Movie and television play information collection method and device, electronic equipment and medium

Abstract

The invention provides a movie and television play information collection method and device, electronic equipment and a medium. The method comprises the steps of obtaining a user file path of a file; classifying the user file paths, and screening out target user file paths of target files belonging to movie and television play categories; inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model; and performing multi-level information matching on the target key information and movie and television play information in a pre-configured movie and television play resource library, and determining a target movie and television play with the highest matching degree with the target key information, thereby accurately determining the target movie and television play corresponding to the target file, and facilitating visual display of more movie and television play information.

Description

Technical Field

[0001] The present application relates to the field of information processing technology, and more specifically, to a method, device, electronic device and medium for collecting film and television drama information. Background Art

[0002] Multimedia players are used to play video resources stored locally or on a network disk by the user. However, when a multimedia player opens the folder where the video resource is located, it usually only displays the name of the multimedia player. Some players will display image frames in the video resource, but cannot clearly display the relevant film and television information of the video resource. Compared with the online video resources displayed by the multimedia player, the information displayed is insufficient and not intuitive enough, and the visual effect is poor, which is not conducive to users quickly finding and playing the required video files. Even if the existing technology retrieves the relevant information of the video based on the file name, there is still a problem of low retrieval accuracy. Summary of the invention

[0003] In view of this, the purpose of this application is to provide a film and television information collection method, device, electronic device and medium, which can accurately extract corresponding film and television information based on files, thereby intuitively displaying more film and television information, which is conducive to users quickly finding and playing the required video files.

[0004] An embodiment of the present application provides a method for collecting information about a film or TV series, the method comprising:

[0005] Get the user file path of the file;

[0006] Classify the user file paths and filter out the target user file paths of the target files belonging to the film and television drama category;

[0007] Inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0008] The target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0009] In some embodiments, a device for collecting information about film and television dramas is also provided, the device comprising:

[0010] The acquisition module is used to obtain the user file path of the file;

[0011] A classification module is used to classify user file paths and filter out target user file paths of target files belonging to the film and television drama category;

[0012] An extraction module, used for inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0013] The matching module is used to perform multiple-level information matching between the target key information and the film and television drama information in the pre-configured film and television resource library, determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0014] In some embodiments, an electronic device is also provided, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the method for collecting film and television drama information are performed.

[0015] In some embodiments, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method for collecting film and television drama information are executed.

[0016] The present application provides a method, device, electronic device and medium for collecting film and television drama information. The method obtains a user file path of a file; classifies the user file path and screens out a target user file path of a target file belonging to a film and television drama category; inputs the target user file path into a trained key information extraction model, and extracts target key information of the target user file path through the trained key information extraction model; performs multiple levels of information matching between the target key information and the film and television drama information in a pre-configured film and television resource library, determines the target film and television drama with the highest matching degree with the target key information, and obtains the target film and television drama information of the target film and television drama, thereby obtaining specific film and television drama information from the user file, and accurately determines the target film and television drama corresponding to the target file by classifying and identifying the user file path, accurately extracting the target key information of the target user file path through an intelligent model, and matching and retrieving multiple levels of key information and a film and television drama resource library, so as to facilitate the intuitive display of more film and television drama information, help users quickly find and play the required video files, and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flowchart of the method for collecting film and television drama information according to an embodiment of the present application is shown;

[0019] Figure 2 A flow chart of a method for classifying user file paths according to an embodiment of the present application is shown;

[0020] Figure 3 A flowchart of a method for training a key information extraction model according to an embodiment of the present application is shown;

[0021] Figure 4 A flow chart of a method for determining a target film or TV series with the highest matching degree with the target key information according to an embodiment of the present application is shown;

[0022] Figure 5 A schematic diagram showing a graphical user interface of a multimedia player according to an embodiment of the present application is shown;

[0023] Figure 6 A schematic diagram showing the structure of a device for collecting information about film and television dramas according to an embodiment of the present application is shown;

[0024] Figure 7 A schematic structural diagram of an electronic device described in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0026] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0027] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0028] Video players are used to play video resources stored locally or on a network disk. However, when a video player opens the folder where a video resource is located, it usually only displays the name of the video resource. Some players display image frames in the video resource, but cannot clearly display the relevant film and television information of the video resource. Compared with online video resources displayed by video players, the information displayed is insufficient and not intuitive enough, and the visual effect is poor, which is not conducive to users quickly finding and playing the required video files.

[0029] Based on this, in an embodiment of the present application, a method, device, electronic device and medium for collecting film and television drama information are provided, wherein the method obtains a user file path of a file; classifies the user file path, and screens out a target user file path of a target file belonging to a film and television drama category; inputs the target user file path into a trained key information extraction model, and extracts target key information of the target user file path through the trained key information extraction model; performs multiple levels of information matching between the target key information and the film and television drama information in a pre-configured film and television resource library, determines the target film and television drama with the highest matching degree with the target key information, and obtains the target film and television drama information of the target film and television drama, thereby obtaining specific film and television drama information from the user file, and accurately determines the target film and television drama corresponding to the target file by classifying and identifying the user file path, accurately extracting the target key information of the target user file path through an intelligent model, and matching and retrieving multiple levels of key information and a film and television drama resource library, so as to facilitate the intuitive display of more film and television drama information, help users quickly find and play the required video files, and improve user experience.

[0030] The method for collecting film and television drama information described in the embodiment of the present application can be applied to various types of multimedia playback applications.

[0031] Optionally, the multimedia playback application using the generation method of the film and television drama information collection method provided by the present invention can be a video playback application, an application with video playback function, etc., but is not limited to this. The present invention does not limit the application type of the multimedia playback application.

[0032] Taking the multimedia playback application as a video playback application as an example, the film and television information collection method provided by the present invention can be used to determine the film and television information corresponding to the user's film and television file, and display the film and television information in the file display area, so that the user can intuitively understand the content of the film and television file, as well as more convenient film and television information in addition to the name, so as to quickly find and play the files required by the user.

[0033] The above-mentioned film and television drama information collection method can be executed by a computing device, and the computing device can be a server, such as one server, multiple servers, a server cluster, a cloud computing platform, etc., but is not limited to this. The embodiment of the present application does not limit the device type of the computing device.

[0034] The method for collecting film and television drama information described in the embodiment of the present application is to analyze the user's files and extract some core information to determine whether the file is a movie or TV series. If it is a movie or TV series, it is also necessary to extract specific information such as the film and television drama name, season, episode and year. After obtaining this information, search and match are performed based on the existing film and television drama resource library, and finally the correct film and television series is located, thereby obtaining various relevant information such as the title, introduction, actors and directors corresponding to the file. The process from a user file to obtaining specific film and television drama information can be called film and television drama scraping.

[0035] Please refer to Figure 1 , Figure 1 The flowchart of the method for collecting film and television drama information according to the embodiment of the present application is shown, and the method comprises the following steps S101-S104:

[0036] S101, obtaining a user file path of a file;

[0037] S102, classifying user file paths, and filtering out target user file paths of target files belonging to the category of film and television dramas;

[0038] S103, inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0039] S104, performing information matching at multiple levels on the target key information and the film and television drama information in the pre-configured film and television resource library, determining the target film and television drama with the highest matching degree with the target key information, and obtaining the target film and television drama information of the target film and television drama.

[0040] In step S101, the user file path of the file is obtained.

[0041] Here, the file is a local file stored in the terminal device, or a file stored in an application such as a network disk.

[0042] The user file path is an absolute path starting from the root directory.

[0043] Because so many levels of directories are not needed in subsequent classification and other processing processes, in an optional embodiment, the user file path of the file is clipped, the user file path of the last preset number of levels is retained, and the user file path is updated.

[0044] The user file path of the last preset number of levels may be level 4, level 5, and so on.

[0045] As an example only, / resource disk / learning materials / Python / Python / 30 iterators and list analysis.mp4. The resource disk is useless in subsequent judgments. Therefore, cut " / resource disk" and the updated user file path is: / learning materials / Python / Python / 30 iterators and list analysis.mp4.

[0046] In step S102, the user file paths are classified to filter out target user file paths of target files belonging to the category of film and television dramas.

[0047] The film and television drama categories include TV series, movies, variety shows, MVs, etc.

[0048] The embodiment of the present application performs file pre-classification on the file based on the user file path, first determining whether the user file is a film or TV drama file or something else. If it is something else, it can be quickly skipped without occupying subsequent resources.

[0049] For details, please refer to Figure 2 , Figure 2 A flow chart of a method for classifying user file paths according to an embodiment of the present application is shown; the method of classifying user file paths and screening out target user file paths of target files belonging to the category of film and television dramas includes the following steps S201-S202:

[0050] S201, pre-judge the user file path according to the rules, and filter out the target user file path of the target file belonging to the film and television drama category;

[0051] S202: input the remaining user file paths after the target user file paths are filtered out into a pre-trained classification model, and filter out the target user file paths of the target files belonging to the film and television drama category from the remaining user file paths based on the classification result of the classification model.

[0052] A rule pre-judgment is performed on the user file path. Specifically, a regular judgment is performed on the user file path to filter out target user file paths of target files belonging to the film and television drama category.

[0053] Just as examples, here are some file paths for other videos:

[0054] / 2024 Exam Data Compilation / System Class / Data Analysis / Exercise / 03.20230703Exercise 3.mp4;

[0055] / xxxx / My Sharing / Daily Update / Total Directory / Video Name Example / .mp4.

[0056] Here are some file paths for movies and TV series:

[0057] / Backup disk / Movie / Animation / China / Animation name / Animation Season 2 name / Episode 58 Blu-ray (1080P)-45-14.mp4;

[0058] / Selected TLF-MiniSD500 (xxx share) / 1981-1990.100 old movie collection.mkv;

[0059] / Resource disk / movie / TV series name / 46_4K.mp4;

[0060] / 115V / TV Series / Watched / TV Series Name 4K.H265.Complete.Internal Subtitles / TV Series Name.Episode 34.4K.H265.2023-01-30-Punch;Actor A, Actor B-More【111111.com】.mkv;

[0061] It can be seen that among various file paths, for many film and television drama files, there are already some obvious rules in the naming of their file paths, such as season and episode identifiers such as S*E*. According to these rules, some simple regular judgments are designed in advance to quickly filter out some obvious files.

[0062] As an example only, including the season identifier "Season * and / or S*E*" will determine it as a TV series category.

[0063] The remaining user file paths after filtering out the target user file paths are input into a pre-trained classification model, and the target user file paths of the target files belonging to the film and television drama category in the remaining user file paths are filtered out based on the classification results of the classification model; that is, for most file paths, they still cannot be identified through regular judgment because they are judged by the AI ​​model.

[0064] In some optional embodiments, a classification network is designed based on the Bert model, a classic architecture commonly used in the NLP field. The classification network receives a user file path as input, first converts it into a text vector, and then passes it through a trained binary classification head to output the final classification result, i.e., film and television drama or others.

[0065] For files that are judged as other by the AI ​​classification model, no further processing will be performed and the collection results will be directly output as empty. However, files that are judged as film and television drama types by rules or classification models will continue to the next key information extraction module to extract core information.

[0066] It should be noted that regular judgment may make mistakes. Even if some user directories carry season labels such as S*E*, they are not movies or TV series, but diaries named by the users themselves. The output results of the AI ​​classification model are probabilistic in themselves.

[0067] Here, the essence is to intercept videos that are obviously of other types. As long as the file name is slightly close to the naming rules of film and television dramas, it will be allowed to pass as much as possible; because if the file is intercepted in the pre-classification, no subsequent processing will be performed, but if a video of other types passes here, the subsequent key information extraction and matching search can also identify that it may be other videos or no search results.

[0068] In the step S103, the target user file path is input into the trained key information extraction model, and the target key information of the target user file path is extracted by the trained key information extraction model.

[0069] The key information, i.e., the information related to the film and television series in the target user's file path, can describe the type, name, actors, number of episodes, and other characteristics of the film and television series.

[0070] The key information extraction model is a large language model (LLM).

[0071] The key information includes multiple types of keywords, and the types of keywords include: film and television drama type, film and television drama name, and film and television drama specified information; among which, the film and television drama specified information corresponding to different film and television drama types is determined based on the characteristics of the film and television drama of that type; different levels of information matching correspond to different types of keywords.

[0072] Inputting the target user file path into a trained key information extraction model, wherein the trained key information extraction model extracts multiple types of keywords from the target user file path, specifically includes the following steps:

[0073] The trained key information extraction model determines the type of film and television drama in the target user's file path;

[0074] The trained key information extraction model extracts the title of the film and television drama, and the designated information of the film and television drama that matches the type of the film and television drama.

[0075] Specifically, if LLM considers the file to be a TV series, it needs to extract the title, year, and season information. If it is judged to be a movie, it extracts the movie title and year. Information that does not appear in the file path is left blank.

[0076] Using open source or commercial LLM models in the existing technology (such as ChatGPT, etc.) alone cannot meet the requirements of accurate and fast information extraction. Based on this, in an optional embodiment, a vertical field LLM for film and television information collection services is provided. The overall steps can be divided into two parts: the construction of the key information extraction model and the online key information extraction logic.

[0077] In an optional embodiment, please refer to Figure 3 , Figure 3 A flowchart of a training method for a key information extraction model according to an embodiment of the present application is shown; the key information extraction model is based on the following steps S301-S304:

[0078] S301, obtaining sample file paths of various types of film and television dramas and other types of videos;

[0079] S302, processing the sample file path through multiple information extraction models to obtain keywords extracted from each sample file path by the multiple information extraction models;

[0080] S303, aggregating keywords extracted by multiple information extraction models for each sample file path, and determining a key information annotation result of the sample file path based on a multi-person voting method;

[0081] S304: construct a training data set based on sample file paths of various types of film and television dramas and key information annotation results corresponding to the sample file paths, train a key information extraction model, and obtain a trained key information extraction model.

[0082] Specifically, the various types of film and television dramas include TV dramas, movies, variety shows, etc.

[0083] The other types of videos are video types other than movies and TV series.

[0084] In order to construct a key information extraction model, the overall process can be divided into two parts: training data set construction and model training.

[0085] During the construction of the training set, hundreds of thousands of sample file paths were collected from various aspects. The paths here include some other files, TV series files, movie files, and variety show files.

[0086] However, these sample file data are not labeled and cannot be used to train the model.

[0087] Based on this, in an optional embodiment, a multi-model automatic data labeling solution is used to fully automatically label the collected sample file paths.

[0088] Specifically, the sample file path is processed by multiple information extraction models to obtain keywords extracted by each sample file path by multiple information extraction models; the keywords extracted by each sample file path by multiple information extraction models are aggregated, and the key information labeling result of the sample file path is determined based on a multi-person voting method.

[0089] The keywords extracted by the information extraction model include various types, such as TV series title, movie title, episode number, etc.

[0090] The types of keywords extracted by each information extraction model may be the same or different, and the specific content of each keyword may be the same or different.

[0091] The aggregating keywords extracted by multiple information extraction models for each sample file path and determining the key information annotation result of the sample file path based on a multi-person voting method includes:

[0092] Aggregate various types of keywords extracted by various information extraction models for each sample file path, and determine the annotation results of each type of keyword based on the multi-person voting method;

[0093] Based on the annotation results of each type of keyword, the key information annotation results of the sample file path are determined.

[0094] In other words, the multi-model automatic data labeling solution selects multiple commercial or open source LLM models to extract key information from each file data, then aggregates all the results and screens them through multi-person voting, and finally generates a labeling result for each file path.

[0095] Exemplarily, three information extraction models are used to extract the following key information of the file path respectively.

[0096] File path: / Backup Disk / Domestic Dramas / Cartoons / China / The Legend of Zhang San / The Legend of Zhang San: The Extraordinary Legend / Episode 58 Blu-ray (1080P)-45-14.mp4.

[0097] All three information extraction models extracted the type of the film and television series as a domestic drama; two of the information extraction models extracted the name as Zhang San Legend, and one extracted the name as an animation, so the name of the film and television series was determined to be Zhang San Legend; all three information extraction models extracted the specified information as episode 58; thus, after screening by multi-person voting, the key information annotation results of the file path were determined to be: domestic drama, Zhang San Legend, episode 58.

[0098] However, this construction method will still lead to many errors. Therefore, in an optional embodiment, a set of discrimination services is further developed based on LLM+ retrieval enhanced generation (RAG) technology to verify the correctness of each annotation. The specific approach here is: based on the key information annotation results generated by multi-person voting, the most similar results are retrieved from the pre-prepared film and television resource library, and then the retrieved real film and television information and the generated annotation results are input into the LLM together, so that it can combine this information to determine whether the generated annotation results are correct. Based on this method, we can filter out some erroneous data and ensure the correctness of the training set, thereby improving the recognition accuracy of the keyword information extraction model.

[0099] Based on this, in an optional embodiment, after determining the key information annotation result of the sample file path based on the multi-person voting method, the real film and television drama information matching the key information annotation result is retrieved from the pre-configured film and television resource library;

[0100] The retrieved real film and television drama information, sample file path, and key information annotation results corresponding to the sample file path are input into the annotation result discrimination model, and the annotation result discrimination model is used to judge whether the generated key information annotation results are correct;

[0101] Based on the discrimination results output by the labeling result discrimination model, key information labeling results and sample file paths whose discrimination results are wrong are screened out.

[0102] That is to say, based on retrieval enhanced generation (RAG), some additional and accurate real film and television drama information is introduced to assist LLM in making judgments, and accurate judgment results for key information labeling results are given, thereby filtering out sample file paths with incorrect labeling results and improving the accuracy of sample file path labeling.

[0103] Based on the sample file paths of various types of film and television dramas and the key information annotation results corresponding to the sample file paths, a training data set is constructed to train the key information extraction model to obtain a trained key information extraction model. Specifically, the instruction tuning method is introduced to fine-tune the open source LLM. The training goal is to enable the model to understand and align the key information that needs to be extracted by scraping, and to ensure that the format of the extracted information is stable.

[0104] In order to fine-tune the instructions, the training data sets are first converted into instructions, and then the model training begins. After multiple rounds of training, the best key information extraction model is obtained.

[0105] After obtaining the trained key information extraction model, the key information extraction model is used online to extract the key information of the target user's file path.

[0106] Specifically, corresponding to a given target user file path, path preprocessing is first performed to retain the final preset data-level directory information; then, the processed target user file path is input into a trained key information extraction model to extract the key information of the target user file path.

[0107] As an example only, the given file path is trimmed to only take the last five levels of directory information.

[0108] In step S104, the target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0109] The multiple levels of information matching include: recalling results in a film and television resource library based on target key information; sorting the recalled results by similarity; and filtering erroneous results based on the target key information.

[0110] Through multi-level information matching, the matching degree between the target file and the target film and television series is improved, and the accuracy of extracting film and television series information based on files is improved.

[0111] In an optional embodiment, please refer to Figure 4 , Figure 4 A flow chart of a method for determining a target film or TV series with the highest matching degree with the target key information according to an embodiment of the present application is shown; the target key information is matched with the film or TV series information in a pre-configured film or TV resource library at multiple levels to determine the target film or TV series with the highest matching degree with the target key information, including the following steps S401-S402:

[0112] S401, based on the target key information, recalling a plurality of film and television dramas matching the target key information from a pre-configured film and television resource library;

[0113] S402: Determine the target film or TV series that has the highest matching degree with the target key information from the recalled multiple film or TV series, and obtain the target film or TV series information of the target film or TV series.

[0114] In some embodiments, an ES search engine is first established, and all existing film and television drama information is written into it. Then, a DSL statement for the ES search is written based on the extracted target key information (based on the type and name in the key information, if the type is a TV series or variety show, the season information will also be added).

[0115] That is to say, based on the target key information, recalling a plurality of films and TV series matching the target key information from a pre-configured film and TV resource library includes:

[0116] Based on the film and television drama type in the target key information, determine the film and television drama that matches the film and television drama type from a pre-configured film and television resource library;

[0117] Based on the film and television drama titles in the target key information, multiple film and television dramas matching the target key information are retrieved from the film and television dramas matching the film and television drama types.

[0118] Based on this dsl statement, we can recall M potentially related TV series and movies from the ES engine.

[0119] There may be multiple languages ​​in the user's file path. Even if it is Chinese, it may be a combination of Chinese and pinyin, or some homophones may appear. If the recall is conducted based solely on the Chinese characters in the titles of TV shows and movies, the recall is likely to fail.

[0120] Based on this, in the embodiment of the present application, based on the film and television drama name in the target key information, retrieving multiple film and television dramas matching the target key information from the film and television dramas matching the film and television drama type includes:

[0121] Based on a word segmenter that matches the text type and / or language type of the film and television drama title in the target key information, the film and television drama title is segmented to determine a word segmentation result of the film and television drama title;

[0122] Based on the word segmentation result of the film and television drama name, multiple film and television dramas matching the word segmentation result are retrieved from the film and television dramas matching the film and television drama type.

[0123] The word segmenter based on the text type matching of the film and television drama title in the target key information performs word segmentation processing on the film and television drama title to determine the word segmentation result of the film and television drama title, including:

[0124] Convert the TV series name in the target key information into Chinese and Pinyin, and determine the Pinyin name corresponding to the Chinese name of the TV series name;

[0125] The pinyin name corresponding to the Chinese name of the film and television drama is segmented to determine the segmentation result of the film and television drama name.

[0126] In this way, as many TV series and movies with the same or similar names as the TV series and movies can be recalled as much as possible, increasing the probability and accuracy of matching files to TV series and movies.

[0127] For example, for Zhang San Chuan Qi, it is converted into zhangsanchuanqi, and both Zhang San Chuan Qi and zhangsanchuanqi are processed by word segmentation, and the results of word segmentation are used for recall. In this way, the final recall results include Zhang San Chuan Qi, Zhang San Chuan Qi, and so on.

[0128] After recalling multiple film and television dramas matching the target key information from a pre-configured film and television resource library, determining a target film and television drama with the highest matching degree with the target key information from the recalled multiple film and television dramas, including:

[0129] Calculate the name of the TV series in the target key information and the name similarities between the multiple TV series, and determine a preset number of TV series whose name similarities meet the preset similarity requirements;

[0130] Filter out the TV series and movies that match the specified information in the target key information from the preset number of TV series and movies;

[0131] The target film or TV series with the highest name similarity is determined from the film or TV series that match the designated information in the target key information.

[0132] Here, the specified information is other key information besides the type and title of the TV series, such as actors, year, director, and if it is a TV series or variety show, the number of episodes and / or seasons.

[0133] The preset similarity requirement may be a preset number of TV series ranked first in terms of name similarity and / or a name similarity greater than a preset similarity threshold, and so on.

[0134] In an optional embodiment, the film and television drama information collection method recalls M potentially relevant film and television dramas from ES based on the key information of the user file path, but the sorting result of ES is not very reliable, so an AI sorting model is introduced to recalculate the correlation between all the recalled film and television dramas and the extracted key information, and finally give a ranking in order of relevance.

[0135] Calculate the similarity between the names of the film and television dramas in the target key information and the names of multiple film and television dramas respectively, and input the two target texts (the names of the film and television dramas in the target key information and the names of the recalled film and television dramas) into the AI ​​sorting model to calculate a similarity score; calculate the names of the film and television dramas extracted from the key information in turn with the names of the recalled film and television dramas, and finally sort them by similarity.

[0136] In practical applications, a preset similarity threshold can be empirically set. If the similarity is higher than this threshold, it is considered that the key information and the recalled TV series are the same. In this way, a search list is finally obtained, which contains TV series that meet the preset similarity threshold.

[0137] The search list obtained based on recall and similarity calculation cannot be used as the final result. There are two main reasons. On the one hand, there are errors in the search results. On the other hand, there may be multiple search results (possibly due to the same name).

[0138] Therefore, in some embodiments, it is also necessary to combine some rules for screening.

[0139] Specifically, based on the specified information in the target key information (year, director, actor, etc., if it is a TV series, there are also episodes and seasons), etc., judge from the search list in turn whether each search result meets the specified information such as year and episode, eliminate some results that do not meet the requirements, and take the first result with the highest similarity in the final search list as the final matching result.

[0140] Determine the target film and television drama that has the highest match with the target key information, and obtain the target film and television drama information of the target film and television drama; so far, the film and television drama information collection method described in the embodiment of the present application has successfully matched a film and television drama in the film and television library from a file, and then retrieved various information of the film and television drama and returned it, thus completing the scraping process.

[0141] The target film or TV series information refers to information related to the target film or TV series, including introduction, posters, actors, etc.

[0142] The target film and television drama information is used to be displayed in the display area of ​​the target file in the graphical user interface of the multimedia player.

[0143] Please refer to Figure 5 , Figure 5 A schematic diagram of a graphical user interface of a multimedia player according to an embodiment of the present application is shown.

[0144] That is to say, the multimedia player originally could only display the file name and / or video frame, but now can simultaneously display the introduction, poster, actors and other information at the file location; the information is more, more intuitive, and has a good visual effect.

[0145] Based on the same inventive concept, the embodiment of the present application also provides a film and television drama information collection device corresponding to the film and television drama information collection method. Since the principle of solving the problem by the device in the embodiment of the present application is similar to the above-mentioned film and television drama information collection method in the embodiment of the present application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0146] Please refer to Figure 6 , Figure 6 The structure diagram of the device for collecting information about film and television dramas according to an embodiment of the present application is shown; the device for collecting information about film and television dramas, the device comprising:

[0147] An acquisition module 601 is used to acquire a user file path of a file;

[0148] The classification module 602 is used to classify the user file paths and filter out the target user file paths of the target files belonging to the film and television drama category;

[0149] An extraction module 603, used for inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0150] The matching module 604 is used to perform multiple-level information matching between the target key information and the film and television drama information in the pre-configured film and television resource library, determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0151] In some embodiments, in the device for collecting information about film and television dramas, the acquisition module, when acquiring the user file path of a file, is specifically used to:

[0152] The updating module is used to perform path clipping on the user file path of the file, retain the user file path of the last preset number of levels, and update the user file path.

[0153] In some embodiments, in the film and television drama information collection device, the classification module, when classifying the user file paths and screening out the target user file paths of the target files belonging to the film and television drama category, is specifically used to:

[0154] Performing rule pre-judgment on the user file path to filter out the target user file path of the target file belonging to the film and television drama category;

[0155] The remaining user file paths after the target user file paths are filtered out are input into a pre-trained classification model, and the target user file paths of the target files belonging to the film and television drama category in the remaining user file paths are filtered out based on the classification results of the classification model.

[0156] In some embodiments, in the film and television drama information collection device, the extraction module, when inputting the target user file path into the trained key information extraction model and extracting the target key information of the target user file path through the trained key information extraction model, is specifically used to:

[0157] Inputting the target user file path into a trained key information extraction model, wherein the trained key information extraction model extracts multiple types of keywords from the target user file path;

[0158] Based on the multiple types of keywords, target key information of the target user file path is determined.

[0159] In some embodiments, in the film and television information collection device, the types of keywords include: film and television type, film and television name, and film and television specified information; wherein, the film and television specified information corresponding to different film and television types is determined based on the characteristics of the type of film and television; different levels of information matching correspond to different types of keywords.

[0160] In some embodiments, the device for collecting information about film and television dramas further includes:

[0161] The training module is used to obtain sample file paths of various types of movies, TV series and other types of videos;

[0162] Process the sample file paths through multiple information extraction models to obtain keywords extracted from each sample file path by the multiple information extraction models;

[0163] Aggregate the keywords extracted by multiple information extraction models for each sample file path, and determine the key information annotation result of the sample file path based on the multi-person voting method;

[0164] A training data set is constructed based on sample file paths of various types of film and television dramas and key information annotation results corresponding to the sample file paths, and a key information extraction model is trained to obtain a trained key information extraction model.

[0165] In some embodiments, in the device for collecting information about film and television dramas, the training module, when aggregating keywords extracted by multiple information extraction models for each sample file path and determining the key information labeling result of the sample file path based on a multi-person voting method, is specifically used to:

[0166] Aggregate various types of keywords extracted by various information extraction models for each sample file path, and determine the annotation results of each type of keyword based on the multi-person voting method;

[0167] Based on the annotation results of each type of keyword, the key information annotation results of the sample file path are determined.

[0168] In some embodiments, in the film and television drama information collection device, the training module is further used to retrieve real film and television drama information matching the key information annotation result from a pre-configured film and television resource library after determining the key information annotation result of the sample file path based on the multi-person voting method;

[0169] The retrieved real film and television drama information, sample file path, and key information annotation results corresponding to the sample file path are input into the annotation result discrimination model, and the annotation result discrimination model is used to judge whether the generated key information annotation results are correct;

[0170] Based on the discrimination results output by the labeling result discrimination model, key information labeling results and sample file paths whose discrimination results are wrong are screened out.

[0171] In some embodiments, in the film and television drama information collection device, the matching module, when performing multiple levels of information matching between the target key information and the film and television drama information in the pre-configured film and television resource library to determine the target film and television drama with the highest matching degree with the target key information, is specifically used to:

[0172] Based on the target key information, recall multiple film and television dramas matching the target key information from a pre-configured film and television resource library;

[0173] A target film or TV series having the highest matching degree with the target key information is determined from the recalled multiple film or TV series, and target film or TV series information of the target film or TV series is obtained.

[0174] In some embodiments, in the film and television drama information collection device, the matching module, when determining the target film and television drama with the highest matching degree with the target key information from the recalled multiple film and television dramas, is specifically used to:

[0175] Calculate the name of the TV series in the target key information and the name similarities between the multiple TV series, and determine a preset number of TV series whose name similarities meet the preset similarity requirements;

[0176] Filter out the TV series and movies that match the specified information in the target key information from the preset number of TV series and movies;

[0177] The target film or TV series with the highest name similarity is determined from the film or TV series that match the designated information in the target key information.

[0178] In some embodiments, in the film and television drama information collection device, the matching module, when recalling multiple film and television dramas matching the target key information from a pre-configured film and television resource library based on the target key information, is specifically used to:

[0179] Based on the film and television drama type in the target key information, determine the film and television drama that matches the film and television drama type from a pre-configured film and television resource library;

[0180] Based on the film and television drama titles in the target key information, multiple film and television dramas matching the target key information are retrieved from the film and television dramas matching the film and television drama types.

[0181] In some embodiments, in the film and television drama information collection device, the matching module, when retrieving multiple film and television dramas matching the target key information from the film and television dramas matching the film and television drama type based on the film and television drama names in the target key information, is specifically used to:

[0182] Based on a word segmenter that matches the text type and / or language type of the film and television drama title in the target key information, the film and television drama title is segmented to determine a word segmentation result of the film and television drama title;

[0183] Based on the word segmentation result of the film and television drama name, multiple film and television dramas matching the word segmentation result are retrieved from the film and television dramas matching the film and television drama type.

[0184] In some embodiments, in the film and television drama information collection device, the matching module, when performing word segmentation processing on the film and television drama name based on the word type matching of the film and television drama name in the target key information and determining the word segmentation result of the film and television drama name, is specifically used to:

[0185] Convert the TV series name in the target key information into Chinese and Pinyin, and determine the Pinyin name corresponding to the Chinese name of the TV series name;

[0186] The pinyin name corresponding to the Chinese name of the film and television drama is segmented to determine the segmentation result of the film and television drama name.

[0187] In some embodiments, in the film and television drama information collection device, the target film and television drama information is used to display in the display area of ​​the target file in the graphical user interface of the multimedia player.

[0188] Based on the same inventive concept, an electronic device corresponding to the method for collecting film and television drama information is also provided in the embodiment of the present application. Since the principle of solving the problem by the electronic device in the embodiment of the present application is similar to the method for collecting film and television drama information in the embodiment of the present application, the implementation of the electronic device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0189] Please refer to Figure 7 , Figure 7 The structure diagram of the electronic device described in the embodiment of the present application is shown. The electronic device 700 includes: a processor 702, a memory 701 and a bus. The memory 701 stores machine-readable instructions executable by the processor 702. When the electronic device 700 is running, the processor 702 communicates with the memory 701 through the bus. When the machine-readable instructions are executed by the processor 702, the processor 702 executes the steps of the film and television drama information collection method, which specifically includes:

[0190] Get the user file path of the file;

[0191] Classify the user file paths and filter out the target user file paths of the target files belonging to the film and television drama category;

[0192] Inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0193] The target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0194] In some embodiments, when the processor executes the step of obtaining the user file path of the file, the processor specifically executes the following steps:

[0195] The user file path of the file is clipped, the user file path of the last preset number of levels is retained, and the user file path is updated.

[0196] In some embodiments, when the processor performs the step of classifying the user file paths and filtering out the target user file paths of the target files belonging to the film and television drama category, the following steps are specifically performed:

[0197] Performing rule pre-judgment on the user file path to filter out the target user file path of the target file belonging to the film and television drama category;

[0198] The remaining user file paths after the target user file paths are filtered out are input into a pre-trained classification model, and the target user file paths of the target files belonging to the film and television drama category in the remaining user file paths are filtered out based on the classification results of the classification model.

[0199] In some embodiments, when the processor executes the step of inputting the target user file path into the trained key information extraction model and extracting the target key information of the target user file path through the trained key information extraction model, the following steps are specifically performed:

[0200] Inputting the target user file path into a trained key information extraction model, wherein the trained key information extraction model extracts multiple types of keywords from the target user file path;

[0201] Based on the multiple types of keywords, target key information of the target user file path is determined.

[0202] In some embodiments, the types of keywords include: film and television drama type, film and television drama name, and film and television drama specified information; among which, the film and television drama specified information corresponding to different film and television drama types is determined based on the characteristics of the type of film and television drama; different levels of information matching correspond to different types of keywords.

[0203] In some embodiments, the processor further performs the following steps:

[0204] Get sample file paths for various types of movies, TV series and other types of videos;

[0205] Process the sample file paths through multiple information extraction models to obtain keywords extracted from each sample file path by the multiple information extraction models;

[0206] Aggregate the keywords extracted by multiple information extraction models for each sample file path, and determine the key information annotation result of the sample file path based on the multi-person voting method;

[0207] A training data set is constructed based on sample file paths of various types of film and television dramas and key information annotation results corresponding to the sample file paths, and a key information extraction model is trained to obtain a trained key information extraction model.

[0208] In some embodiments, when the processor performs the step of aggregating keywords extracted by multiple information extraction models for each sample file path and determining the key information annotation result of the sample file path based on a multi-person voting method, the following steps are specifically performed:

[0209] Aggregate various types of keywords extracted by various information extraction models for each sample file path, and determine the annotation results of each type of keyword based on the multi-person voting method;

[0210] Based on the annotation results of each type of keyword, the key information annotation results of the sample file path are determined.

[0211] In some embodiments, after the processor executes the step of determining the key information annotation result of the sample file path based on the multi-person voting method, the processor further executes the following steps:

[0212] Retrieving real film and television drama information matching the key information annotation result from a pre-configured film and television resource library;

[0213] The retrieved real film and television drama information, sample file path, and key information annotation results corresponding to the sample file path are input into the annotation result discrimination model, and the annotation result discrimination model is used to judge whether the generated key information annotation results are correct;

[0214] Based on the discrimination results output by the labeling result discrimination model, key information labeling results and sample file paths whose discrimination results are wrong are screened out.

[0215] In some embodiments, when the processor performs the step of performing multiple-level information matching between the target key information and the film and television drama information in the pre-configured film and television resource library to determine the target film and television drama with the highest matching degree with the target key information, the following steps are specifically performed:

[0216] Based on the target key information, recall multiple film and television dramas matching the target key information from a pre-configured film and television resource library;

[0217] A target film or TV series having the highest matching degree with the target key information is determined from the recalled multiple film or TV series, and target film or TV series information of the target film or TV series is obtained.

[0218] In some embodiments, when the processor executes the step of determining the target film or TV series with the highest matching degree with the target key information from the recalled multiple film or TV series, the processor specifically executes the following steps:

[0219] Calculate the name of the TV series in the target key information and the name similarities between the multiple TV series, and determine a preset number of TV series whose name similarities meet the preset similarity requirements;

[0220] Filter out the TV series and movies that match the specified information in the target key information from the preset number of TV series and movies;

[0221] The target film or TV series with the highest name similarity is determined from the film or TV series that match the designated information in the target key information.

[0222] In some embodiments, when the processor executes the step of recalling a plurality of film and television dramas matching the target key information from a pre-configured film and television resource library based on the target key information, the following steps are specifically performed:

[0223] Based on the film and television drama type in the target key information, determine the film and television drama that matches the film and television drama type from a pre-configured film and television resource library;

[0224] Based on the film and television drama titles in the target key information, multiple film and television dramas matching the target key information are retrieved from the film and television dramas matching the film and television drama types.

[0225] In some embodiments, when the processor executes the step of retrieving multiple film and television dramas matching the target key information from the film and television dramas matching the film and television drama type based on the film and television drama names in the target key information, the following steps are specifically performed:

[0226] Based on a word segmenter that matches the text type and / or language type of the film and television drama title in the target key information, the film and television drama title is segmented to determine a word segmentation result of the film and television drama title;

[0227] Based on the word segmentation result of the film and television drama name, multiple film and television dramas matching the word segmentation result are retrieved from the film and television dramas matching the film and television drama type.

[0228] In some embodiments, the processor executes the word segmenter based on the text type matching of the film and television drama name in the target key information, performs word segmentation processing on the film and television drama name, and determines the word segmentation result of the film and television drama name, specifically performing the following steps:

[0229] Convert the TV series name in the target key information into Chinese and Pinyin, and determine the Pinyin name corresponding to the Chinese name of the TV series name;

[0230] The pinyin name corresponding to the Chinese name of the film and television drama is segmented to determine the segmentation result of the film and television drama name.

[0231] In some embodiments, the target film or TV series information is used to be displayed in a display area of ​​the target file in a graphical user interface of a multimedia player.

[0232] Based on the same inventive concept, a computer-readable storage medium corresponding to the method for collecting film and television drama information is also provided in the embodiment of the present application. Since the principle of solving the problem by the computer-readable storage medium in the embodiment of the present application is similar to the method for collecting film and television drama information in the embodiment of the present application, the implementation of the computer-readable storage medium can refer to the implementation of the method, and the repeated parts will not be repeated.

[0233] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor performs the steps of a method for collecting film and television drama information, which are specifically as follows:

[0234] Get the user file path of the file;

[0235] Classify the user file paths and filter out the target user file paths of the target files belonging to the film and television drama category;

[0236] Inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model;

[0237] The target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

[0238] In some embodiments, when the processor executes the step of obtaining the user file path of the file, the processor specifically executes the following steps:

[0239] The user file path of the file is clipped, the user file path of the last preset number of levels is retained, and the user file path is updated.

[0240] In some embodiments, when the processor performs the step of classifying the user file paths and filtering out the target user file paths of the target files belonging to the film and television drama category, the following steps are specifically performed:

[0241] Performing rule pre-judgment on the user file path to filter out the target user file path of the target file belonging to the film and television drama category;

[0242] The remaining user file paths after the target user file paths are filtered out are input into a pre-trained classification model, and the target user file paths of the target files belonging to the film and television drama category in the remaining user file paths are filtered out based on the classification results of the classification model.

[0243] In some embodiments, when the processor executes the step of inputting the target user file path into the trained key information extraction model and extracting the target key information of the target user file path through the trained key information extraction model, the following steps are specifically performed:

[0244] Inputting the target user file path into a trained key information extraction model, wherein the trained key information extraction model extracts multiple types of keywords from the target user file path;

[0245] Based on the multiple types of keywords, target key information of the target user file path is determined.

[0246] In some embodiments, the types of keywords include: film and television drama type, film and television drama name, and film and television drama specified information; among which, the film and television drama specified information corresponding to different film and television drama types is determined based on the characteristics of the type of film and television drama; different levels of information matching correspond to different types of keywords.

[0247] In some embodiments, the processor further performs the following steps:

[0248] Get sample file paths for various types of movies, TV series and other types of videos;

[0249] Process the sample file paths through multiple information extraction models to obtain keywords extracted from each sample file path by the multiple information extraction models;

[0250] Aggregate the keywords extracted by multiple information extraction models for each sample file path, and determine the key information annotation result of the sample file path based on the multi-person voting method;

[0251] A training data set is constructed based on sample file paths of various types of film and television dramas and key information annotation results corresponding to the sample file paths, and a key information extraction model is trained to obtain a trained key information extraction model.

[0252] In some embodiments, when the processor performs the step of aggregating keywords extracted by multiple information extraction models for each sample file path and determining the key information annotation result of the sample file path based on a multi-person voting method, the following steps are specifically performed:

[0253] Aggregate various types of keywords extracted by various information extraction models for each sample file path, and determine the annotation results of each type of keyword based on the multi-person voting method;

[0254] Based on the annotation results of each type of keyword, the key information annotation results of the sample file path are determined.

[0255] In some embodiments, after the processor executes the step of determining the key information annotation result of the sample file path based on the multi-person voting method, the processor further executes the following steps:

[0256] Retrieving real film and television drama information matching the key information annotation result from a pre-configured film and television resource library;

[0257] The retrieved real film and television drama information, sample file path, and key information annotation results corresponding to the sample file path are input into the annotation result discrimination model, and the annotation result discrimination model is used to judge whether the generated key information annotation results are correct;

[0258] Based on the discrimination results output by the labeling result discrimination model, key information labeling results and sample file paths whose discrimination results are wrong are screened out.

[0259] In some embodiments, when the processor performs the step of performing multiple-level information matching between the target key information and the film and television drama information in the pre-configured film and television resource library to determine the target film and television drama with the highest matching degree with the target key information, the following steps are specifically performed:

[0260] Based on the target key information, recall multiple film and television dramas matching the target key information from a pre-configured film and television resource library;

[0261] A target film or TV series having the highest matching degree with the target key information is determined from the recalled multiple film or TV series, and target film or TV series information of the target film or TV series is obtained.

[0262] In some embodiments, when the processor executes the step of determining the target film or TV series with the highest matching degree with the target key information from the recalled multiple film or TV series, the processor specifically executes the following steps:

[0263] Calculate the name of the TV series in the target key information and the name similarities between the multiple TV series, and determine a preset number of TV series whose name similarities meet the preset similarity requirements;

[0264] Filter out the TV series and movies that match the specified information in the target key information from the preset number of TV series and movies;

[0265] The target film or TV series with the highest name similarity is determined from the film or TV series that match the designated information in the target key information.

[0266] In some embodiments, when the processor executes the step of recalling a plurality of film and television dramas matching the target key information from a pre-configured film and television resource library based on the target key information, the following steps are specifically performed:

[0267] Based on the film and television drama type in the target key information, determine the film and television drama that matches the film and television drama type from a pre-configured film and television resource library;

[0268] Based on the film and television drama titles in the target key information, multiple film and television dramas matching the target key information are retrieved from the film and television dramas matching the film and television drama types.

[0269] In some embodiments, when the processor executes the step of retrieving multiple film and television dramas matching the target key information from the film and television dramas matching the film and television drama type based on the film and television drama names in the target key information, the following steps are specifically performed:

[0270] Based on a word segmenter that matches the text type and / or language type of the film and television drama title in the target key information, the film and television drama title is segmented to determine a word segmentation result of the film and television drama title;

[0271] Based on the word segmentation result of the film and television drama name, multiple film and television dramas matching the word segmentation result are retrieved from the film and television dramas matching the film and television drama type.

[0272] In some embodiments, the processor executes the word segmenter based on the text type matching of the film and television drama name in the target key information, performs word segmentation processing on the film and television drama name, and determines the word segmentation result of the film and television drama name, specifically performing the following steps:

[0273] Convert the TV series name in the target key information into Chinese and Pinyin, and determine the Pinyin name corresponding to the Chinese name of the TV series name;

[0274] The pinyin name corresponding to the Chinese name of the film and television drama is segmented to determine the segmentation result of the film and television drama name.

[0275] In some embodiments, the target film or TV series information is used to be displayed in a display area of ​​the target file in a graphical user interface of a multimedia player.

[0276] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0277] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0278] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0279] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a platform server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0280] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for collecting information on film and television dramas, characterized in that: The method comprises: Get the user file path of the file; Classify the user file paths and filter out the target user file paths of the target files belonging to the film and television drama category; Inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model; The target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

2. The method for collecting film and television drama information according to claim 1, characterized in that: The user file path of the obtained file includes: The user file path of the file is clipped, the user file path of the last preset number of levels is retained, and the user file path is updated.

3. The method for collecting film and television drama information according to claim 1 or 2, characterized in that: The method of classifying the user file paths and filtering out the target user file paths of the target files belonging to the film and television drama category includes: Performing rule pre-judgment on the user file path to filter out the target user file path of the target file belonging to the film and television drama category; The remaining user file paths after the target user file paths are filtered out are input into a pre-trained classification model, and the target user file paths of the target files belonging to the film and television drama category in the remaining user file paths are filtered out based on the classification results of the classification model.

4. The method for collecting film and television drama information according to claim 1, characterized in that: The step of inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model, includes: Inputting the target user file path into a trained key information extraction model, wherein the trained key information extraction model extracts multiple types of keywords from the target user file path; Based on the multiple types of keywords, target key information of the target user file path is determined.

5. The method for collecting film and television drama information according to claim 4, characterized in that: The types of keywords include: film and television drama type, film and television drama name, and film and television drama specified information; among which, the film and television drama specified information corresponding to different film and television drama types is determined based on the characteristics of the film and television drama of that type; different levels of information matching correspond to different types of keywords.

6. The method for collecting film and television drama information according to claim 4, characterized in that: The key information extraction model is trained based on the following method: Get sample file paths for various types of movies, TV series and other types of videos; Process the sample file paths through multiple information extraction models to obtain keywords extracted from each sample file path by the multiple information extraction models; Aggregate the keywords extracted by multiple information extraction models for each sample file path, and determine the key information annotation result of the sample file path based on the multi-person voting method; A training data set is constructed based on sample file paths of various types of film and television dramas and key information annotation results corresponding to the sample file paths, and a key information extraction model is trained to obtain a trained key information extraction model.

7. The method for collecting film and television drama information according to claim 6, characterized in that: The aggregating keywords extracted by multiple information extraction models for each sample file path and determining the key information annotation result of the sample file path based on a multi-person voting method includes: Aggregate various types of keywords extracted by various information extraction models for each sample file path, and determine the annotation results of each type of keyword based on the multi-person voting method; Based on the annotation results of each type of keyword, the key information annotation results of the sample file path are determined.

8. The method for collecting film and television drama information according to claim 6 or 7, characterized in that: After determining the key information annotation result of the sample file path based on the multi-person voting method, retrieving the real film and television drama information matching the key information annotation result from the pre-configured film and television resource library; The retrieved real film and television drama information, sample file path, and key information annotation results corresponding to the sample file path are input into the annotation result discrimination model, and the annotation result discrimination model is used to judge whether the generated key information annotation results are correct; Based on the discrimination results output by the labeling result discrimination model, key information labeling results and sample file paths whose discrimination results are wrong are screened out.

9. The method for collecting film and television drama information according to claim 1, 4 or 5, characterized in that: The target key information is matched with the film and television drama information in the pre-configured film and television resource library at multiple levels to determine the target film and television drama with the highest matching degree with the target key information, including: Based on the target key information, recall multiple film and television dramas matching the target key information from a pre-configured film and television resource library; A target film or TV series having the highest matching degree with the target key information is determined from the recalled multiple film or TV series, and target film or TV series information of the target film or TV series is obtained.

10. The method for collecting film and television drama information according to claim 9, characterized in that: The determining of the target film and television drama with the highest matching degree with the target key information from the recalled multiple film and television dramas includes: Calculate the name of the TV series in the target key information and the name similarities between the multiple TV series, and determine a preset number of TV series whose name similarities meet the preset similarity requirements; Filter out the TV series and movies that match the specified information in the target key information from the preset number of TV series and movies; The target film or TV series with the highest name similarity is determined from the film or TV series that match the designated information in the target key information.

11. The method for collecting film and television drama information according to claim 9, characterized in that: The step of recalling a plurality of films and TV series matching the target key information from a pre-configured film and TV resource library based on the target key information includes: Based on the film and television drama type in the target key information, determine the film and television drama that matches the film and television drama type from a pre-configured film and television resource library; Based on the film and television drama titles in the target key information, multiple film and television dramas matching the target key information are retrieved from the film and television dramas matching the film and television drama types.

12. The method for collecting film and television drama information according to claim 11, characterized in that: The retrieving a plurality of film and television dramas matching the target key information from the film and television dramas matching the film and television drama types based on the film and television drama names in the target key information includes: Based on a word segmenter that matches the text type and / or language type of the film and television drama title in the target key information, the film and television drama title is segmented to determine a word segmentation result of the film and television drama title; Based on the word segmentation result of the film and television drama name, multiple film and television dramas matching the word segmentation result are retrieved from the film and television dramas matching the film and television drama type.

13. The method for collecting film and television drama information according to claim 12, characterized in that: The word segmenter based on the text type matching of the film and television drama title in the target key information performs word segmentation processing on the film and television drama title to determine the word segmentation result of the film and television drama title, including: Convert the TV series name in the target key information into Chinese and Pinyin, and determine the Pinyin name corresponding to the Chinese name of the TV series name; The pinyin name corresponding to the Chinese name of the film and television drama is segmented to determine the segmentation result of the film and television drama name.

14. The method for collecting film and television drama information according to claim 1, characterized in that: The target film and television drama information is used to be displayed in the display area of ​​the target file in the graphical user interface of the multimedia player.

15. A device for collecting information about film and television dramas, characterized in that: The device comprises: The acquisition module is used to obtain the user file path of the file; A classification module is used to classify user file paths and filter out target user file paths of target files belonging to the film and television drama category; An extraction module, used for inputting the target user file path into a trained key information extraction model, and extracting target key information of the target user file path through the trained key information extraction model; The matching module is used to perform multiple-level information matching between the target key information and the film and television drama information in the pre-configured film and television resource library, determine the target film and television drama with the highest matching degree with the target key information, and obtain the target film and television drama information of the target film and television drama.

16. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of a method for collecting film and television drama information as described in any one of claims 1 to 14 are performed.

17. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of a method for collecting film and television drama information as described in any one of claims 1 to 14.

Images