Conversation completion method of social software, terminal equipment and storage medium

Abstract

The invention is suitable for the field of data processing, and discloses a dialogue completion method of social software, terminal equipment and a storage medium. The dialogue completion method of the social software comprises the following steps: executing an interface feature analysis operation on a screenshot to obtain an identification result of target software; if it is judged that the target software is the social software according to the recognition result, dialogue context analysis result information is generated according to the screenshot; and determining a user input fragment according to the current cursor position, and executing an intelligent completion operation according to the user input fragment and the dialogue context information to obtain a response option set. According to the method, the semantic relevance and scene applicability of completion suggestions are remarkably improved, and the problems of response machinery and scene splitting caused by lack of context awareness of a completion function of a traditional input method are solved.

Description

Technical Field

[0001] The present invention belongs to the field of data processing, and in particular relates to a conversation completion method, terminal equipment and storage medium for social software. Background Art

[0002] Dialogue completion refers to the process of completing a conversation by filling in missing sentences or words by understanding the context and linguistic context of the conversation.

[0003] In social media apps like WeChat, traditional completion methods rely solely on the user's input fragments (e.g., the text before the cursor). Examples include Sogou Input Method and Baidu Input Method, and the resulting response options lack contextual adaptability. A new technical approach is needed to address these issues. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a conversation completion method, terminal device and storage medium for social software, which can solve the problem in the related art that the completion only relies on the user's input fragment (such as the text before the cursor), and the generated answer options lack scenario adaptability.

[0005] A first aspect of the present invention provides a method for completing a conversation in a social software, comprising:

[0006] Perform interface feature analysis on the screenshot to obtain the recognition result of the target software;

[0007] If the target software is determined to be social software according to the recognition result, generating conversation context analysis result information according to the screenshot;

[0008] A user input segment is determined according to the current cursor position, and a smart completion operation is performed according to the user input segment and the dialog context information to obtain a set of answer options.

[0009] Optionally, in a first implementation of the first aspect of the present invention, the step of performing an interface feature analysis operation on the screenshot to obtain a target software recognition result includes:

[0010] Performing the feature collection operation on the screenshot according to the visual macro model to obtain interface feature information, wherein the interface feature information includes interface layout, brand logo, and functional components;

[0011] Performing the interface feature analysis operation according to the interface feature information to obtain the software name, version number, interface element features and confidence level;

[0012] The recognition result of the target software is determined according to the software name, the version number, the interface element features and the confidence level.

[0013] Optionally, in a second implementation of the first aspect of the present invention, the step of performing a smart completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes:

[0014] Constructing an emotion mapping matrix based on the conversation context information to obtain an emotion feature vector;

[0015] A set of answer options is generated according to the emotional feature vector and the user input segment.

[0016] Optionally, in a third implementation of the first aspect of the present invention, the step of performing a smart completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes:

[0017] Performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options to be processed;

[0018] According to the portrait label corresponding to the identity of the dialogue object, the to-be-processed response option set is semantically rewritten to obtain the response option set.

[0019] Optionally, in a fourth implementation of the first aspect of the present invention, the step of semantically rewriting the to-be-processed answer option set according to the portrait tag corresponding to the identity of the conversation partner to obtain the answer option set includes:

[0020] Determine sensitive words based on the portrait label corresponding to the identity of the conversation partner;

[0021] According to the sensitive words, semantic rewriting is performed on the to-be-processed answer option set to obtain the answer option set.

[0022] Optionally, in a fifth implementation of the first aspect of the present invention, the step of performing a smart completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes:

[0023] Performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options to be processed;

[0024] Determine the characteristics of the current time period, and determine the current conversation topic tags and implicit scene elements based on the screenshots;

[0025] According to the time period characteristics, the current conversation topic tag and the implicit scene elements, the to-be-processed answer option set is semantically rewritten to obtain the answer option set.

[0026] Optionally, in a sixth implementation of the first aspect of the present invention, the step of performing a smart completion operation based on the user input segment and the conversation context information to obtain a set of answer options further includes:

[0027] Determining a user input segment according to a current cursor position, and performing an intelligent completion operation based on the user input segment and the conversation context information to obtain an intermediate answer option set;

[0028] According to the preset polishing scheme information, a polishing operation is performed on the intermediate answer option set to obtain an answer option set.

[0029] Optionally, in a seventh implementation of the first aspect of the present invention, the step of performing an interface feature analysis operation on the screenshot to obtain an identification result of the target software includes:

[0030] If a screenshot operation triggered by a preset interactive behavior is detected, an interface feature analysis operation is performed on the screenshot captured by the screenshot operation to obtain an identification result of the target software.

[0031] In a second aspect, an embodiment of the present invention provides a terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the conversation completion method of the above-mentioned social software are implemented.

[0032] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned conversation completion method of the social software.

[0033] In a fourth aspect, an embodiment of the present invention provides a computer program product, which, when executed on a terminal device, enables the terminal device to execute the above-mentioned method for completing a conversation in social software.

[0034] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: by introducing screenshot analysis and context-aware technology, the contextual adaptability of social software conversation completion is effectively improved. First, the interface feature analysis accurately identifies the target software type (such as WeChat), which enables the completion operation to be accurately applied to social scenarios; second, the context analysis based on the screenshot captures the global information of the current conversation (such as participants, historical messages, interaction status), breaking through the limitation of the traditional solution that only relies on the text before the cursor; finally, by fusing the user input fragment with the context semantics, a set of response options is generated that is both in line with the immediate input intention and adapted to the conversation scenario. The semantic relevance and scenario applicability of the completion suggestions are significantly improved, solving the problems of mechanical responses and scene fragmentation caused by the lack of context awareness in the traditional input method completion function. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is a schematic diagram of an embodiment of a method for completing a conversation in social software according to an embodiment of the present invention;

[0037] Figure 2 This is a schematic diagram of a specific embodiment of step S101 of the method for completing a conversation in social software according to an embodiment of the present invention;

[0038] Figure 3 This is a schematic diagram of a specific embodiment of step S103 of the method for completing a conversation in social software according to an embodiment of the present invention;

[0039] Figure 4 This is another specific embodiment diagram of step S101 of the method for completing a conversation in social software according to an embodiment of the present invention;

[0040] Figure 5 Schematic diagram of a terminal device in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are protected by the present invention.

[0042] It should be noted that the terms "include", "comprising" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, terminal, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices. In the claims, specification and drawings of the present invention, relational terms such as "first" and "second" are merely used to distinguish one entity / operation / object from another entity / operation / object, and do not necessarily require or imply any such real-time relationship or order between these entities / operations / objects.

[0043] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0044] Dialogue completion refers to the process of completing a conversation by filling in missing sentences or words by understanding the context and linguistic context of the conversation.

[0045] In social media apps like WeChat, traditional completion methods rely solely on the user's input fragments (e.g., the text before the cursor). Examples include Sogou Input Method and Baidu Input Method, and the resulting response options lack contextual adaptability. A new technical approach is needed to address these issues.

[0046] In view of this, the embodiment of the present invention provides a conversation completion method, terminal device and storage medium for social software, which effectively improves the context adaptability of social software conversation completion by introducing screenshot analysis and context awareness technology. First, the interface feature analysis accurately identifies the target software type (such as WeChat), which enables the completion operation to be accurately applied to social scenarios; secondly, the context analysis based on the screenshot captures the global information of the current conversation (such as participants, historical messages, interaction status), breaking through the limitation of the traditional solution that only relies on the text before the cursor; finally, by fusing the user input fragment with the context semantics, a set of response options is generated that is both in line with the immediate input intention and adapted to the conversation scenario. The semantic relevance and scenario applicability of the completion suggestions are significantly improved, solving the problems of mechanical responses and scene fragmentation caused by the lack of context awareness in the traditional input method completion function.

[0047] In order to illustrate the technical solution of the present invention, specific embodiments are provided below.

[0048] Figure 1 The present invention provides a flow chart of a method for completing conversations in social software, which can be applied to a terminal device, such as a mobile phone, tablet computer, laptop computer, ultra-mobile personal computer (UMPC), or netbook.

[0049] Specifically, the above-mentioned method for completing a conversation in social software may include the following steps S101 to S103.

[0050] Step S101: performing an interface feature analysis operation on the screenshot to obtain an identification result of the target software.

[0051] In an embodiment of the present invention, a terminal device triggers a screenshot capture mechanism, acquiring complete image data of the current screen through a screenshot interface. This screenshot operation can be initiated by a preset automatic trigger condition (e.g., detecting an input box focus event) or by a user-manual trigger command. The device loads the captured screen image into a memory buffer and activates an interface feature analysis module to perform real-time image analysis.

[0052] The terminal device executes an interface feature analysis algorithm to perform pattern recognition on the visual elements in the screenshot. This process compares the image to a pre-set software feature database (containing key features of common applications, such as interface layout, brand logos, and functional components), using image matching and machine learning classifiers to determine the type of target software currently running. If the recognition result meets the social software matching threshold (e.g., a confidence level exceeding 0.85), the software name and version information are recorded, and the contextual analysis service is activated.

[0053] Step S102: If the target software is determined to be social software according to the recognition result, conversation context analysis result information is generated according to the screenshot.

[0054] In an embodiment of the present invention, after confirming that the target software is a social application, the terminal device calls a conversation context extraction engine to perform region segmentation and optical character recognition (OCR) on the screenshot.

[0055] First, the chat window area is located. Using a time-series analysis algorithm, the device extracts the message sequence within the visible area, including sender ID, timestamp, and text content. The device then integrates the structured conversation history with the current input box state (e.g., partially entered content) to generate contextual analysis results, building a semantic model encompassing recent conversation content, participant relationships, and unanswered messages.

[0056] Step S103 : determining the user input segment according to the current cursor position, and performing a smart completion operation according to the user input segment and the dialog context information to obtain a set of answer options.

[0057] In an embodiment of the present invention, a terminal device monitors changes in the cursor position of an input box, capturing the user's input text fragment in real time. Combined with the semantic relationship graph from the context analysis results, the current input fragment and the conversation history are input into a pre-trained large completion model. This model analyzes contextual relevance based on an attention mechanism, generates multiple candidate completion solutions, and applies a style transfer algorithm to adjust the wording to suit the conversational context (e.g., formal or informal tone). Ultimately, it outputs a set of response options containing different expression styles.

[0058] Optionally, the terminal device uses a graphical interface rendering engine to display the generated response options in a floating panel around the input area. The user can click or swipe to select the completed content, and the terminal device automatically inserts the selected text into the input box while keeping the cursor position editable, completing the intelligent completion process for the conversation content.

[0059] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: by introducing screenshot analysis and context-aware technology, the contextual adaptability of social software conversation completion is effectively improved. First, the interface feature analysis accurately identifies the target software type (such as WeChat), which enables the completion operation to be accurately applied to social scenarios; second, the context analysis based on the screenshot captures the global information of the current conversation (such as participants, historical messages, interaction status), breaking through the limitation of the traditional solution that only relies on the text before the cursor; finally, by fusing the user input fragment with the context semantics, a set of response options is generated that is both in line with the immediate input intention and adapted to the conversation scenario. The semantic relevance and scenario applicability of the completion suggestions are significantly improved, solving the problems of mechanical responses and scene fragmentation caused by the lack of context awareness in the traditional input method completion function.

[0060] Traditional social software recognition relies heavily on fixed-region text OCR or simple icon matching, which makes it difficult to cope with interface dynamics, version differences, and interference from similar software. Based on this, the present invention proposes an alternative embodiment.

[0061] Reference Figure 2 , Figure 2 This is a schematic diagram of a specific embodiment of step S101 of the method for completing a conversation in social software in an embodiment of the present invention. Step S101 also includes the following specific implementation methods.

[0062] Step S1011 , performing the feature acquisition operation on the screenshot according to the visual macro model to obtain interface feature information, wherein the interface feature information includes interface layout, brand logo, and functional components.

[0063] In an embodiment of the present invention, when a terminal device performs a feature acquisition operation on a screenshot using a visual macro model, it first feeds the captured screen image into a pre-trained visual macro model inference engine. This model performs multi-level feature extraction on the image, using a convolutional neural network to identify interface layout features (including menu bar arrangement and functional block distribution), detect brand identity (such as WeChat's green title bar and logo location and style), and locate core functional components (such as the visual features of the chat input box and message bubbles).

[0064] Step S1012: performing the interface feature analysis operation according to the interface feature information to obtain the software name, version number, interface element features and confidence level.

[0065] In an embodiment of the present invention, the terminal device inputs the extracted feature vectors of interface layout, brand logo, and functional components into the interface feature analysis module, where they are matched against a pre-set software feature database. This database stores interface element templates and version identification rules for different versions of social software. A similarity calculation is performed to generate a list of software names, version numbers, and interface element features, and a confidence score is output (e.g., 0.92 indicates a high match).

[0066] Step S1013: determining the recognition result of the target software according to the software name, the version number, the interface element features and the confidence level.

[0067] In this embodiment of the present invention, the terminal device determines the identity of the target software based on the software name, version number, and a confidence threshold (e.g., confidence ≥ 0.85). If the match result is WeChat and the confidence level meets the threshold, the subsequent process is activated; otherwise, the feature analysis is retriggered or the status returns to "unknown software."

[0068] In an embodiment of the present invention, through deep feature extraction of the visual large model and matching of multi-dimensional interface elements, the accuracy of social software recognition can be significantly improved, and it can adapt to the dynamic changes of the interface of different versions of software (such as UI adjustments caused by WeChat version updates), effectively avoiding the misjudgment caused by traditional OCR technology due to position offset or style changes of interface elements. At the same time, the confidence mechanism can effectively distinguish similar software (such as enterprise WeChat and ordinary WeChat), so that context analysis can be accurately triggered.

[0069] Traditional input method completion technologies (such as Sogou and Baidu Input Method) rely primarily on word frequency statistics and basic semantic associations, lacking in-depth analysis of the emotional state of the conversation. This can easily lead to sentiment misjudgment, context mismatch, and neglect of relationships. Based on this, the present invention proposes an alternative embodiment.

[0070] Reference Figure 3 , Figure 3 This is a schematic diagram of a specific embodiment of step S103 of the method for completing a conversation in social software in an embodiment of the present invention. Step S103 also includes the following specific implementation methods.

[0071] Step S1031: construct an emotion mapping matrix based on the conversation context information to obtain an emotion feature vector.

[0072] In this embodiment of the present invention, when generating a set of response options, the terminal device first performs sentiment feature modeling on the results of the conversation context analysis. Using a natural language processing model, the text content in the conversation history is parsed, and the sentiment tendency indicator (e.g., positive, neutral, negative) of each message is extracted. This is combined with the frequency of emoticons and punctuation (e.g., the number of exclamation marks) to construct a multidimensional sentiment mapping matrix. This matrix quantifies the overall sentiment intensity of the conversation using a weighted allocation algorithm, ultimately outputting a sentiment feature vector (e.g., [positivity: 0.8, urgency: 0.6]).

[0073] Step S1032: Generate a set of answer options based on the emotional feature vector and the user input segment.

[0074] In an embodiment of the present invention, the terminal device jointly encodes the sentiment feature vector and the user's current input fragment. Using a pre-trained language model (such as BERT), the two are embedded in the same semantic space. The attention mechanism calculates the association weight between the sentiment vector and the input text, generating candidate completion sentences that match the current emotional context.

[0075] For example, when a highly positive feature is detected, the model prioritizes generating options containing affirmative particles ("I strongly agree"); if a negative sentiment is identified, mitigating wording ("Maybe you can consider it.") is automatically added.

[0076] Finally, the terminal device sorts the generated candidate sentences by sentiment matching, retains the top N best options (such as Top 5), and displays them as a set of selectable response options through the interface rendering module.

[0077] In this embodiment of the present invention, through dynamic modeling of the emotion mapping matrix and guidance from emotion eigenvectors, the emotional adaptability of conversation completion content can be significantly improved, effectively avoiding the mechanical response issues caused by traditional completion technologies that rely solely on semantic associations (such as generating positive responses while ignoring the other party's negative emotions). This makes the generated completion options more consistent with the emotional communication needs of interpersonal communication, reducing communication risks, especially when dealing with sensitive topics or complex social relationships.

[0078] Traditionally, input method completion technologies typically generate suggestions based on a general corpus, without distinguishing the identity characteristics of the conversation partner, leading to identity mismatches, relationship confusion, and template rigidity. Based on this, the present invention proposes an alternative embodiment.

[0079] Reference Figure 4 , Figure 4 This is another specific embodiment diagram of step S101 of the method for completing a conversation in social software in an embodiment of the present invention. Step S103 also includes the following specific implementation methods.

[0080] Step S1033 : performing an intelligent completion operation based on the user input segment and the dialog context information to obtain a set of answer options to be processed.

[0081] In this embodiment of the present invention, when performing intelligent completion, a terminal device feeds the user input fragment and conversation context information into a pre-trained language generation model to generate an initial set of response options to be processed. This model uses an attention mechanism to analyze the semantic relevance of the input text and outputs a set of candidate sentences (e.g., 5-10 initial completion suggestions) that contain basic semantic matches.

[0082] Step S1034: semantically rewrite the pending response option set according to the portrait label corresponding to the identity of the conversation object to obtain the response option set.

[0083] In this embodiment of the present invention, the terminal device invokes a conversation partner identity profiling module to construct a profile tag by analyzing participant information within the conversation context (e.g., WeChat name, group chat role tags, and historical chat history). The conversation partner is categorized according to pre-set identity classification rules (e.g., "superior," "client," "relative," and "friend") and the corresponding semantic rewriting template library is loaded (e.g., using honorific templates for "superior" and embedding professional terminology for "client").

[0084] Finally, the terminal device feeds each candidate sentence in the set of responses into a semantic rewriting engine, which then adjusts the wording based on the persona tags. For example, the original completion suggestion "The solution can be changed like this" can be rewritten based on the "customer" identity to "We recommend the following optimization solution for your reference." This creates the final set of responses tailored to the target identity and displays it on the user interface.

[0085] In this embodiment of the present invention, through dynamic matching of conversation partner identity profile tags and a semantic rewriting mechanism, the adaptability of suggested completions to social scenarios is significantly improved, addressing the problem of inappropriate expression (such as using casual language to superiors) caused by traditional techniques that ignore identity differences. This solution automatically adapts the generated completion options to the language standards of different social relationship scenarios, such as workplace communication, customer service, and communication with friends and family, reducing the cost of manual adjustment of expression style for users.

[0086] In traditional technologies, the completion function of social software lacks a targeted special processing mechanism, which easily leads to content risks, cognitive dissonance, and emotional misleading. Based on this, the present invention proposes an optional embodiment.

[0087] Step S1034 also includes the following specific implementation methods.

[0088] Step S10341, determining sensitive words based on the portrait label corresponding to the identity of the conversation partner.

[0089] In an embodiment of the present invention, during the semantic rewriting process, the terminal device first loads a child-specific sensitive word list from the safe word library based on the identity tag of the dialogue object (such as "child"). This word list contains child protection sensitive words (such as violence and adult terms), education banned words (such as "cheating" and "plagiarism"), and simplified expression rules suitable for children to understand (such as replacing complex idioms with vernacular). The system also activates a child-friendly mode, limiting the semantic complexity of generated sentences (such as sentence length ≤ 15 words, vocabulary level ≤ third grade of primary school).

[0090] Step S10342: semantically rewrite the to-be-processed answer option set according to the sensitive words to obtain the answer option set.

[0091] In an embodiment of the present invention, the terminal device uses a semantic analysis model to detect potentially risky content in the set of pending response options. For example, if a candidate sentence contains "You're doing this wrong," it is rewritten as "We can try a better way" based on the child's profile label. If an abstract concept (such as "economic benefit") is detected, it is automatically replaced with "Doing this will make more people happy." At the same time, encouraging words (such as "You're great!" and "Keep up the hard work next time") are embedded to adapt to children's communication scenarios.

[0092] The terminal device scores the readability of the rewritten sentences, filters out options that do not meet the cognitive level of children, and displays the optimized set of response options through a cartoon interface layout (such as bubble fonts and emoticon icons).

[0093] In this embodiment of the present invention, a sensitive word filtering and semantic simplification mechanism driven by child-like tags effectively blocks negative or complex information inappropriate for children, while generating complementary suggestions tailored to children's cognitive levels and psychological characteristics. While ensuring safe communication, encouraging language and intuitive expression enhance the friendliness and positive guidance of children's interactions with social media.

[0094] In conventional technologies, completion suggestions are usually generated based solely on the conversation text, which is prone to time blind spots, scene fragmentation, topic deviation, etc. Based on this, the present invention proposes an optional embodiment.

[0095] Step S103 also includes the following specific implementation methods.

[0096] Step S1035 : performing an intelligent completion operation based on the user input segment and the dialog context information to obtain a set of answer options to be processed.

[0097] In an embodiment of the present invention, when generating a set of response options, the terminal device uses the clock module to obtain the current time period characteristics (for example, weekdays from 9:00 AM to 6:00 PM are marked as "working hours," and after 10:00 PM are marked as "late night hours") and analyzes the current conversation topic tags based on the conversation content in the screenshot. The chat window screenshot is then subjected to scene recognition using a large visual model to extract implicit scene elements (such as whether the chat background is festive or whether there are meeting invitation cards). The conversation text is then clustered using keywords based on a natural language processing model (for example, "contract" and "quotation" are marked as the topic of "business negotiation").

[0098] Step S1036, determine the characteristics of the current time period, and determine the current conversation topic tags and implicit scene elements based on the screenshot.

[0099] In an embodiment of the present invention, the terminal device inputs time period features (such as "late night"), conversation topic tags (such as "family and friends gathering"), and scene elements (such as "birthday cake emoji") into the semantic rewriting engine. The engine adjusts the semantic style of candidate responses based on predefined scene adaptation rules. For example, it automatically adds caring statements such as "go to bed early" during the late night period, or embeds formal honorifics when the topic of "business negotiation" is detected. At the same time, it generates time-sensitive suggestions (such as "need to confirm tomorrow's meeting time") in combination with scene elements (such as identifying "meeting reminder").

[0100] Step S1037: semantically rewrite the to-be-processed answer option set according to the time period feature, the current conversation topic tag, and the implicit scene elements to obtain the answer option set.

[0101] In an embodiment of the present invention, the terminal device sorts the rewritten answer option set according to the scene matching degree, and dynamically displays the completion suggestions that are strongly related to the current time, topic and scene through the interface rendering module.

[0102] In this embodiment, a multi-dimensional semantic rewriting mechanism integrating time, topic, and context can significantly improve the timeliness and contextual adaptability of conversation completion. For example, it can automatically downplay business-like expressions in late-night conversations and embed greetings in holiday conversations. This effectively avoids the problem of unnatural responses caused by traditional technologies that ignore time or contextual factors, making completion suggestions more tailored to the implicit needs of actual communication scenarios.

[0103] In traditional technologies, input method completion technology usually directly outputs the original suggestions generated by the model, which has the following problems: rough expression, single style, and lack of professionalism. Based on this, the present invention proposes an optional embodiment.

[0104] Step S103 also includes the following specific implementation methods.

[0105] Step S1038 , determining the user input segment according to the current cursor position, and performing a smart completion operation according to the user input segment and the dialog context information to obtain an intermediate answer option set.

[0106] In an embodiment of the present invention, when generating a set of response options, the terminal device uses an input box cursor position tracking module to capture the user's input text fragment (e.g., "Regarding the solution, I agree") in real time. This fragment, along with conversation context information (e.g., the three most recent conversation records), is fed into an intelligent completion model (e.g., GPT-4). The model generates an intermediate set of response options based on semantic associations (e.g., basic completion suggestions such as "needs further discussion" and "believes that the implementation details can be optimized").

[0107] Step S1039: Performing a polishing operation on the intermediate answer option set according to the preset polishing scheme information to obtain an answer option set.

[0108] In this embodiment of the present invention, the terminal device invokes a polishing solution loading module to retrieve a preset polishing rule library (e.g., "business formal," "colloquial," and "humorous") from a local or cloud-based source. Based on scenario features within the conversation context (e.g., detecting that a chat window contains a contract file transfer record), the terminal automatically matches a polishing strategy (e.g., selecting the "business formal" template). The semantic conversion model then adjusts the style of the intermediate option set. For example, the phrase "implementation details can be optimized" can be polished to "it is recommended to prioritize improving the implementation details of the implementation plan."

[0109] The terminal device performs a consistency check on the polished set of answer options, removes options with semantic breaks or logical conflicts, and generates the final set of answer options for display through a dynamic sorting algorithm (such as based on the user's historical selection preferences).

[0110] In this embodiment, secondary optimization of completion suggestions through a pre-set polishing scheme improves the linguistic standardization and stylistic adaptability of the generated text, resolving the stiff and unprofessional nature of traditional completion suggestions. For example, this can automatically enhance formal terms and structured expressions in business scenarios, or add emotional phrasing to conversations between friends and family, making the completion content more tailored to the user's actual communication needs.

[0111] In traditional technologies, social software recognition relies on timed screenshots (e.g., every 5 seconds) or forces users to manually take screenshots, which can easily lead to resource waste, recognition delays, and a fragmented experience. Based on this, the present invention proposes an optional embodiment.

[0112] Before step S101, the following specific implementation is also included.

[0113] Step S201: If a screenshot operation triggered by a preset interactive behavior is detected, an interface feature analysis operation is performed on the screenshot captured by the screenshot operation to obtain an identification result of the target software.

[0114] In an embodiment of the present invention, the terminal device monitors user operation behaviors (such as long pressing the screen, sliding down with three fingers, and other preset gestures) in real time through the interactive event monitoring module. When a screenshot trigger action that meets the preset conditions is detected, the system screenshot interface is called to capture the current screen image data.

[0115] Optionally, after the screenshot operation is completed, the device compresses the original image data into a low-resolution copy (such as 720p) to reduce subsequent processing load, and temporarily stores it in a memory cache.

[0116] The terminal device activates the interface feature analysis thread to quickly extract features from the captured screenshot. Using a lightweight inference engine based on a large visual model (such as MobileNet), it identifies core interface elements (such as the WeChat green title bar and input box location) and performs real-time matching against a pre-configured social media feature library to generate a recognition result for the target app. If the confidence level exceeds a threshold, the subsequent process is triggered; otherwise, the current screenshot is discarded and the monitoring state is reset.

[0117] In this embodiment of the present invention, an on-demand screenshot and real-time analysis mechanism triggered by pre-set interactive behaviors can significantly reduce resource consumption while improving the response speed and accuracy of social software recognition. While ensuring functional availability, it effectively solves the problems of device heating and reduced battery life associated with traditional continuous screenshot solutions.

[0118] like Figure 5 FIG2 is a schematic diagram of a terminal device according to an embodiment of the present invention. The terminal device 500 may include a processor 501, a memory 502, and a computer program 503 stored in the memory 502 and executable on the processor 501, such as a conversation completion program for social software. When the processor 501 executes the computer program 503, the steps described in the aforementioned embodiments of conversation completion for social software are implemented.

[0119] The computer program can be divided into one or more modules / units, which are stored in the memory 502 and executed by the processor 501 to implement the present invention. One or more modules / units can be a series of computer program instruction segments that can perform specific functions. These instruction segments are used to describe the execution process of the computer program in the terminal device.

[0120] The terminal device may include, but is not limited to, a processor 501 and a memory 502. Those skilled in the art will appreciate that Figure 5It is only an example of a terminal device and does not constitute a limitation of the terminal device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0121] The processor 501 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0122] The memory 502 can be an internal storage unit of the terminal device, such as a hard drive or memory of the terminal device. The memory 502 can also be an external storage device of the terminal device, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped with the terminal device. Furthermore, the memory 502 can include both an internal storage unit of the terminal device and an external storage device. The memory 502 is used to store computer programs and other programs and data required by the terminal device. The memory 502 can also be used to temporarily store data that has been output or is about to be output.

[0123] It should be noted that, for the convenience and brevity of description, the structure of the above-mentioned terminal device can also refer to the specific description of the structure in the method embodiment, which will not be repeated here.

[0124] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method for completing a conversation in social software can be implemented.

[0125] An embodiment of the present invention provides a computer program product. When the computer program product is run on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned method for completing a conversation in social software.

[0126] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0127] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0128] In the embodiments provided herein, it should be understood that the disclosed terminal devices and methods can be implemented in other ways. For example, the terminal device embodiments described above are merely illustrative. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via some interface, device, or unit, which may be electrical, mechanical, or other means.

[0129] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0130] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0131] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0132] The above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may be modified or some of the technical features thereof may be replaced with equivalents. Such modifications or replacements do not deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention and are therefore intended to be included within the scope of protection of the present invention.

Claims

1. A method for completing a conversation in a social software, characterized in that: include: Perform interface feature analysis on the screenshot to obtain the recognition result of the target software; If the target software is determined to be social software according to the recognition result, generating conversation context analysis result information according to the screenshot; A user input segment is determined according to the current cursor position, and a smart completion operation is performed according to the user input segment and the dialog context information to obtain a set of answer options.

2. The method for completing a conversation in social software according to claim 1, wherein: The step of performing an interface feature analysis operation on the screenshot to obtain a target software recognition result includes: Performing the feature collection operation on the screenshot according to the visual macro model to obtain interface feature information, wherein the interface feature information includes interface layout, brand logo, and functional components; Performing the interface feature analysis operation according to the interface feature information to obtain the software name, version number, interface element features and confidence level; The recognition result of the target software is determined according to the software name, the version number, the interface element features and the confidence level.

3. The method for completing a conversation in social software according to claim 1, wherein: The step of performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes: Constructing an emotion mapping matrix based on the conversation context information to obtain an emotion feature vector; A set of answer options is generated according to the emotional feature vector and the user input segment.

4. The method for completing a conversation in social software according to claim 1, wherein: The step of performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes: Performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options to be processed; According to the portrait label corresponding to the identity of the dialogue object, the to-be-processed response option set is semantically rewritten to obtain the response option set.

5. The method for completing a conversation in social software according to claim 4, wherein: The step of semantically rewriting the to-be-processed answer option set according to the portrait label corresponding to the identity of the conversation object to obtain the answer option set includes: Determine sensitive words based on the portrait label corresponding to the identity of the conversation partner; According to the sensitive words, semantic rewriting is performed on the to-be-processed answer option set to obtain the answer option set.

6. The method for completing a conversation in a social software according to claim 1, wherein: The step of performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options includes: Performing an intelligent completion operation based on the user input segment and the conversation context information to obtain a set of answer options to be processed; Determine the characteristics of the current time period, and determine the current conversation topic tags and implicit scene elements based on the screenshots; According to the time period characteristics, the current conversation topic tag and the implicit scene elements, the to-be-processed answer option set is semantically rewritten to obtain the answer option set.

7. The method for completing a conversation in a social software according to claim 1, wherein: The step of performing a smart completion operation based on the user input segment and the conversation context information to obtain a set of answer options further comprises: Determining a user input segment according to a current cursor position, and performing an intelligent completion operation based on the user input segment and the conversation context information to obtain an intermediate answer option set; According to the preset polishing scheme information, a polishing operation is performed on the intermediate answer option set to obtain an answer option set.

8. The method for completing a conversation in social software according to claim 1, wherein: The step of performing an interface feature analysis operation on the screenshot to obtain a recognition result of the target software includes: If a screenshot operation triggered by a preset interactive behavior is detected, an interface feature analysis operation is performed on the screenshot captured by the screenshot operation to obtain an identification result of the target software.

9. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the method for completing a conversation in social software according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the conversation completion method for social software as claimed in any one of claims 1 to 8 are implemented.

Images