Multi-format purchase order analysis and input method based on artificial intelligence

Abstract

The invention belongs to the technical field of order data processing, and particularly relates to a multi-format purchase order analysis and input method based on artificial intelligence, and the method comprises the steps: inputting a customer order file; according to the customer order file, the adapter automatically identifies the format of the customer order file, and finds a corresponding parser for parsing; according to the text content of the customer order file, using a proper analyzer to extract the text content; the multi-format purchase order input method comprises the following steps: selecting a proper AI model, adapting to a proper Prompt, integrating the text content into a dialogue through the Prompt, and automatically identifying a file format and calling a corresponding analyzer to accurately extract the text content in files of various formats through the arrangement of an adapter. And the text format is converted into a uniform text format, so that manual identification and input are not needed, and the automation degree and efficiency of order processing are greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of order data processing, and in particular relates to a multi-format purchase order parsing and entry method based on artificial intelligence. Background Art

[0002] In the prior art, the methods for parsing and entering purchase orders are nothing more than manual entry, OCR recognition and PDF parsing. The disadvantages of these methods are: 1. Manual data entry: Visual recognition and manual data entry lead to low efficiency, prone to errors, extremely high labor costs, and huge workload.

[0003] 2. OCR recognition and PDF parsing: The parsed content still needs to be manually found to correspond to the system standard template and converted through a series of Excel matching and other operations, which is prone to errors and requires a lot of work.

[0004] In addition, because the formats of purchase orders are different, different templates are required to match them. Exclusive templates must be created for each customer, which requires a large amount of research and development work and is difficult to cover all customers. When a new customer joins or the format of a customer's order changes, research and development time is required to maintain the template separately. The development and maintenance costs are high, the degree of automation is insufficient, and the response speed is slow, making it impossible to meet business needs in a timely manner. There are still a large number of customers' orders that cannot be automatically entered through templates and need to be processed manually, which is inefficient and takes up a lot of manpower.

[0005] Based on this, the applicant proposed an artificial intelligence-based multi-format purchase order parsing and entry method to solve the above problems. Summary of the invention

[0006] Based on this, it is necessary to provide an artificial intelligence-based multi-format purchase order parsing and entry method and its use method that can parse multi-format orders and enter order contents.

[0007] A multi-format purchase order parsing and entry method based on artificial intelligence, the multi-format purchase order parsing and entry method comprising: Enter customer order documents; Based on the customer order file, the adapter automatically identifies its format and finds the corresponding parser for parsing; According to the text content of the customer order file, use a suitable parser to extract the text content; Select a suitable AI model and adapt a suitable prompt. Use the prompt to integrate the above text content into a dialogue, which is input into the AI ​​model for its understanding. The AI ​​model uses the prompt to guide the recognition of information in the order. According to the identified order information, convert it into formatted data of the inherent template; According to the obtained formatted data, the data for feature completion is completed; Based on the completed formatted data, the data is displayed on the user interface, and the user can verify the above data by themselves to ensure whether the order information is compliant; The order information can be exported after the customer confirms that it is correct.

[0008] Furthermore, the parser includes Text Extractor, PDF BOX, POI and OCR.

[0009] Furthermore, after receiving the parsed text content, Prompt analyzes the text content, extracts key information, including product name, quantity, and price, and specifies the output format and sample; Prompt generation module, generates appropriate prompt guide words according to the key information obtained above; The generated Prompt is input into the pre-trained language model. The language model generates the final dialogue text based on the information in the Prompt and its own learned language knowledge and dialogue logic for the AI ​​model to understand. Furthermore, based on the conversation text, the AI ​​model can identify keywords in the conversation text and specify the output generation format, and the formatted data is in a tabular form and / or other data statistical form.

[0010] Furthermore, the formatted data output by the AI ​​model is identified according to a fixed template, and the missing features of the formatted data are found according to the fixed template, and the features are supplemented to the formatted data before being output to the user interface.

[0011] Technical effects and advantages of the present invention: 1. The present invention is provided with an adapter, which can automatically identify the file format and call the corresponding parser to accurately extract the text content in various format files and convert it into a unified text format without manual identification and entry, thereby greatly improving the automation and efficiency of order processing.

[0012] 2. The AI ​​model used in the present invention can convert the extracted text content into formatted data, eliminating the need to create separate templates for each customer and flexibly adapting to the order formats of different customers.

[0013] 3. The present invention provides a user verification interface, through which users can view the order data converted by the AI ​​model, and modify and confirm it to improve accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1The present invention is a schematic diagram of the steps of a method for parsing and entering multi-format purchase orders based on artificial intelligence according to an embodiment of the present invention.

[0015] Figure 2 Schematic diagram of the steps for Prompt to convert the acquired text content into dialogue text.

[0016] Figure 3 This is a schematic diagram of the imported original file.

[0017] Figure 4 This is a schematic diagram of the display and verification interface for formatted data on the user interface.

[0018] Figure 5 Schematic diagram for exporting order data into a common file. DETAILED DESCRIPTION

[0019] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. The preferred embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thoroughly understood.

[0020] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only and do not represent the only implementation method.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0022] Example like Figure 1-5 The present invention relates to a multi-format purchase order parsing and input method based on artificial intelligence, comprising: Enter as Figure 3 Customer order documents in ; According to the customer order file, the adapter automatically identifies its format and finds the corresponding parser for parsing; through the file adapter, intelligent file format recognition is adopted, such as program recognition method, that is, txt corresponds to .txt corresponds to TextExtractor, .pdf corresponds to PDF BOX, .xlsx corresponds to POI, .png, .jpg corresponds to OCR, which can automatically detect the format type of the uploaded file and automatically call the corresponding parser according to the recognition result. Various parsers include Text Extractor, PDF BOX, POI and OCR. After optimization and encapsulation, the parser can extract text content from files of various formats and convert it into a unified text format such as txt.

[0023] According to the text content of the customer order file, use a suitable parser to extract the text content; take PDFBOX's recognition of PDF files as an example to describe its recognition and conversion methods in detail. Load the uploaded PDF document and use the PDDcocument.load method to load the PDF file; Extract text content, create a PDF Text Stripper object, and call its get Text method to extract the file; Format the text content, remove extra blank characters, and use the repalce method of the String class to remove extra spaces, tabs, and line breaks in the text; Further process and convert the format as required, converting the extracted text content into other data structures, such as lists, for subsequent transmission; The above method only takes PDF BOX's recognition of PDF files as an example. Scanned PDF files can also be recognized and output in combination with OCR. In addition, Text Extractor and POI can recognize Text, PDF, Word, EXCEL, PPT and various image files. Their recognition methods are all existing technologies, so they are not described in detail in this document. Select a suitable AI model and adapt a suitable prompt. Use the prompt to integrate the above text content into a dialogue, which is input into the AI ​​model for its understanding. The AI ​​model uses the prompt to guide the recognition of information in the order. After receiving the parsed text content, Prompt analyzes the text content, extracts key information, including product name, quantity, and price, and specifies the output format and sample; Prompt generation module, generates appropriate prompt guide words according to the key information obtained above; The generated prompt is input into the pre-trained language model. The language model generates the final dialogue text based on the information in the prompt and its own learned language knowledge and dialogue logic for the AI ​​model to understand. According to the identified order information, convert it into formatted data of the inherent template; Prompt defines how to extract key information from text, which is then converted into dialogue text. For example, the key information involved in the text content is that the customer needs 100 1L rice cookers in the order content, with a total price of 20,000. The text content and the configured parsing Prompt are input into the AI ​​model together. According to the guide words of Prompt, the text content can be converted into dialogue text. like Text: 100 1L rice cookers, total price 20,000.

[0024] System: Extract product name, quantity, unit price, and brand from text information, and the price needs to be converted into unit price.

[0025] The list form that the AI ​​model can recognize and convert is Product Name quantity unit price brand Rice Cooker 100 200 Based on the conversation text, the AI ​​model can identify the keywords in the conversation text and specify the output generation format, and the formatted data is in table form and / or other statistical form; The fixed templates converted in a unified manner are not only easy for customers to check and correct, but also do not require different templates to be formulated according to customer needs. In addition, there is no need to maintain templates in a timely manner due to the addition of new customers or changes in the order format of old customers, otherwise matching failures or extraction errors will occur. According to the obtained formatted data, the data for feature completion is completed; Take the fixed template above as an example Product Name quantity unit price brand Xxxx Xxxx Xxxx XXX The output data becomes Product Name quantity unit price brand Rice Cooker 100 200 Supor Data can be completed based on the supplier's database. For example, if the product is identified as a rice cooker, its brand name can be completed based on the database. If the missing feature that can be identified in the above is the brand, and the brand name is Supor, it can be completed and output; like Figure 4As shown, according to the completed formatted data, the data is displayed on the user interface, and the user can verify the above data by himself to ensure whether the order information is compliant; the formatted data output by the AI ​​model is identified according to the fixed template, and the missing features after the formatted data are found according to the fixed template, and the features are filled in the formatted data and then output to the user interface. The transmission method can use HTML, CSS or JavaScript technology to build a page to present it to the customer. The above methods are all relatively common display page construction methods in the prior art, so they are not repeated in this article; like Figure 5 As shown, the order information can be exported after the customer confirms that it is correct.

[0026] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0027] The above-mentioned embodiments only express several embodiments of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached claims.

Claims

1. A multi-format purchase order parsing and entry method based on artificial intelligence, characterized in that: The multi-format purchase order parsing and input method includes: Enter customer order documents; Based on the customer order file, the adapter automatically identifies its format and finds the corresponding parser for parsing; According to the text content of the customer order file, use a suitable parser to extract the text content; Select a suitable AI model and adapt a suitable prompt. Use the prompt to integrate the above text content into a dialogue, which is input into the AI ​​model for its understanding. The AI ​​model uses the prompt to guide the recognition of information in the order. According to the identified order information, convert it into formatted data of the inherent template; According to the obtained formatted data, the data for feature completion is completed; Based on the completed formatted data, the data is displayed on the user interface, and the user can verify the above data by themselves to ensure whether the order information is compliant; The order information can be exported after the customer confirms that it is correct.

2. The method for parsing and entering multi-format purchase orders based on artificial intelligence according to claim 1 is characterized in that: The parser includes Text Extractor, PDF BOX, POI and OCR.

3. The method for parsing and entering multi-format purchase orders based on artificial intelligence according to claim 1 is characterized in that: After receiving the parsed text content, Prompt analyzes the text content, extracts key information, including product name, quantity, and price, and specifies the output format and sample; Prompt generation module, generates appropriate prompt guide words according to the key information obtained above; The generated Prompt is input into the pre-trained language model. The language model generates the final dialogue text based on the information in the Prompt and its own learned language knowledge and dialogue logic for the AI ​​model to understand.

4. The method for parsing and entering multi-format purchase orders based on artificial intelligence according to claim 3 is characterized in that: Based on the conversation text, the AI ​​model can identify keywords in the conversation text and specify the output generation format, and the formatted data is in table form and / or other data statistical form.

5. The method for parsing and entering multi-format purchase orders based on artificial intelligence according to claim 1 is characterized in that: The formatted data output by the AI ​​model is identified according to a fixed template, and the missing features of the formatted data are found according to the fixed template. The features are then added to the formatted data before being output to the user interface.

Images