Method and device for automatically notifying user in network cutover

Abstract

The invention provides a method and device for automatically notifying a user through network cutover, and the method comprises the steps: obtaining all target network cutover work order attachments, analyzing the target network cutover work order attachments, obtaining a corresponding analysis field text, and enabling the analysis field text to at least comprise a customer name and cutover information which are affected by a network cutover operation; vectorizing all the customer names to obtain customer name vectors corresponding to the customer names; performing similarity calculation on all the customer name vectors and the standard customer name vector to obtain corresponding similarity values; the target clients with the similarity values larger than a set similarity threshold value are extracted, cutover notifications are sent to all the target clients, and the cutover notifications are used for informing the target clients of network cutover operation and corresponding cutover information in advance. The problem of low efficiency caused by manual auditing and manual notification sending during network cutover in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of network crawlers, and in particular to a method for automatically notifying users of network cutover, an apparatus for automatically notifying users of network cutover, a computer-readable storage medium, and a computer program product. Background Art

[0002] There are many inconveniences and inefficiencies in the current process of notifying customers of network cutover. First, staff need to log in to multiple different systems and manually download the attachments of the network cutover work order. This process is not only time-consuming, but also prone to information omission or confusion due to switching between systems. Second, after downloading the attachments, staff need to identify and select the customers who need to be notified one by one. This step relies on manual identification, which is not only inefficient but also prone to errors. Subsequently, staff also need to jump to the CRM (Customer Relationship Management) system to retrieve and obtain the contact information of these customers. This process is also cumbersome and prone to errors. Finally, the time of network cutover and the possible impact on the business are not only inefficient, but also difficult to ensure the timeliness and accuracy of the notification. Overall, this process relies entirely on manual turnover and interspersed between multiple systems, which is not only inefficient but also has a low accuracy rate. Summary of the invention

[0003] The main purpose of the present application is to provide a method for automatically notifying users of network cutover, an apparatus for automatically notifying users of network cutover, a computer-readable storage medium, and a computer program product, so as to at least solve the problem of low efficiency caused by manual review and manual notification during network cutover in the prior art.

[0004] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for automatically notifying users of network cutover is provided, comprising: obtaining all target network cutover work order attachments, and parsing the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, the target network cutover operation is a network cutover operation to be performed, the parsed field text at least includes the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number and the cutover start time of the network cutover operation. and the cutover end time; vectorizing all the customer names to obtain customer name vectors corresponding to each customer name, one customer name corresponding to one customer name vector; calculating the similarity between all the customer name vectors and the standard customer name vector to obtain corresponding similarity values, wherein the similarity values ​​represent the degree of matching between the customer name vector and the standard customer name vector; extracting the target customers whose similarity values ​​are greater than the set similarity threshold, and sending a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0005] Optionally, all target network cutover work order attachments are obtained and parsed to obtain parsed field texts, including: automatically downloading the target network cutover work order attachments related to the target network cutover operation through a web crawler script; calling a semantic embedding model to perform semantic recognition on the target network cutover work order attachments to parse the column field content in the target network cutover work order attachments, and extracting the parsed field texts, wherein the column field content is the network line number, the customer name, the network cutover type, the cutover start time, and the cutover end time. The semantic embedding model is a convergence model obtained by iteratively training a large language model using a deep learning algorithm using sample training data, and the sample training data at least includes a sample customer name and a sample similarity corresponding to the sample customer name, and the sample similarity is the cosine similarity between the vector corresponding to the sample customer name and the standard customer name vector.

[0006] Optionally, all of the customer names are vectorized to obtain customer name vectors corresponding to each of the customer names, including: using a semantic embedding model to vector encode all of the customer names to generate the customer name vectors corresponding to each of the customer names, the semantic embedding model is a convergent model obtained by iteratively training a large language model using a deep learning algorithm using sample training data, the sample training data at least includes sample customer names and sample similarities corresponding to the sample customer names, and the customer name vector is an intermediate variable generated by the semantic embedding model.

[0007] Optionally, all the customer names are vectorized to obtain corresponding customer name vectors, and the method further includes: when the customer name contains multiple words, each word is vectorized using a vector aggregation strategy to obtain a word vector, and the vector aggregation strategy is at least one of the following: an average pooling strategy, a weighted average pooling strategy, and a maximum pooling strategy; all the word vectors are combined according to the word order to obtain the customer name vector.

[0008] Optionally, after all the customer names are vectorized to obtain corresponding customer name vectors, the method further includes: storing the customer name vectors in a vector database of the semantic embedding model.

[0009] Optionally, similarity calculation is performed on all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, including: substituting the customer name vector and the standard customer name vector into a first formula in the semantic embedding model for calculation to obtain the corresponding similarity value, wherein the first formula is a calculation formula for similarity calculation in the semantic embedding model, and the first formula is Similarity(A,B) represents the similarity value, A represents the customer name vector, B represents the standard customer name vector, n represents the total number of elements in the customer name vector or the standard customer name vector, and i represents the element number in the customer name vector or the standard customer name vector.

[0010] Optionally, sending a cutover notification to all the target customers includes: querying the contact information corresponding to each of the target customers; automatically calling a service interface to automatically distribute the cutover notification to the corresponding target customers through the contact information, and the service interface is at least one of the following: a text message interface, a voice interface, and a mini-program interface.

[0011] According to another aspect of the present application, a device for automatically notifying a user of a network cutover is provided, the device comprising: an acquisition and parsing unit, configured to acquire all target network cutover work order attachments, and parse the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with a target network cutover operation in a network cutover work order stored in a system, the target network cutover operation is a network cutover operation to be performed, the parsed field texts at least include a customer name and cutover information affected by the network cutover operation, and the cutover information at least includes a network line number, a cutover start time, and a cutover end time of the network cutover operation; A vectorization unit is used to vectorize all the customer names to obtain customer name vectors corresponding to each customer name, where one customer name corresponds to one customer name vector; a calculation unit is used to calculate the similarity between all the customer name vectors and the standard customer name vector to obtain corresponding similarity values, wherein the similarity values ​​represent the degree of matching between the customer name vectors and the standard customer name vectors; a notification unit is used to extract target customers whose similarity values ​​are greater than a set similarity threshold, and send a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation will be performed and the corresponding cutover information.

[0012] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the methods described.

[0013] According to another aspect of the present application, a computer program product is provided, comprising computer instructions, wherein when the computer instructions are executed by a processor, any one of the methods described above is implemented.

[0014] By applying the technical solution of the present application, in a method for automatically notifying a user of a network cutover, first, all target network cutover work order attachments are obtained, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is a network cutover operation to be executed; then, the target network cutover work order attachments are parsed to obtain corresponding parsed field texts, wherein the parsed field texts at least include the customer name affected by the network cutover operation and the cutover information, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation; Then, all the customer names are vectorized to obtain customer name vectors corresponding to each customer name, and one customer name corresponds to one customer name vector; then, all the customer name vectors are similarity calculated with the standard customer name vector to obtain corresponding similarity values, and the similarity values ​​represent the degree of matching between the customer name vector and the standard customer name vector; finally, the target customers whose similarity values ​​are greater than the set similarity threshold are extracted, and a cutover notification is sent to all the target customers, and the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed. This application automatically crawls the attachments of the network cutover work order (i.e., the target network cutover work order attachments), parses the target network cutover work order attachments, extracts the customer name list affected by this network cutover and the cutover information such as the relevant time period, content information, etc.; vectorizes and stores the customer name list; calculates the vector similarity between the obtained customer name list and the customer names already in the customer management system, and the customer whose similarity value is greater than the set similarity threshold (80% or other value) is regarded as the target customer, and the system automatically calls the SMS, voice and other automatic push interface services to push the cutover information to the target customer. This application solves the problem of low efficiency caused by the need for manual review and manual notification sending during network cutover in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for automatically notifying a user of network cutover provided in an embodiment of the present application is shown;

[0016] Figure 2 A flow chart of a method for automatically notifying a user of network cutover provided in accordance with an embodiment of the present application is shown;

[0017] Figure 3 A schematic diagram showing a method for automatically notifying a user of network cutover according to an embodiment of the present application is shown;

[0018] Figure 4A schematic diagram showing the compilation result of a Python script for automatically downloading a network cutover work order attachment according to an embodiment of the present application is shown;

[0019] Figure 5 A structural block diagram of a device for automatically notifying a user of a network cutover provided according to an embodiment of the present application is shown.

[0020] The above drawings include the following reference numerals:

[0021] 102, processor; 104, memory; 106, transmission device; 108, input and output devices. DETAILED DESCRIPTION

[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution 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. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0026] Network cutover: operations on the lines and devices in use.

[0027] As introduced in the background technology, in the prior art, staff need to log in to multiple different systems and manually download the attachments of the network cutover work order. Secondly, after downloading the attachments, staff need to identify and select the customers who need to be notified one by one. Finally, they need to notify the customers one by one of the time of network cutover and the possible impact on the business by phone. In order to solve the problem of low efficiency caused by manual review and manual sending of notifications during network cutover in the prior art, the embodiments of the present application provide a method for automatically notifying users of network cutover, an apparatus for automatically notifying users of network cutover, a computer-readable storage medium and a computer program product.

[0028] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0029] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal for executing a method for automatically notifying a user of network cutover according to an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0030] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for automatically notifying users of network cutover in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the mobile terminal via a network. Examples of the above networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof. The transmission device 106 is used to receive or send data via a network. The specific examples of the above networks may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0031] In this embodiment, a method for automatically notifying users of network cutover running on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0032] Figure 2 FIG. 1 is a flow chart of a method for automatically notifying a user of network cutover according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0033] Step S201, obtain all target network cutover work order attachments, and parse the above target network cutover work order attachments to obtain corresponding parsed field texts, wherein the above target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the above target network cutover operation is a network cutover operation to be executed, and the above parsed field texts at least include the customer name and cutover information affected by the above network cutover operation, and the above cutover information at least includes the network line number, the cutover start time and the cutover end time of the above network cutover operation.

[0034] Specifically, Figure 3 As shown in the figure, a Python web crawler script is written to automatically log in and crawl all work order attachments related to the target network cutover operation in the government and enterprise network operation and maintenance system, including files in PDF, Word, Excel and other formats. You can use corresponding document parsing tools, such as PDFMiner, Python-docx or pandas, or directly use the semantic embedding model to parse the downloaded attachments and extract key field information, including the name of the customer affected by the cutover, the network line number, the start and end time of the cutover, etc. This avoids the tedious process of manually downloading and parsing attachments, and improves the efficiency and accuracy of data acquisition. It ensures that all customer information related to the cutover is completely extracted, providing a comprehensive data foundation for subsequent vectorization and matching.

[0035] Step S202, vectorize all the above customer names to obtain customer name vectors corresponding to each of the above customer names, where one customer name corresponds to one customer name vector.

[0036] Specifically, a pre-trained semantic embedding model (such as BCE-Embedding) is called to process each extracted customer name and convert it into a vector representation of a fixed dimension. The generated customer name vector is stored in the vector database in the system for subsequent similarity calculation and retrieval. The unstructured customer name text is converted into a numerical vector for easy machine understanding and processing. Through vectorization, the model can capture the semantic information in the customer name, including lexical meaning and contextual connection, providing a semantic basis for similarity calculation.

[0037] Step S203 , performing similarity calculation on all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector.

[0038] Specifically, obtain the canonical customer name vector from the CRM system, and use the similarity search function of the vector database to calculate the similarity value between the customer name vector and the canonical customer name vector, such as through the cosine similarity formula. Set a similarity threshold (such as 80%) as a standard for judging the matching degree of the customer name. By calculating the similarity value, it is possible to find the canonical customer information with the highest matching degree with the customer name affected by the cutover. Using the efficient retrieval capability of the vector database, the matching process is significantly accelerated and the processing time is reduced.

[0039] Step S204, extracting the target customers whose similarity values ​​are greater than the set similarity threshold, and sending a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0040] Specifically, from the similarity calculation results, customer name vectors with similarity values ​​higher than the threshold are selected, and the customers corresponding to these vectors are the target customers that need to be notified. Automatically generate a cutover notification text containing cutover information, and automatically call SMS, voice or mini-program interfaces through the system to send a cutover notification to the selected target customers, informing them in advance of the time and possible impact of network cutover. Through automated notification, ensure that customers can prepare for network cutover in advance and reduce the impact of business interruptions. Timely and accurate notifications reduce customer complaints caused by poor information flow and improve customers' service perception.

[0041] The core points of this application include the application of automated notification services for network cutover, the use of automated crawling of attachments, intelligent identification of affected users, and a fast and accurate notification mechanism that actively associates CRM user data. By writing an efficient crawler program, attachments of network cutover work orders are automatically downloaded from multiple systems without manual intervention. Then, using advanced natural language processing and machine learning technologies, user groups affected by network cutover are intelligently identified. Finally, the system automatically associates this user information with the data in the CRM system, quickly filters out contact information, and actively and accurately informs customers of the time, scope of impact, and response measures of the network cutover through preset communication channels, such as SMS, email, or phone.

[0042] Corresponding beneficial effects:

[0043] 1. Improve generation efficiency and flexibility: Automatically identify and associate CRM data through automated web crawler scripts and customer names, effectively avoid the complex workflow of manual processing between multiple systems, and achieve "one-stop" automation.

[0044] 2. Enhanced accuracy: Use the large language model BCE-Embedding semantic embedding model to perform named entity recognition similarity matching to automatically associate accurate CRM customer relationships and automatically send SMS notifications.

[0045] 3. Promote innovation and application: The present invention provides more convenient and efficient tool support for network operation and maintenance, which helps to promote innovation and development in related fields.

[0046] In summary, the automation and intelligence level of automatic notification of network cutover to customers has been significantly improved, while reducing labor input costs and eliminating invalid and cumbersome operating procedures.

[0047] In this embodiment, first, all target network cutover work order attachments are obtained, the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is the network cutover operation to be executed; then, the target network cutover work order attachments are parsed to obtain corresponding parsed field texts, the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation; then, all the customer names are parsed. The above customer names are vectorized to obtain customer name vectors corresponding to each of the above customer names, and one customer name corresponds to one customer name vector; then, all the above customer name vectors are similarly calculated with the standard customer name vector to obtain the corresponding similarity value, and the above similarity value represents the matching degree between the above customer name vector and the above standard customer name vector; finally, the target customers corresponding to the above similarity value greater than the set similarity threshold are extracted, and a cutover notice is sent to all the above target customers, and the above cutover notice is used to inform the above target customers in advance that the above network cutover operation and the corresponding above cutover information will be performed. This application automatically crawls the attachments of the network cutover work order (i.e., the target network cutover work order attachments), parses the target network cutover work order attachments, extracts the customer name list affected by this network cutover and the related time period, content information and other cutover information; vectorizes and stores the customer name list; calculates the vector similarity between the obtained customer name list and the customer names already in the customer management system, and the customers whose similarity values ​​are greater than the set similarity threshold (80% or other values) are used as target customers, and the system automatically calls SMS, voice and other automatic push interface services to push the cutover information to the target customers. The present application solves the problem in the prior art that manual review and manual notification are required during network cutover, resulting in low efficiency.

[0048] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the method for automatically notifying users of network cutover of the present application will be described in detail below in conjunction with specific embodiments.

[0049] In order to improve the automation and efficiency of downloading and processing the attachment information of the cutover work order, in an optional implementation manner, the above step S201 includes:

[0050] Step S2011, automatically downloading the target network cutover work order attachment related to the target network cutover operation through a web crawler script;

[0051] Step S2012, calling the semantic embedding model to perform semantic recognition on the above-mentioned target network cutover work order attachment to parse the column field content in the above-mentioned target network cutover work order attachment, and extracting the above-mentioned parsed field text, the above-mentioned column field content is the above-mentioned network line number, the above-mentioned customer name, the network cutover type, the above-mentioned cutover start time and the above-mentioned cutover end time, the above-mentioned semantic embedding model is a convergence model obtained by iteratively training a large language model using a deep learning algorithm using sample training data, the above-mentioned sample training data at least includes a sample customer name and a sample similarity corresponding to the above-mentioned sample customer name, and the above-mentioned sample similarity is the cosine similarity between the vector corresponding to the above-mentioned sample customer name and the above-mentioned standard customer name vector.

[0052] In the above embodiment, if Figure 4 As shown, the attachments in the network cutover work order are crawled in the comprehensive scheduling system through the Python web crawler script to obtain the attachments of the target network cutover work order, in order to prepare for obtaining the data such as the users of the affected business list in the attachments.

[0053] The Python code snippet is as follows:

[0054] import os

[0055] import requests

[0056] from bs4 import BeautifulSoup

[0057] #Target website URL

[0058] url='http: / / 10.152.8.202:30028 / '

[0059] check_requests(url)

[0060] #Traverse the download link and download the file

[0061] for file_url in file_urls:

[0062] file_response=requests.get(file_url,headers=headers,stream=True)

[0063] file_name=os.path.basename(file_url)

[0064] file_path=os.path.join(download_dir,file_name)

[0065] print('Start downloading file')

[0066] with open(file_path,'wb')as f:

[0067] for chunk in file_response.iter_content(chunk_size=1024):

[0068] if chunk:

[0069] f.write(chunk)

[0070] print(f'File {file_name} download completed')

[0071] num++;

[0072] print('All files downloaded')

[0073] def check_requests(url):

[0074] try:

[0075] response=requests.get(url,timeout=10)

[0076] response.raise_for_status()

[0077]

[0078] The downloaded work order attachments may be in PDF, Word, or Excel formats, and the corresponding document parsing libraries (such as pdfplumber, docx, or pandas) need to be used to convert the unstructured data into processable text or table formats. Next, the parsed columns in the affected user list are parsed through the BCE-Embedding semantic embedding model named entity recognition (NER), mainly customer names, and customer name recognition and extraction.

[0079] Column fields: cutover start time, cutover end time, circuit number, customer name, impact duration, cutover type, and code snippet are as follows:

[0080]

[0081] continue

[0082] cnt+=1

[0083] question=f"""

[0084] By applying named entity recognition (NER) technology, key fields such as network line number, customer name, network cutover type, cutover start time, and cutover end time can be intelligently identified. Through the semantic embedding model, the system can intelligently parse the text in the cutover work order attachment, and accurately identify key fields even if the information is not standardized or ambiguous, thus improving the accuracy and robustness of information extraction.

[0085] Through the above embodiments, the entire system can automatically and efficiently process the attachment information of the cutover work order, improve the accuracy and timeliness of the automatic notification of network cutover to customers, thereby reducing the impact of cutover on the business of government and enterprise customers, and improving service perception and customer satisfaction.

[0086] In order to improve the accuracy and robustness of customer name recognition, in an optional implementation, the above step S202 includes:

[0087] Step S2021, use the semantic embedding model to vector encode all the above-mentioned customer names, and generate the above-mentioned customer name vectors corresponding to each of the above-mentioned customer names. The above-mentioned semantic embedding model is a convergent model obtained by iteratively training a large language model using a deep learning algorithm using sample training data. The above-mentioned sample training data at least includes sample customer names and sample similarities corresponding to the above-mentioned sample customer names. The above-mentioned customer name vector is an intermediate variable generated by the above-mentioned semantic embedding model.

[0088] In the above embodiment, the collected field information (main field customer name) data is converted into a form suitable for storage in a vector database. For each customer name extracted from the target network cutover work order attachment, a semantic embedding model is used for vector encoding, and each customer name is converted into its corresponding vector representation. The model outputs a vector of fixed length, which carries the semantic features of the customer name. The generated customer name vector is an intermediate variable, which does not directly represent the text information of the customer name, but captures the semantic features of the customer name in the form of a numerical vector. These vectors can be stored in a vector database to facilitate subsequent similarity calculations and customer matching. By converting customer names into vectors, the system can identify and match customer names based on semantics, rather than just literal text matching. This helps to handle abbreviations, aliases, and spelling differences in customer names, and improves the accuracy and robustness of customer name recognition.

[0089] In order to improve the recognition and matching accuracy of complex customer names, in an optional implementation, the above step S202 further includes:

[0090] In step S2022, in the case where the above customer name contains multiple words, a vector aggregation strategy is used to vectorize each word to obtain word vectors. The above vector aggregation strategy is at least one of the following: average pooling strategy, weighted average pooling strategy, and max pooling strategy;

[0091] In step S2023, all the above word vectors are combined in the word order to obtain the above customer name vector.

[0092] In the above embodiment, when processing a customer name composed of multiple words, aggregating the vectors of individual words into a vector representing the entire customer name is a key step. First, use a Chinese word segmentation tool (such as jieba, hanlp, etc.) to segment each customer name, splitting the compound customer name into individual words or phrases. Using a semantic embedding model (such as BCE-Embedding), each segmented word is transformed into a vector representation. Through its internal deep learning structure, the model can take into account the semantics and context information of the word and generate a high-dimensional vector representation of the word. For a customer name composed of multiple words, it is necessary to aggregate the vectors of these words into one vector. There are several strategies to choose from:

[0093] Average pooling strategy: Simply average each word vector to obtain a vector representation of the entire customer name. This method assumes that each word makes an equal semantic contribution to the customer name.

[0094] Weighted average pooling strategy: Weight each word according to its importance in the customer name and then take the average. The importance of a word can be based on its frequency in the corpus, position (for example, words like "Co., Ltd." in a customer name usually may have a lower weight), or determined through additional training data.

[0095] Max pooling strategy: Select the maximum value of each dimension in each word vector as the value of the corresponding dimension in the final customer name vector. This method emphasizes the most representative features in the words.

[0096] By aggregating word vectors into a customer name vector, the system can comprehensively consider the semantic information of all words in the customer name, rather than just individual words. This is crucial for understanding complex names composed of multiple words. The aggregation strategy can improve the accuracy of customer name matching. For example, through weighted averaging, the system can pay more attention to key parts, such as "Agricultural Bank", and reduce the dependence on secondary information, such as "China" (if it is not so crucial in the matching context), so as to still be able to identify the correct customer in ambiguous or incomplete information. There may be some missing, replaced, or abbreviated words in the customer name. Through the vector aggregation strategy, the system can have stronger adaptability to these changes and still be able to accurately match the customer.

[0097] The word vectors obtained are combined according to the word order in the customer name. This process can directly use the result of vector aggregation, or in some complex cases, further processing may be required, such as using models such as recurrent neural networks (RNNs) or Transformers to consider the influence of word order and context. In some customer names, word order may carry important information. For example, some words contain the same words, but the word order is different and the meaning is different. By combining according to word order, this information is retained, which helps to make more accurate judgments in subsequent customer matching.

[0098] Through the above embodiments, the system can convert a complex customer name consisting of multiple words into a unified vector representation that can comprehensively reflect the semantics of the entire name, thereby improving the recognition and matching accuracy of the customer name and providing a solid technical foundation for the network cutover automation notification process.

[0099] In order to improve the search and matching speed and efficiency, in an optional implementation manner, after the above step S202, the method further includes:

[0100] Step S301, storing the above customer name vector into the vector database of the above semantic embedding model.

[0101] In the above embodiment, in the automatic notification of the customer to the network cutover, a corresponding vector is made for each customer name in the affected service list and stored in a vector database. This makes it convenient to use the similarity calculation capability of the semantic embedding model when similar customers need to be found, and the vector database can be used for efficient retrieval. The above vector database can be ChromaDB, Faiss, Milvus, etc. These databases can efficiently process high-dimensional vector data and provide fast similarity search capabilities. The customer name vectors stored in the vector database can support efficient similarity retrieval, which means that when matching CRM standard customer names in the future, the system can quickly find the most similar records without traversing the entire database, greatly improving the matching speed and efficiency.

[0102] Moreover, when storing, some metadata may be attached to each vector, such as customer ID, cutover time and other information, to facilitate the associated use in subsequent retrieval and matching.

[0103] In order to improve the processing speed and efficiency of customer search matching, in an optional implementation, the above step S203 includes:

[0104] Step S2031, in the semantic embedding model, the customer name vector and the standard customer name vector are substituted into the first formula for calculation to obtain the corresponding similarity value. The first formula is a calculation formula for similarity calculation in the semantic embedding model. The first formula is Similarity(A,B) represents the similarity value, A represents the customer name vector, B represents the standard customer name vector, n represents the total number of elements in the customer name vector or the standard customer name vector, and i represents the element number in the customer name vector or the standard customer name vector.

[0105] In the above embodiment, the similarity calculation capability of the BCE-embedding semantic embedding model is called to perform similarity calculation (M:N) with the standardized customer names in the existing CRM system, that is, the customer names (M) extracted by the crawled BCE-embedding semantic embedding model learning and the customer names (N) in the CRM system are calculated by cosine similarity. If the similarity of the matched customer is greater than 80%, the matched customer is considered to be the target customer. Using the cosine similarity formula, the system can quickly calculate the similarity between two vectors, which can significantly improve the processing speed and efficiency for large-scale customer information matching. By calculating the similarity value, the system can identify the CRM customer record that best matches the customer name in the cutover work order, even if the customer name has slight changes or abbreviations in the expression, it can be correctly matched based on semantic similarity. Set a predefined similarity threshold, such as 80% (or 0.8). For matching results below this threshold, the system can exclude or mark them as requiring manual review, thereby ensuring the accuracy of the notification and avoiding mis-sending or omissions.

[0106] In order to reduce the burden of manual operations and improve the efficiency of notification, in an optional implementation, the above step S204 includes:

[0107] Step S2041, querying the contact information corresponding to each of the above target customers;

[0108] Step S2042: automatically calling a service interface to automatically send the cutover notification to the corresponding target customer through the contact information, wherein the service interface is at least one of the following: a text message interface, a voice interface, and a mini-program interface.

[0109] In the above embodiment, the system automatically calls SMS, voice, WeChat mini-program and other service interfaces for the matched customer list together with the cutover information (time period, circuit, etc.) to automatically distribute the cutover notification. The contact information is extracted from the matched customer records, including phone number, email address, WeChat mini-program ID, etc. This information will be used for the subsequent cutover notification distribution. Through integration with the CRM system, the system can quickly locate the customers who need to be notified, avoiding the tedious process of manually finding contact information, and improving the efficiency and accuracy of the notification. It ensures that the contact information used in the notification is the latest and consistent with the information stored in the CRM system, reducing notification failures or delays caused by expired or incorrect contact information. Configure SMS interface, voice interface and / or mini-program interface to ensure that the cutover notification can be sent to customers through these channels. This usually involves docking with the communication service provider to obtain API keys, authentication information and sending rules. Based on information such as the type, start and end time of the network cutover, the content of the cutover notification is automatically generated. The content should include the cutover time, possible impact and recommended countermeasures to ensure that customers can clearly understand the content of the notification and the action suggestions. Based on the extracted customer contact information, the system automatically calls the corresponding service interface, such as the SMS interface to send SMS notifications, the voice interface to make notification calls, or the mini-program interface to push messages to the customer's mini-program end. By automatically calling the service interface, the system can immediately and accurately notify customers of information about network cutover, reducing the burden of manual operations and improving the efficiency of notifications. The use of multiple notification channels such as SMS, voice, and mini-programs ensures coverage of different customer preferences and improves the arrival rate and customer perception of notifications.

[0110] In addition, the status of notification delivery is monitored, and the contact information of failed delivery is recorded, which may require subsequent manual review or retry. The system's monitoring mechanism can ensure the success rate of notification delivery. In the case of failed delivery, the system can automatically try other contact methods or record them for subsequent processing, which improves the reliability of the entire notification process and customer satisfaction.

[0111] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0112] The embodiment of the present application also provides a device for automatically notifying users of network cutover. It should be noted that the device for automatically notifying users of network cutover in the embodiment of the present application can be used to execute the method for automatically notifying users of network cutover provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation in hardware, or a combination of software and hardware, is also possible and conceivable.

[0113] The following is an introduction to the device for automatically notifying users of network cutover provided in an embodiment of the present application.

[0114] Figure 5 1 is a structural block diagram of a device for automatically notifying a user of network cutover according to an embodiment of the present application. Figure 5 As shown, the device comprises:

[0115] The acquisition parsing unit 10 is used to obtain all target network cutover work order attachments, and parse the above target network cutover work order attachments to obtain corresponding parsing field texts, wherein the above target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the above target network cutover operation is a network cutover operation to be executed, and the above parsing field text at least includes the customer name and cutover information affected by the above network cutover operation, and the above cutover information at least includes the network line number, the cutover start time and the cutover end time of the above network cutover operation.

[0116] Specifically, Figure 3 As shown in the figure, a Python web crawler script is written to automatically log in and crawl all work order attachments related to the target network cutover operation in the government and enterprise network operation and maintenance system, including files in PDF, Word, Excel and other formats. You can use corresponding document parsing tools, such as PDFMiner, Python-docx or pandas, or directly use the semantic embedding model to parse the downloaded attachments and extract key field information, including the name of the customer affected by the cutover, the network line number, the start and end time of the cutover, etc. This avoids the tedious process of manually downloading and parsing attachments, and improves the efficiency and accuracy of data acquisition. It ensures that all customer information related to the cutover is completely extracted, providing a comprehensive data foundation for subsequent vectorization and matching.

[0117] The vectorization unit 20 is used to vectorize all the above customer names to obtain customer name vectors corresponding to each of the above customer names, where one customer name corresponds to one customer name vector.

[0118] Specifically, a pre-trained semantic embedding model (such as BCE-Embedding) is called to process each extracted customer name and convert it into a vector representation of a fixed dimension. The generated customer name vector is stored in the vector database in the system for subsequent similarity calculation and retrieval. The unstructured customer name text is converted into a numerical vector for easy machine understanding and processing. Through vectorization, the model can capture the semantic information in the customer name, including lexical meaning and contextual connection, providing a semantic basis for similarity calculation.

[0119] The calculation unit 30 is used to perform similarity calculation on all the above customer name vectors and the standard customer name vector to obtain corresponding similarity values, wherein the similarity values ​​represent the matching degree between the above customer name vectors and the standard customer name vectors.

[0120] Specifically, obtain the canonical customer name vector from the CRM system, and use the similarity search function of the vector database to calculate the similarity value between the customer name vector and the canonical customer name vector, such as through the cosine similarity formula. Set a similarity threshold (such as 80%) as a standard for judging the matching degree of the customer name. By calculating the similarity value, it is possible to find the canonical customer information with the highest matching degree with the customer name affected by the cutover. Using the efficient retrieval capability of the vector database, the matching process is significantly accelerated and the processing time is reduced.

[0121] The notification unit 40 is used to extract the target customers whose similarity values ​​are greater than the set similarity threshold and send a cutover notification to all the target customers. The cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0122] Specifically, from the similarity calculation results, customer name vectors with similarity values ​​higher than the threshold are selected, and the customers corresponding to these vectors are the target customers that need to be notified. Automatically generate a cutover notification text containing cutover information, and automatically call SMS, voice or mini-program interfaces through the system to send a cutover notification to the selected target customers, informing them in advance of the time and possible impact of network cutover. Through automated notification, ensure that customers can prepare for network cutover in advance and reduce the impact of business interruptions. Timely and accurate notifications reduce customer complaints caused by poor information flow and improve customers' service perception.

[0123] In this embodiment, an acquisition parsing unit is used to acquire all target network cutover work order attachments, and parse the target network cutover work order attachments to obtain corresponding parsing field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, the target network cutover operation is a network cutover operation to be executed, and the parsing field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation; a vectorization unit is used to convert all the above customers The names are vectorized to obtain customer name vectors corresponding to the above customer names, and one customer name corresponds to one customer name vector; a calculation unit is used to calculate the similarity of all the above customer name vectors with the standard customer name vector to obtain corresponding similarity values, and the above similarity values ​​represent the matching degree between the above customer name vectors and the above standard customer name vectors; a notification unit is used to extract the target customers corresponding to the similarity values ​​greater than the set similarity threshold, and send a cutover notification to all the above target customers, and the above cutover notification is used to inform the above target customers in advance that the above network cutover operation and the corresponding above cutover information will be performed. This application automatically crawls the attachments of the network cutover work order (i.e., the target network cutover work order attachments), parses the target network cutover work order attachments, extracts the customer name list affected by this network cutover and the cutover information such as the relevant time period, content information, etc.; vectorizes and stores the customer name list; calculates the vector similarity between the obtained customer name list and the customer names already in the customer management system, and the customer whose similarity value is greater than the set similarity threshold (80% or other value) is regarded as the target customer, and the system automatically calls the SMS, voice and other automatic push interface services to push the cutover information to the target customer. This application solves the problem of low efficiency caused by the need for manual review and manual notification sending during network cutover in the prior art.

[0124] In order to improve the automation and efficiency of downloading and processing the attachment information of the cutover work order, in an optional implementation manner, the acquisition and parsing unit includes:

[0125] A download module, used to automatically download the target network cutover work order attachments related to the target network cutover operation through a web crawler script;

[0126] The semantic recognition module is used to call the semantic embedding model to perform semantic recognition on the above-mentioned target network cutover work order attachment, so as to parse the column field content in the above-mentioned target network cutover work order attachment, and extract the above-mentioned parsed field text, wherein the above-mentioned column field content is the above-mentioned network line number, the above-mentioned customer name, the network cutover type, the above-mentioned cutover start time and the above-mentioned cutover end time. The above-mentioned semantic embedding model is a convergence model obtained by iteratively training a large language model using a deep learning algorithm using sample training data. The above-mentioned sample training data at least includes a sample customer name and a sample similarity corresponding to the above-mentioned sample customer name. The above-mentioned sample similarity is the cosine similarity between the vector corresponding to the above-mentioned sample customer name and the above-mentioned standard customer name vector.

[0127] In the above embodiment, if Figure 4 As shown, the attachments in the network cutover work order are crawled in the comprehensive scheduling system through the Python web crawler script to obtain the attachments of the target network cutover work order, in order to prepare for obtaining the data such as the users of the affected business list in the attachments.

[0128] The Python code snippet is as follows:

[0129] import os

[0130] import requests

[0131] from bs4 import BeautifulSoup

[0132] #Target website URL

[0133] url='http: / / 10.152.8.202:30028 / '

[0134] check_requests(url)

[0135] #Traverse the download link and download the file

[0136] for file_url in file_urls:

[0137] file_response=requests.get(file_url,headers=headers,stream=True)

[0138] file_name=os.path.basename(file_url)

[0139] file_path=os.path.join(download_dir,file_name)

[0140] print('Start downloading file')

[0141] with open(file_path,'wb')as f:

[0142] for chunk in file_response.iter_content(chunk_size=1024):

[0143] if chunk:

[0144] f.write(chunk)

[0145] print(f'File {file_name} download completed')

[0146] num++;

[0147] print('All files downloaded')

[0148] def check_requests(url):

[0149] try:

[0150] response=requests.get(url,timeout=10)

[0151] response.raise_for_status()

[0152] print(f"Requests to{url}successful.")

[0153] print('Check network connection')

[0154] return True

[0155] except requests.RequestException as err:

[0156] print(f"Requests to{url}failed:{err}")

[0157] return False

[0158] The downloaded work order attachments may be in PDF, Word, or Excel formats, and the corresponding document parsing libraries (such as pdfplumber, docx, or pandas) need to be used to convert the unstructured data into processable text or table formats. Next, the parsed columns in the affected user list are parsed through the BCE-Embedding semantic embedding model named entity recognition (NER), mainly customer names, and customer name recognition and extraction.

[0159] Column fields: cutover start time, cutover end time, circuit number, customer name, impact duration, cutover type, and code snippet are as follows:

[0160] sys_info=None

[0161] with open('data / zip_tkts.txt','r',encoding='utf-8')as f:

[0162] sys_info = f.read()

[0163] def ask_can_zip(rmk:str="):

[0164] #rmk = 'Cutover start time'

[0165] #rmk = 'Cutover end time'

[0166] #rmk = 'Circuit number'

[0167] #rmk = 'Customer Name'

[0168] #rmk = 'Duration of impact'

[0169] #rmk = 'Cutover type'

[0170]

[0171] By applying named entity recognition (NER) technology, key fields such as network line number, customer name, network cutover type, cutover start time, and cutover end time can be intelligently identified. Through the semantic embedding model, the system can intelligently parse the text in the cutover work order attachment, and accurately identify key fields even if the information is not standardized or ambiguous, thus improving the accuracy and robustness of information extraction.

[0172] Through the above embodiments, the entire system can automatically and efficiently process the attachment information of the cutover work order, improve the accuracy and timeliness of the automatic notification of network cutover to customers, thereby reducing the impact of cutover on the business of government and enterprise customers, and improving service perception and customer satisfaction.

[0173] In order to improve the accuracy and robustness of customer name recognition, in an optional implementation manner, the vectorization unit includes:

[0174] The first vectorization module is used to use a semantic embedding model to vector encode all the above-mentioned customer names and generate the above-mentioned customer name vectors corresponding to each of the above-mentioned customer names. The above-mentioned semantic embedding model is a convergence model obtained by iteratively training a large language model using a deep learning algorithm using sample training data. The above-mentioned sample training data at least includes sample customer names and sample similarities corresponding to the above-mentioned sample customer names. The above-mentioned customer name vector is an intermediate variable generated by the above-mentioned semantic embedding model.

[0175] In the above embodiment, the collected field information (main field customer name) data is converted into a form suitable for storage in a vector database. For each customer name extracted from the target network cutover work order attachment, a semantic embedding model is used for vector encoding, and each customer name is converted into its corresponding vector representation. The model outputs a vector of fixed length, which carries the semantic features of the customer name. The generated customer name vector is an intermediate variable, which does not directly represent the text information of the customer name, but captures the semantic features of the customer name in the form of a numerical vector. These vectors can be stored in a vector database to facilitate subsequent similarity calculations and customer matching. By converting customer names into vectors, the system can identify and match customer names based on semantics, rather than just literal text matching. This helps to handle abbreviations, aliases, and spelling differences in customer names, and improves the accuracy and robustness of customer name recognition.

[0176] In order to improve the recognition and matching accuracy of complex customer names, in an optional implementation manner, the vectorization unit further includes:

[0177] A second vectorization module is used for, when the customer name contains multiple words, vectorizing each word by adopting a vector aggregation strategy to obtain a word vector, wherein the vector aggregation strategy is at least one of the following: an average pooling strategy, a weighted average pooling strategy, and a maximum pooling strategy;

[0178] The vector combination module is used to combine all the above word vectors according to the word order to obtain the above customer name vector.

[0179] In the above embodiments, when processing a customer name consisting of multiple words, aggregating the vectors of individual words into a vector representing the entire customer name is a crucial step. First, use a Chinese word segmentation tool (such as jieba, hanlp, etc.) to segment each customer name, splitting the compound customer name into individual words or phrases. Utilize a semantic embedding model (such as BCE-Embedding) to convert each segmented word into a vector representation. Through its internal deep learning structure, the model can take into account the semantics and context information of the word and generate a high-dimensional vector representation of the word. For a customer name consisting of multiple words, it is necessary to aggregate the vectors of these words into one vector. There are several strategies to choose from:

[0180] Average pooling strategy: Simply average each word vector to obtain a vector representation of the entire customer name. This method assumes that each word contributes equally to the semantics of the customer name.

[0181] Weighted average pooling strategy: Weight each word according to its importance in the customer name and then take the average. The importance of a word can be based on its frequency in the corpus, position (for example, words indicating the company type in a customer name such as "Limited Company" may have a lower weight), or determined through additional training data.

[0182] Max pooling strategy: Select the maximum value of each dimension in each word vector as the value of the corresponding dimension in the final customer name vector. This method emphasizes the most representative features in the words.

[0183] By aggregating word vectors into a customer name vector, the system can comprehensively consider the semantic information of all words in the customer name, rather than just individual words. This is crucial for understanding complex names consisting of multiple words. The aggregation strategy can improve the accuracy of customer name matching. For example, through weighted average, the system can pay more attention to key parts, such as "Agricultural Bank", and reduce the dependence on secondary information, such as "China" (if it is not so crucial in the matching context), so as to still be able to identify the correct customer in ambiguous or incomplete information. There may be some missing or replaced words or abbreviations in the customer name. Through the vector aggregation strategy, the system can have a stronger adaptability to these changes and still be able to accurately match the customer.

[0184] The word vectors obtained are combined according to the word order in the customer name. This process can directly use the result of vector aggregation, or in some complex cases, further processing may be required, such as using models such as recurrent neural networks (RNNs) or Transformers to consider the influence of word order and context. In some customer names, word order may carry important information. For example, some words contain the same words, but the word order is different and the meaning is different. By combining according to word order, this information is retained, which helps to make more accurate judgments in subsequent customer matching.

[0185] Through the above embodiments, the system can convert a complex customer name consisting of multiple words into a unified vector representation that can comprehensively reflect the semantics of the entire name, thereby improving the recognition and matching accuracy of the customer name and providing a solid technical foundation for the network cutover automation notification process.

[0186] In order to improve the retrieval matching speed and efficiency, in an optional implementation manner, the device further includes:

[0187] The storage unit is used to vectorize all the above customer names to obtain corresponding customer name vectors, and then store the above customer name vectors in the vector database of the above semantic embedding model.

[0188] In the above embodiment, in the automatic notification of the customer to the network cutover, a corresponding vector is made for each customer name in the affected service list and stored in a vector database. This makes it convenient to use the similarity calculation capability of the semantic embedding model when similar customers need to be found, and the vector database can be used for efficient retrieval. The above vector database can be ChromaDB, Faiss, Milvus, etc. These databases can efficiently process high-dimensional vector data and provide fast similarity search capabilities. The customer name vectors stored in the vector database can support efficient similarity retrieval, which means that when matching CRM standard customer names in the future, the system can quickly find the most similar records without traversing the entire database, greatly improving the matching speed and efficiency.

[0189] Moreover, when storing, some metadata may be attached to each vector, such as customer ID, cutover time and other information, to facilitate the associated use in subsequent retrieval and matching.

[0190] In order to improve the processing speed and efficiency of customer search matching, in an optional implementation, the above-mentioned calculation unit includes:

[0191] A calculation module is used to substitute the customer name vector and the standard customer name vector into the first formula in the semantic embedding model to calculate and obtain the corresponding similarity value. The first formula is the calculation formula for similarity calculation in the semantic embedding model. The first formula is Similarity(A,B) represents the similarity value, A represents the customer name vector, B represents the standard customer name vector, n represents the total number of elements in the customer name vector or the standard customer name vector, and i represents the element number in the customer name vector or the standard customer name vector.

[0192] In the above embodiment, the similarity calculation capability of the BCE-embedding semantic embedding model is called to perform similarity calculation (M:N) with the standardized customer names in the existing CRM system, that is, the customer names (M) extracted by the crawled BCE-embedding semantic embedding model learning and the customer names (N) in the CRM system are calculated by cosine similarity. If the similarity of the matched customer is greater than 80%, the matched customer is considered to be the target customer. Using the cosine similarity formula, the system can quickly calculate the similarity between two vectors, which can significantly improve the processing speed and efficiency for large-scale customer information matching. By calculating the similarity value, the system can identify the CRM customer record that best matches the customer name in the cutover work order, even if the customer name has slight changes or abbreviations in the expression, it can be correctly matched based on semantic similarity. Set a predefined similarity threshold, such as 80% (or 0.8). For matching results below this threshold, the system can exclude or mark them as requiring manual review, thereby ensuring the accuracy of the notification and avoiding mis-sending or omissions.

[0193] In order to reduce the burden of manual operations and improve the efficiency of notification, in an optional implementation, the notification unit includes:

[0194] A query module, used to query the contact information corresponding to each of the above target customers;

[0195] The notification module is used to automatically call the service interface to automatically distribute the above-mentioned cutover notification to the corresponding above-mentioned target customer through the above-mentioned contact method, and the above-mentioned service interface is at least one of the following: SMS interface, voice interface and mini-program interface.

[0196] In the above embodiment, the system automatically calls SMS, voice, WeChat mini-program and other service interfaces for the matched customer list together with the cutover information (time period, circuit, etc.) to automatically distribute the cutover notification. The contact information is extracted from the matched customer records, including phone number, email address, WeChat mini-program ID, etc. This information will be used for the subsequent cutover notification distribution. Through integration with the CRM system, the system can quickly locate the customers who need to be notified, avoiding the tedious process of manually finding contact information, and improving the efficiency and accuracy of the notification. It ensures that the contact information used in the notification is the latest and consistent with the information stored in the CRM system, reducing notification failures or delays caused by expired or incorrect contact information. Configure SMS interface, voice interface and / or mini-program interface to ensure that the cutover notification can be sent to customers through these channels. This usually involves docking with the communication service provider to obtain API keys, authentication information and sending rules. Based on information such as the type, start and end time of the network cutover, the content of the cutover notification is automatically generated. The content should include the cutover time, possible impact and recommended countermeasures to ensure that customers can clearly understand the content of the notification and the action suggestions. Based on the extracted customer contact information, the system automatically calls the corresponding service interface, such as the SMS interface to send SMS notifications, the voice interface to make notification calls, or the mini-program interface to push messages to the customer's mini-program end. By automatically calling the service interface, the system can immediately and accurately notify customers of information about network cutover, reducing the burden of manual operations and improving the efficiency of notifications. The use of multiple notification channels such as SMS, voice, and mini-programs ensures coverage of different customer preferences and improves the arrival rate and customer perception of notifications.

[0197] In addition, the status of notification delivery is monitored, and the contact information of failed delivery is recorded, which may require subsequent manual review or retry. The system's monitoring mechanism can ensure the success rate of notification delivery. In the case of failed delivery, the system can automatically try other contact methods or record them for subsequent processing, which improves the reliability of the entire notification process and customer satisfaction.

[0198] The above-mentioned device for automatically notifying users of network cutover includes a processor and a memory. The above-mentioned acquisition and parsing unit, vectorization unit and calculation unit are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in any combination.

[0199] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the problem of low efficiency caused by manual review and manual notification during network cutover in the prior art can be solved by adjusting kernel parameters.

[0200] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0201] An embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for automatically notifying users of network cutover.

[0202] Specifically, the method for automatically notifying users of network cutover includes:

[0203] Step S201, obtaining all target network cutover work order attachments, and parsing the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is a network cutover operation to be performed, and the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation;

[0204] Step S202, vectorizing all the customer names to obtain customer name vectors corresponding to each customer name, where one customer name corresponds to one customer name vector;

[0205] Step S203, performing similarity calculation on all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector;

[0206] Step S204, extracting the target customers whose similarity values ​​are greater than the set similarity threshold, and sending a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0207] An embodiment of the present invention provides a processor, and the processor is used to run a program, wherein the method for automatically notifying a user of network cutover is executed when the program is run.

[0208] An embodiment of the present invention provides a network cutover automation notification system, the network cutover automation notification system includes a processor, a memory, and a program stored in the memory and executable on the processor, and when the processor executes the program, at least the following steps are implemented:

[0209] Step S201, obtaining all target network cutover work order attachments, and parsing the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is a network cutover operation to be performed, and the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation;

[0210] Step S202, vectorizing all the customer names to obtain customer name vectors corresponding to each customer name, where one customer name corresponds to one customer name vector;

[0211] Step S203, performing similarity calculation on all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector;

[0212] Step S204, extracting the target customers whose similarity values ​​are greater than the set similarity threshold, and sending a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0213] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program for initializing at least the following method steps:

[0214] Step S201, obtaining all target network cutover work order attachments, and parsing the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is a network cutover operation to be performed, and the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation;

[0215] Step S202, vectorizing all the customer names to obtain customer name vectors corresponding to each customer name, where one customer name corresponds to one customer name vector;

[0216] Step S203, performing similarity calculation on all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector;

[0217] Step S204, extracting the target customers whose similarity values ​​are greater than the set similarity threshold, and sending a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

[0218] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0219] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0220] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0221] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0222] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0223] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0224] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0225] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0226] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0227] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0228] 1) The method of automatically notifying users of network cutover of the present application comprises: first, obtaining all target network cutover work order attachments, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is the network cutover operation to be executed; then, parsing the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation; then, All the above customer names are vectorized to obtain customer name vectors corresponding to each of the above customer names, and one customer name corresponds to one customer name vector; then, all the above customer name vectors are similarity calculated with the standard customer name vector to obtain the corresponding similarity value, and the above similarity value represents the degree of match between the above customer name vector and the above standard customer name vector; finally, the target customers corresponding to the above similarity value greater than the set similarity threshold are extracted, and a cutover notice is sent to all the above target customers, and the above cutover notice is used to inform the above target customers in advance that the above network cutover operation and the corresponding above cutover information will be performed. This application automatically crawls the attachments of the network cutover work order (i.e., the target network cutover work order attachments), parses the target network cutover work order attachments, extracts the customer name list affected by this network cutover and the relevant time period, content information and other cutover information; vectorizes and stores the customer name list; calculates the vector similarity between the obtained customer name list and the customer names already in the customer management system, and the customers whose similarity values ​​are greater than the set similarity threshold (80% or other values) are used as target customers, and the system automatically calls SMS, voice and other automatic push interface services to push the cutover information to the target customers. The present application solves the problem in the prior art that manual review and manual notification are required during network cutover, resulting in low efficiency.

[0229] 2) The apparatus for automatically notifying users of network cutover of the present application comprises: an acquisition parsing unit for acquiring all target network cutover work order attachments, and parsing the target network cutover work order attachments to obtain corresponding parsing field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, and the target network cutover operation is a network cutover operation to be performed, and the parsing field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time and the cutover end time of the network cutover operation; and a vectorization unit for All the above-mentioned customer names are vectorized to obtain customer name vectors corresponding to each of the above-mentioned customer names, and one customer name corresponds to one customer name vector; a calculation unit is used to calculate the similarity of all the above-mentioned customer name vectors with the standard customer name vector to obtain corresponding similarity values, and the above-mentioned similarity values ​​represent the matching degree between the above-mentioned customer name vectors and the above-mentioned standard customer name vectors; a notification unit is used to extract the target customers corresponding to the above-mentioned similarity values ​​greater than the set similarity threshold, and send a cutover notification to all the above-mentioned target customers, and the above-mentioned cutover notification is used to inform the above-mentioned target customers in advance that the above-mentioned network cutover operation and the corresponding above-mentioned cutover information will be performed. This application automatically crawls the attachments of the network cutover work order (i.e., the target network cutover work order attachments), parses the target network cutover work order attachments, extracts the customer name list affected by this network cutover and the cutover information such as the relevant time period, content information, etc.; vectorizes and stores the customer name list; calculates the vector similarity between the obtained customer name list and the customer names already in the customer management system, and the customer whose similarity value is greater than the set similarity threshold (80% or other value) is regarded as the target customer, and the system automatically calls the SMS, voice and other automatic push interface services to push the cutover information to the target customer. This application solves the problem of low efficiency caused by the need for manual review and manual notification sending during network cutover in the prior art.

[0230] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for automatically notifying users of network cutover, characterized in that: include: Acquire all target network cutover work order attachments, and parse the target network cutover work order attachments to obtain corresponding parsed field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, the target network cutover operation is a network cutover operation to be performed, and the parsed field texts at least include the customer name and cutover information affected by the network cutover operation, and the cutover information at least includes the network line number, the cutover start time, and the cutover end time of the network cutover operation; Vectorize all the customer names to obtain customer name vectors corresponding to each customer name, where one customer name corresponds to one customer name vector; Calculate the similarity between all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector; Target customers whose similarity values ​​are greater than a set similarity threshold are extracted, and a cutover notification is sent to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

2. The method according to claim 1, characterized in that Obtain all target network cutover work order attachments and parse the target network cutover work order attachments to obtain parsed field texts, including: Automatically downloading the target network cutover work order attachment related to the target network cutover operation through a web crawler script; A semantic embedding model is called to perform semantic recognition on the target network cutover work order attachment to parse the column field content in the target network cutover work order attachment to extract the parsed field text, wherein the column field content is the network line number, the customer name, the network cutover type, the cutover start time and the cutover end time. The semantic embedding model is a convergence model obtained by iteratively training a large language model using a deep learning algorithm using sample training data. The sample training data includes at least a sample customer name and a sample similarity corresponding to the sample customer name. The sample similarity is the cosine similarity between the vector corresponding to the sample customer name and the standard customer name vector.

3. The method according to claim 1, characterized in that All the customer names are vectorized to obtain customer name vectors corresponding to each customer name, including: A semantic embedding model is used to perform vector encoding on all the customer names to generate the customer name vector corresponding to each customer name. The semantic embedding model is a convergent model obtained by iteratively training a large language model using a deep learning algorithm using sample training data. The sample training data at least includes sample customer names and sample similarities corresponding to the sample customer names. The customer name vector is an intermediate variable generated by the semantic embedding model.

4. The method according to claim 3, characterized in that All the customer names are vectorized to obtain corresponding customer name vectors, and also include: In the case where the customer name contains multiple words, each word is vectorized by using a vector aggregation strategy to obtain a word vector, and the vector aggregation strategy is at least one of the following: an average pooling strategy, a weighted average pooling strategy, and a maximum pooling strategy; All the word vectors are combined according to the word order to obtain the customer name vector.

5. The method according to claim 3, characterized in that: After all the customer names are vectorized to obtain corresponding customer name vectors, the method further includes: The customer name vector is stored in the vector database of the semantic embedding model.

6. The method according to claim 3, characterized in that All the customer name vectors are similarly calculated with the standard customer name vector to obtain corresponding similarity values, including: In the semantic embedding model, the customer name vector and the standard customer name vector are substituted into the first formula for calculation to obtain the corresponding similarity value, wherein the first formula is a calculation formula for similarity calculation in the semantic embedding model, and the first formula is Similarity(A,B) represents the similarity value, A represents the customer name vector, B represents the standard customer name vector, n represents the total number of elements in the customer name vector or the standard customer name vector, and i represents the element number in the customer name vector or the standard customer name vector.

7. The method according to claim 1, characterized in that Send a cutover notice to all target customers, including: Query the contact information corresponding to each of the target customers; The service interface is automatically called to automatically send the cutover notification to the corresponding target customer through the contact method, and the service interface is at least one of the following: a text message interface, a voice interface, and a mini-program interface.

8. A device for automatically notifying users of network cutover, characterized in that: The device comprises: an acquisition and parsing unit, configured to acquire all target network cutover work order attachments, and parse the target network cutover work order attachments to obtain corresponding parsing field texts, wherein the target network cutover work order attachments are text attachments associated with the target network cutover operation in the network cutover work order stored in the system, the target network cutover operation is a network cutover operation to be performed, the parsing field texts at least include a customer name and cutover information affected by the network cutover operation, and the cutover information at least includes a network line number, a cutover start time, and a cutover end time of the network cutover operation; A vectorization unit, used for vectorizing all the customer names to obtain a customer name vector corresponding to each customer name, where one customer name corresponds to one customer name vector; A calculation unit, used for calculating the similarity between all the customer name vectors and the standard customer name vector to obtain a corresponding similarity value, wherein the similarity value represents the degree of matching between the customer name vector and the standard customer name vector; The notification unit is used to extract the target customers whose similarity values ​​are greater than the set similarity threshold and send a cutover notification to all the target customers, wherein the cutover notification is used to inform the target customers in advance that the network cutover operation and the corresponding cutover information will be performed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images