the semantic description. The data layer iterates through the placeholder name field of each placeholder entry in the image placeholder list of the preset poster template, performs text similarity matching between the semantic description and each placeholder name, determines the placeholder entry with the highest matching degree as the target image placeholder, and reads the horizontal coordinate, vertical coordinate, width, and height values ​​stored in the entry as the coordinates of the image placeholder and returns them to the Agent core layer.

[0116] When the preset poster template contains multiple image placeholders and the editing instructions do not explicitly specify the target image area, the Agent core layer can determine the target image placeholder according to preset default rules (e.g., prioritize matching the first image placeholder), or return all possible placeholder options to the user through the interaction layer for selection and confirmation.

[0117] Optionally, the Agent core layer initiates an image metadata read request to an object storage service or CDN (Content Delivery Network) via the image access link. The object storage service or CDN returns metadata information of the image to be edited based on the request, including at least the image's original pixel width and original pixel height. The Agent core layer extracts the original pixel width and original pixel height from the metadata information as the actual size parameters.

[0118] The Agent core layer obtains the placeholder width and height defined in the image placeholder coordinates. Based on the actual size parameters and the placeholder width and height, the Agent core layer determines an adaptation strategy: when the actual aspect ratio of the image to be edited matches the aspect ratio of the placeholder coordinates, the Agent core layer scales the width of the image to be edited to the placeholder width and the height to the placeholder height, generating an adapted image using a proportional scaling method.

[0119] Optionally, when the actual aspect ratio of the image to be edited is inconsistent with the aspect ratio of the placeholder coordinates, the Agent core layer performs adaptation processing according to the centering cropping strategy: First, scale the image to be edited proportionally according to the ratio of the placeholder width to the original width, so that the width of the scaled image is equal to the placeholder width. Then, crop the portion of the scaled image from the center vertically, matching the height of the placeholder, as the adapted image. Alternatively, first, scale the image to be edited proportionally according to the ratio of the placeholder height to the original height, so that the height of the scaled image is equal to the placeholder height. Then, crop the portion of the scaled image from the center horizontally, matching the width of the placeholder, as the adapted image.

[0120] The Agent core layer generates adapted images based on the above adaptation strategy and temporarily stores them in the current session context.

[0121] Optionally, the system inserts the adapted image into the position corresponding to the image placeholder coordinates to replace the original image in the target image area. By writing the adapted image into the image node of the target image area, the image node is switched from the original image to the image to be edited, and rendered and displayed at the image placeholder coordinate position according to the adapted size, thus realizing automatic replacement of image content.

[0122] Optionally, the Agent core layer writes the binary data of the adapted image or the image access link into the image node position corresponding to the target image area in the template layout data of the first poster. Specifically, the image node includes at least the following fields: an image link field, used to store the access link or storage path of the adapted image; an image width field, used to store the pixel width value of the adapted image; an image height field, used to store the pixel height value of the adapted image; and image x-coordinate field and image y-coordinate field, used to store the x-coordinate and y-coordinate values ​​in the image placeholder coordinates.

[0123] Optionally, the Agent core layer overwrites all the original data of the image node through a data write operation, using the data of the adapted image as the new content data of the image node, while maintaining the coordinate position of the image node consistent with the coordinates of the image placeholder. After completing the above replacement operation, the Agent core layer marks the updated template layout data as pending rendering. When subsequent layout parameters are updated, the rendering engine will render and display the adapted image at the coordinate position of the image placeholder according to the access link and size parameters of the adapted image.

[0124] In this embodiment, the user uploads an image and specifies the target area through a dialog box. The system then automatically completes the entire process of image storage, area positioning, size adaptation, and content replacement. This eliminates the need for the user to have design software operation skills or for designers to be involved, greatly reducing the operational threshold and labor costs of image editing.

[0125] In one feasible implementation, the method of inserting the adapted image into the position corresponding to the coordinates of the image placeholder to replace the original image in the target image area can be: Obtain the first size parameter of the original image and the second size parameter of the image to be edited; determine whether the first size parameter and the second size parameter are consistent; when the first size parameter and the second size parameter are inconsistent, calculate the position offset of at least one second element in the preset poster template that is associated with the target image area according to the second size parameter; update the coordinates of at least one second element according to the position offset, and update the height of the overall poster canvas according to the second size parameter to complete the replacement of the original image in the target image area.

[0126] Optionally, the template layout data of the first poster stores the width and height values ​​of the original image corresponding to the target image area. The Agent core layer obtains the first size parameters of the original image, i.e., the width and height values ​​of the original image, by reading the width and height fields of the image node corresponding to the target image area in the template layout data.

[0127] Meanwhile, the Agent core layer reads the actual pixel width and actual pixel height of the adapted image from the image metadata of the adapted image, and uses them as the second size parameter of the image to be edited.

[0128] Optionally, the Agent core layer compares the width value in the first size parameter with the width value in the second size parameter, and simultaneously compares the height value in the first size parameter with the height value in the second size parameter. When the width value of the first size parameter is equal to the width value of the second size parameter, and the height value of the first size parameter is equal to the height value of the second size parameter, the Agent core layer determines that the first size parameter and the second size parameter are consistent; when the width values ​​are not equal or the height values ​​are not equal, the Agent core layer determines that the first size parameter and the second size parameter are inconsistent.

[0129] Optionally, after determining that the first size parameter and the second size parameter are inconsistent, the Agent core layer subtracts the height value in the second size parameter from the height value in the first size parameter to obtain a height difference. This height difference is the position offset. When the height value of the second size parameter is greater than the height value of the first size parameter, the height difference is positive, indicating that the subsequent second element needs to move downwards; when the height value of the second size parameter is less than the height value of the first size parameter, the height difference is negative, indicating that the subsequent second element needs to move upwards.

[0130] Optionally, the overall poster canvas refers to the total canvas area used to carry all elements in the preset poster template, and the total canvas area has an adjustable total height value.

[0131] Optionally, the Agent core layer first determines all second elements located after the target image area in the layout structure of the preset poster template. Specifically, the Agent core layer traverses all element nodes in the template layout data and filters out all element nodes whose ordinate values ​​are greater than the ordinate value of the target image area, as the second elements. Since the elements in the preset poster template are arranged in a top-down order, the ordinate values ​​of elements located after the target image area are necessarily greater than the ordinate value of the target image area.

[0132] Optionally, after the Agent core layer determines all the second elements, it adds the position offset to the current ordinate value of each second element to generate an updated ordinate value, and then updates the ordinate field of each second element to the updated ordinate value. When the position offset is positive, all second elements move down; when the position offset is negative, all second elements move up, thereby maintaining the original spacing between them and the target image region.

[0133] Simultaneously, the Agent core layer obtains the current total height value of the entire poster canvas. The Agent core layer adds the position offset to the current total height value to obtain the updated total canvas height value, and updates the canvas height field of the entire poster canvas to the updated total canvas height value to ensure that all elements remain completely within the canvas boundary after the position offset.

[0134] Optionally, after the Agent core layer completes the update of all the above layout parameters, it marks the updated template layout data as ready for rendering. Subsequently, the rendering engine re-synthesizes the poster image based on the updated template layout data to generate the second poster. In the second poster, the image to be edited has replaced the original image in the target image area, and the positions of all second elements and the total height of the canvas have been adapted and adjusted according to the size of the image to be edited, maintaining a consistent overall layout.

[0135] In this embodiment, the system decouples image data from various microservices by storing the user-uploaded image to be edited in an object storage service and generating an image access link, ensuring stable access across services and rounds. Based on this, the system automatically matches the coordinates of predefined image placeholders in a preset poster template according to the target image region specified in the user's natural language command. After obtaining the actual size parameters of the image to be edited and performing size adaptation processing, the insertion is completed. This achieves automatic mapping from user's natural language description to specific pixel coordinates, eliminating the need for users to manually specify coordinate parameters or drag and drop positioning using design tools.

[0136] For example, to help understand the implementation process of the poster production method obtained by combining this embodiment with the above embodiments, please refer to... Figures 6-13 , specifically: In a specific application scenario, a user inputs a poster generation request through an interaction layer. This request includes trending materials, market materials, and product information. The system responds to the user's poster generation command and parses it. After obtaining the poster generation requirements, it calls an external data interface to retrieve relevant information data. Based on this information data, it determines the poster title and theme color, and finally generates a first poster according to a preset poster template. Figure 6 As shown, the generated first poster includes at least the following: a title area titled "Hot News Flash", a main and subtitle area titled "Gold Price Volatility Increases, Gold's Safe-Haven Capacity Remains Undiminished", a text area titled "Market Interpretation", and an area titled "Related Products".

[0137] After previewing the first poster, the user initiates an editing command through the dialog interface: "Change the poster's theme color to gold." For example... Figure 7 As shown, the updated poster's main color has changed to gold, while other elements in the poster remain unchanged.

[0138] The system initiates an editing command via the dialog interface: "Change the font color of the main title to gold." Upon receiving this command, the system first performs semantic analysis to determine that the content to be edited is "modify the color of the main title text," the area to be edited is "the main title text area," and further determines that the type of the text to be edited is rich text (applying the gold style only to all text in the main title, essentially a global text edit), the target substring is "gold price volatility intensifies, gold's safe-haven appeal remains undiminished," and the second style attribute is "color: gold."

[0139] The system applies a gold style to the entire text to be edited based on the basic style attributes corresponding to the target text area in the preset text style table, stores the replaced text, updates the template layout data, and renders the updated poster. For example... Figure 8 As shown, the updated main title text color has changed to gold, while the text style in other areas of the poster remains unchanged.

[0140] The user continued to issue a command through the chat interface: "Help me change the font color of the words 'Oscillating Decline' in the Hot Topics section to red." The updated text "Oscillating Decline" turned gold, while the other elements in the poster remained unchanged. Figure 9 Not shown in the image.

[0141] The user continues to issue a command through the dialog interface: "Please insert two more sets of fund information for me." The system performs structured parsing of the fund information carried in the user's command, extracting the fund name, fund code, and risk level data, and then formats and combines them according to the preset display format corresponding to the target text area. The generated structured text is then inserted into the specified position in the "Related Products" area. For example... Figure 10 As shown, the inserted fund information is arranged and displayed in the relevant product area in a uniform format, which meets the standardized requirements of the financial industry for the display of fund information.

[0142] A user uploads a new image file through a dialog interface and initiates an editing command: "Please insert an image in the 'Hotspot Frontline' area." The system receives the uploaded image to be edited and stores it in an object storage service to obtain an image access link. Simultaneously, it matches the predefined image placeholder coordinates in a preset poster template based on the "Hotspot Frontline" area specified in the editing command. For example... Figure 11 As shown, after the system obtains the actual size parameters of the image to be edited and performs size adaptation processing, it inserts the adapted image into the image placeholder coordinate position of the hotspot front area.

[0143] During this process, because the size parameters of the image to be edited are inconsistent with the original image, the system automatically calculates the height difference as the position offset, uniformly updates the vertical coordinates of all related elements (including the "Market Interpretation" text area, the "Related Products" area, and other decorative elements) located after the hot topic image area, and simultaneously adjusts the overall height of the poster canvas, ultimately generating a third poster. Figure 12 As shown, in the replaced poster, the new image has been successfully inserted into the hot topic area, and the positions of all elements below it and the total height of the canvas have been automatically adapted and adjusted according to the size of the new image. The overall layout remains consistent and there are no overlapping images or text or layout errors.

[0144] Throughout the process, users do not need to communicate with designers, operations staff, or compliance personnel across departments. They can independently generate, modify, and improve posters simply by communicating with the system's natural language. The average time from inputting requirements to producing the final product for a single trending poster has been reduced from several days in the traditional model to within a few minutes, significantly improving the production efficiency of trending news posters and greatly reducing cross-departmental communication and labor costs.

[0145] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the poster production method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0146] This application provides a poster making device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the poster making method in the above embodiments.

[0147] The following is for reference. Figure 13 The diagram illustrates a structural schematic of a poster-making device suitable for implementing embodiments of this application. The poster-making device in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, tablets, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital televisions and desktop computers. Figure 13 The poster-making equipment shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of this application.

[0148] like Figure 13As shown, the poster-making device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the poster-making device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication device 1009 allows the poster-making equipment to communicate wirelessly or wiredly with other devices to exchange data. Although poster-making equipment with various systems is shown in the figures, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0149] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0150] The poster production equipment provided in this application, employing the poster production method described in the above embodiments, can solve the technical problem that traditional poster production methods result in the poster material being uploaded after the hot topic has expired, affecting traffic conversion. Compared with the prior art, the beneficial effects of the poster production equipment provided in this application are the same as those of the poster production method provided in the above embodiments, and other technical features of this poster production equipment are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0151] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0152] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0153] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the poster making method in the above embodiments.

[0154] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0155] The aforementioned computer-readable storage medium may be included in the poster-making equipment; or it may exist independently and not be assembled into the poster-making equipment.

[0156] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the poster-making device, cause the poster-making device to perform the aforementioned poster-making method.

[0157] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0158] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0159] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0160] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described poster production method. This solves the technical problem that traditional poster production methods result in the poster material being uploaded after its trending period has expired, affecting traffic conversion. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the poster production method provided in the above embodiments, and will not be repeated here.

[0161] The above descriptions are merely some embodiments of this application and do not limit the scope of protection of this application. Any equivalent structural transformations made based on the technical concept of this application and the content of this specification and drawings, or direct / indirect applications in other related technical fields, are included within the scope of protection of this application. All actions involving the acquisition of signals, information, or data in this application are performed in accordance with the relevant data protection laws and policies of the country where the application is located and with authorization from the owner of the corresponding device.

Claims

1. A poster production method, characterized in that, The poster production method includes: In response to the poster generation command, the poster generation command is parsed to obtain the poster generation requirements; According to the poster generation requirements, an external data interface is called to obtain relevant information data through the external data interface; Based on the poster generation requirements and the relevant information data, determine the poster title and poster theme color; A first poster is generated based on the poster title, poster theme color, relevant information data, and preset poster template.

2. The poster production method as described in claim 1, characterized in that, After the step of generating the first poster based on the poster title, poster theme color, relevant information data, and preset poster template, the method further includes: In response to the editing instruction for the first poster, the editing instruction is parsed to determine the content to be edited and the area to be edited; The content to be edited is applied to the area to be edited to replace the original content in the area to be edited; Based on the application result of the content to be edited, update the layout parameters of at least one first element in the first poster that has a positional association with the area to be edited, and generate a second poster.

3. The poster production method as described in claim 2, characterized in that, The content to be edited includes text to be edited, the type of the text to be edited includes global text, the area to be edited includes a target text area, and the step of applying the content to be edited to the area to replace the original content in the area to be edited includes: When the type of the text to be edited is global text, the text to be edited is applied as a whole according to the first style attribute corresponding to the target text area in the preset text style table, so as to replace the original text in the target text area with the text to be edited as a whole. The text style table is used to record the styles of different text areas.

4. The poster production method as described in claim 3, characterized in that, The type of text to be edited includes rich text, and the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited further includes: When the type of the text to be edited is rich text, determine the target substring in the text to be edited that needs to have the style applied and the second style attribute corresponding to the target substring; Locate at least one matching substring in the text to be edited that semantically matches the target substring, and embed the second style attribute as a style tag into at least one of the matching substrings; Based on the basic style attributes corresponding to the target text region in the preset text style table, the text to be edited is applied as a whole, and the applied text is stored as a sequence of fragments with style tags; Replace the original text in the target text region with the fragment sequence; After the step of replacing the original content in the target text region with the fragment sequence, the method further includes: Based on the style information of the replaced text, update the style record in the preset text style table corresponding to the target text area.

5. The poster production method as described in claim 3, characterized in that, The text to be edited includes fund information. Before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the method further includes: The fund information is parsed to obtain the fund name, fund code, and fund performance data; Based on the preset display format corresponding to the target text area, the fund name, the fund code, and the fund performance data are formatted and combined to generate structured text, which is then applied to the target text area.

6. The poster production method as described in claim 2, characterized in that, The content to be edited also includes an image to be edited, and the area to be edited includes a target image area. The step of applying the content to be edited to the area to replace the original content in the area to be edited includes: Obtain the image access link of the image to be edited; Based on the target image region, match the coordinates of the predefined image placeholder in the preset poster template; Obtain the actual size parameters of the image to be edited based on the image access link, and perform size adaptation processing on the image to be edited based on the actual size parameters; The adapted image is inserted into the position corresponding to the coordinates of the image placeholder to replace the original image in the target image area.

7. The poster production method as described in claim 6, characterized in that, The step of inserting the adapted image into the position corresponding to the coordinates of the image placeholder to replace the original image in the target image area includes: Obtain the first size parameter of the original image and the second size parameter of the image to be edited; Determine whether the first dimension parameter is consistent with the second dimension parameter; When the first size parameter is inconsistent with the second size parameter, the position offset of at least one second element in the preset poster template that is associated with the target image area is calculated based on the second size parameter. The coordinates of at least one of the second elements are updated according to the position offset, and the height of the overall poster canvas is updated according to the second size parameter, thereby completing the replacement of the original image in the target image area.

8. The poster production method as described in claim 2, characterized in that, Before the step of applying the content to be edited to the area to be edited to replace the original content in the area to be edited, the method further includes: The content to be edited is subjected to compliance checks based on a preset compliance thesaurus. When the content to be edited matches any word in the compliance thesaurus, it is determined that the content to be edited has failed the compliance check, the application operation of the content to be edited is terminated, and a prompt message indicating that the review has not been approved is returned; When the content to be edited does not match any word in the compliant thesaurus, it is determined that the content to be edited passes the compliance check, and the step of applying the content to be edited to the editing area is allowed.

9. A poster making device, characterized in that, The poster making device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the poster making method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the poster production method as described in any one of claims 1 to 8.

Images