Method for constructing whole-chain meat safety knowledge graph and retrieval device

Abstract

The application provides a full-chain meat safety knowledge graph construction method and retrieval device, which has the following characteristics: steps S1-S2, entity and relationship extraction is performed on all meat safety standard files; step S3, meat product nodes and corresponding subordinate nodes are constructed and relationships are set; steps S4-S6, corresponding multistage nodes are set for each subordinate node; step S7, the extracted entities and relationships are assigned to the corresponding nodes; step S8, CCP selection nodes including specific hazards and corresponding control measures are set according to the HACCP principle; steps S9-S10, non-directly connected nodes are directly connected as needed through artificial relationships. In summary, the method can construct a more comprehensive meat safety knowledge graph, and the retrieval device can more quickly, conveniently and accurately retrieve and display the data of the meat safety knowledge graph.

Description

Technical Field

[0001] This invention belongs to the field of food safety information big data technology, specifically involving a method for constructing a full-chain meat safety knowledge graph and a retrieval device. Background Technology

[0002] Meat products are an important component of my country's food structure, and their safety is closely related to consumer health. The meat product industry chain involves multiple stages, including breeding, slaughtering, processing, storage, transportation, and sales. Any disruption or lack of oversight at any stage can impact meat product safety. Hazard Analysis and Critical Control Point (HACCP) is a risk-based food safety management system designed to ensure hygiene and safety. HACCP helps food manufacturers identify hazards at each stage of food production and processing and implement effective control measures, ensuring product safety and reducing food safety incidents. On the other hand, food safety standards are the only mandatory food standards. They represent the minimum requirements that food producers and operators must follow, the legal permit for food production and consumption, and the basis for government regulatory enforcement. Linking HACCP information throughout the entire meat product processing chain with food safety standards, establishing a connection between food safety standards and relevant meat product production stages and key control measures, is of great significance for ensuring meat product safety. However, there are numerous food safety standard documents covering the entire meat product chain, and these documents are lengthy and contain a lot of repetitive content. Manual loading and querying methods cannot obtain the correlation between standards, making it difficult to directly obtain relevant information about related standards. Knowledge service methods are relatively limited, time-consuming, and inefficient.

[0003] A knowledge graph is a semantic network knowledge base that describes concepts, entities, events, and relationships in the objective world, providing a channel for users to quickly obtain relevant knowledge and their logical relationships.

[0004] However, existing knowledge graphs related to meat products suffer from disorganized content integration and a lack of HACCP information, resulting in low query efficiency and accuracy during use. Summary of the Invention

[0005] This invention is made to solve the above-mentioned problems, and aims to provide a method for constructing a knowledge graph of meat safety across the entire supply chain and a retrieval device.

[0006] This invention provides a method for constructing a full-chain meat safety knowledge graph based on HACCP principles. This method constructs a full-chain meat safety knowledge graph based on multiple meat safety standard documents and includes the following steps: Step S1, extracting entities from each specific standard in all meat safety standard documents to obtain multiple standard testing objects and corresponding standard requirements; Step S2, extracting relationships from all meat safety standard documents to obtain multiple existing relationships; Step S3, constructing a meat product node, and then constructing a meat product contaminant limit node, a meat product pathogenic bacteria limit node, and various meat product sub-category nodes as lower-level nodes of the meat product node, and using the corresponding existing relationships as nodes. Step S4: For each meat product sub-category node, construct corresponding raw meat nodes, safety standard nodes, and CCP selection nodes as lower-level nodes of the meat product sub-category node, and use the corresponding existing relationships as the relationships between nodes; Step S5: For each raw meat node, construct corresponding raw meat contaminant limit nodes, raw meat pathogenic bacteria limit nodes, and raw meat pesticide and veterinary drug residue nodes as lower-level nodes of the raw meat node, and use the corresponding existing relationships as the relationships between nodes; Step S6: For each safety standard node, construct corresponding sensory requirement nodes, physicochemical requirement nodes, and microbiological requirement nodes as lower-level nodes of the safety standard node, and use the corresponding existing relationships as the relationships between nodes. Relationship; Step S7, take the corresponding standard test objects as the lower-level nodes of meat product contaminant limit nodes, meat product pathogenic bacteria limit nodes, raw meat contaminant limit nodes, raw meat pathogenic bacteria limit nodes, raw meat pesticide and veterinary drug residue nodes, sensory requirements nodes, physicochemical requirements nodes, and microbiological requirements nodes, and then take the standard requirements as the lower-level nodes of the corresponding standard test objects, and take the corresponding existing relationships as the relationships between nodes; Step S8, for each CCP selection node, according to the HACCP principle, perform hazard analysis on the raw meat and processing process of the meat product subclass corresponding to the CCP selection node, and obtain multiple critical control points as the lower-level nodes of the CCP selection node. The key control points include specific hazards and corresponding control measures; Step S9, for each raw meat node and meat product node, directly connect the node to the corresponding related pathogenic bacteria and contaminant nodes through the first artificial relationship; Step S10, for each raw meat node and meat product node, directly connect the node to the corresponding related pathogenic bacteria and contaminant standard requirements nodes through the second artificial relationship. All nodes and relationships between nodes constructed through steps S1 to S10 are used as a whole-chain meat safety knowledge graph, where each node has primary and secondary attributes. The primary attribute is the specific content of the node, and the secondary attribute is the specific content of the node's parent node.

[0007] The method for constructing a full-chain meat safety knowledge graph based on HACCP principles provided by this invention may also have the following features: the meat safety standard document is a national meat safety standard document, which includes food safety standards, limits for contaminants, limits for pathogens, limits for pesticide and veterinary drug residues, and some local safety standards.

[0008] The method for constructing a full-chain meat safety knowledge graph based on HACCP principles provided by this invention may also have the following features: In step S1, the specific standards include the specific content of sensory requirements, technical requirements, microbial limits, pathogenic bacteria limits, and contaminant limits that meat products should meet.

[0009] The method for constructing a full-chain meat safety knowledge graph based on HACCP principles provided by this invention may also have the following feature: in step S3, the classification standard of the meat product subclass corresponding to the meat product subclass node is based on the food production license classification catalog.

[0010] The present invention also provides a whole-chain meat safety knowledge graph retrieval device, characterized by comprising: a user input unit for a user to input a retrieval statement containing specified relationships and specified nodes; a knowledge graph storage unit containing a database storing data of the whole-chain meat safety knowledge graph; a retrieval unit for retrieving nodes corresponding to specified relationships and specified nodes from the whole-chain meat safety knowledge graph according to the retrieval statement as retrieval result nodes; and a display unit for displaying the content of the specified nodes and retrieval result nodes, as well as the specified relationships between the specified nodes and retrieval result nodes, wherein the whole-chain meat safety knowledge graph is constructed according to steps S1 to S10 of the whole-chain meat safety knowledge graph construction method based on HACCP principles described above.

[0011] The whole-chain meat safety knowledge graph retrieval device provided by the present invention may also have the following feature: the content of the retrieval result node includes the main attributes and secondary attributes of the node.

[0012] The full-chain meat safety knowledge graph retrieval device provided by this invention may also have the following features: the database is a Neo4j database, and the data storage method of the full-chain meat safety knowledge graph is as follows: each node in the full-chain meat safety knowledge graph is numbered to obtain an entity number, and then the node is converted into a form represented by the entity number, main attributes, and secondary attributes and saved as different files according to different meat safety standard documents. The relationship between two nodes is saved as different files according to different relationships, represented by the entity number corresponding to one node, the relationship, and the entity number corresponding to the other node.

[0013] The role and effect of invention

[0014] According to the method and retrieval device for constructing a full-chain meat safety knowledge graph based on the present invention, on the one hand, hazard analysis is performed on the raw meat and processing of each meat product subcategory based on HACCP principles to identify hazards and corresponding control measures as key control points for constructing the full-chain meat safety knowledge graph. This combines the food itself with safety standards and HACCP, thereby integrating safety information across the entire meat supply chain. On the other hand, pathogens and contaminants, along with their corresponding specific standards, are connected to the corresponding raw meat nodes of the meat products and meat product subcategories. This allows users to directly and clearly obtain the relevant regulations on pathogens and contaminants most relevant to food safety for meat products or raw meat during the retrieval process, facilitating user queries. Therefore, the method for constructing a full-chain meat safety knowledge graph of the present invention can build a more comprehensive meat safety knowledge graph, and the retrieval device can retrieve and display the data of the meat safety knowledge graph more quickly, conveniently, and accurately. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the method for constructing a full-chain meat safety knowledge graph based on HACCP principles in an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the architecture of the meat product sub-nodes in the full-chain meat safety knowledge graph in an embodiment of the present invention;

[0017] Figure 3 This is a schematic diagram of the framework of the whole-chain meat safety knowledge graph retrieval device in an embodiment of the present invention;

[0018] Figure 4 This is a schematic diagram of the retrieval results of pathogenic bacteria and contaminants in meat products related to the meat product node in an embodiment of the present invention;

[0019] Figure 5 This is a schematic diagram illustrating the search results for specific labeling requirements of meat product pathogens and meat product contaminants related to the meat product node in an embodiment of the present invention. Detailed Implementation

[0020] To make the technical means, creative features, objectives and effects of this invention easy to understand, the following embodiments, in conjunction with the accompanying drawings, specifically illustrate the method for constructing a full-chain meat safety knowledge graph and the retrieval device of this invention.

[0021] In this embodiment, the method for constructing a full-chain meat safety knowledge graph based on HACCP principles is used to construct a full-chain meat safety knowledge graph based on multiple meat safety standard documents.

[0022] Among them, the meat safety standard document is the national meat safety standard document, which includes food safety standards, limits for contaminants, limits for pathogens, limits for pesticide and veterinary drug residues, and some local safety standards.

[0023] Figure 1 This is a flowchart illustrating the method for constructing a full-chain meat safety knowledge graph based on HACCP principles in an embodiment of the present invention.

[0024] like Figure 1 As shown, the method for constructing a full-chain meat safety knowledge graph based on HACCP principles is used to construct a full-chain meat safety knowledge graph based on multiple meat safety standard documents, including the following steps:

[0025] Step S1 involves extracting entities from all specific standards in all meat safety standard documents to obtain multiple standard testing objects and corresponding standard requirements.

[0026] The specific standards include the sensory requirements, technical requirements, microbial limits, pathogenic bacteria limits, and contaminant limits that meat products should meet.

[0027] For example, if a specific standard states that "the starch content of this product should meet the limit of 6g / 100g", then the standard test object obtained by physical sampling is "starch / (g / 100g)", and the corresponding standard requirement is "≤6.0".

[0028] Step S2: Extract relations from all meat safety standard documents to obtain multiple existing relations.

[0029] In this embodiment, the existing relationships include those mentioned in meat safety standard documents, such as "includes", "satisfies", and "complies".

[0030] Step S3: Construct a meat product node, then construct a meat product contaminant limit node, a meat product pathogen limit node, and various meat product sub-category nodes as lower-level nodes of the meat product node, and use the corresponding existing relationships as relationships between nodes.

[0031] The classification standard for the meat products sub-category corresponding to the meat products sub-category node is based on the food production license classification catalog. In this embodiment, the meat products sub-category node includes fermented meat products node, pre-prepared and cooked meat products node, cooked meat products node, and cured meat products node.

[0032] Step S4: For each meat product sub-category node, construct the corresponding raw meat node, safety standard node, and CCP selection node as the lower-level nodes of the meat product sub-category node, and use the corresponding existing relationships as the relationships between nodes.

[0033] Step S5: For each raw meat node, construct the corresponding raw meat contaminant limit node, raw meat pathogen limit node, and raw meat pesticide and veterinary drug residue node as the lower-level nodes of the raw meat node, and use the corresponding existing relationships as the relationships between nodes.

[0034] Step S6: For each safety standard node, construct the corresponding sensory requirement node, physicochemical requirement node, and microbiological requirement node as the lower-level nodes of the safety standard node, and use the corresponding existing relationships as the relationships between nodes.

[0035] Step S7: Take the corresponding standard test objects as the lower-level nodes of meat product contaminant limit node, meat product pathogenic bacteria limit node, raw meat contaminant limit node, raw meat pathogenic bacteria limit node, raw meat pesticide and veterinary drug residue node, sensory requirements node, physicochemical requirements node and microbiological requirements node, and then take the standard requirements as the lower-level nodes of the corresponding standard test objects, and take the corresponding existing relationships as the relationships between nodes.

[0036] Step S8: For each CCP selection node, conduct hazard analysis on the raw meat and processing of the meat product sub-category corresponding to the CCP selection node according to the HACCP principle, and obtain multiple critical control points as the lower-level nodes of the CCP selection node. The critical control points include specific hazards and corresponding control measures.

[0037] In this embodiment, when the meat product sub-category node can be further subdivided according to the food production license classification catalog, the lower-level nodes of the CCP selection node of the meat product sub-category node are the corresponding sub-category nodes, and the lower-level nodes of each sub-category node are the corresponding critical control points. For example, when the meat product sub-category node is the cured meat product node, the lower-level nodes of the corresponding CCP selection node include the meat filling product node and the ham product node. The lower-level nodes of the meat filling product node and the ham product node are the corresponding critical control points obtained according to the HACCP principle.

[0038] Step S9: For each raw meat node and meat product node, the node is directly connected to the corresponding related pathogenic bacteria and contaminant nodes through the first artificial relationship. That is, the raw meat node is directly connected to each of the subordinate nodes of the raw meat contaminant limit node and the raw meat pathogenic bacteria limit node corresponding to the node, and the meat product node is directly connected to each of the subordinate nodes of the meat product contaminant limit node and the meat product pathogenic bacteria limit node corresponding to the node. In this embodiment, the first artificial relationship is set to related.

[0039] Step S10: For each raw meat node and meat product node, the node is directly connected to the corresponding standard requirement nodes for pathogens and contaminants through the second artificial relationship. That is, the raw meat node is directly connected to the standard requirement nodes corresponding to each of the lower-level nodes of the raw meat contaminant limit node and the raw meat pathogen limit node, and the meat product node is directly connected to the standard requirement nodes corresponding to each of the lower-level nodes of the meat product contaminant limit node and the meat product pathogen limit node. In this embodiment, the first artificial relationship is set to must.

[0040] In this embodiment, artificial relationships can be added between nodes that are not directly hierarchical according to actual needs, so as to facilitate the query of relevant content. For example, an artificial relationship can be added between the cooked meat products node and the corresponding physicochemical requirements node, which is the lower-level node of the protein content node, so as to avoid interference from other physicochemical requirements standards and directly query the protein content requirements standard corresponding to cooked meat products.

[0041] All nodes and relationships between nodes constructed through steps S1 to S10 will be used as a knowledge graph for the entire meat safety chain.

[0042] Each node has a primary attribute and a secondary attribute. The primary attribute is the specific content of the node, and the secondary attribute is the specific content of the node's parent node. In this embodiment, the secondary attribute of the node without a parent node is set to foodgraph, so that all nodes have a secondary attribute.

[0043] Figure 2 This is a schematic diagram of the architecture of the meat product sub-nodes in the full-chain meat safety knowledge graph of this invention.

[0044] like Figure 2 As shown, the two nodes connected by the arrow curves indicate a relationship between them. The node in the direction of the arrow is a subordinate node of the other node. Within the framework of a meat product sub-category node, the meat product node is the superior node of that meat product sub-category node. The subordinate nodes of the meat product node include meat product contaminant limit nodes and meat product pathogen limit nodes. The subordinate nodes of this meat product sub-category node include raw meat nodes, safety labeling nodes, and CCP selection nodes. The subordinate nodes of the raw meat node include raw meat contaminant limit nodes, raw meat pathogen limit nodes, and pesticide and veterinary drug residue nodes. The subordinate nodes of the safety standard node include sensory requirements nodes, physicochemical requirements nodes, and microbiological requirements nodes. The subordinate nodes of the CCP selection node include CCP1 nodes, CCP2 nodes, etc., which correspond to multiple key control points of this meat product sub-category.

[0045] To facilitate the retrieval of the full-chain meat safety knowledge graph constructed in this embodiment, this embodiment also provides a full-chain meat safety knowledge graph retrieval device.

[0046] Figure 3 This is a schematic diagram of the framework of the whole-chain meat safety knowledge graph retrieval device in an embodiment of the present invention.

[0047] like Figure 3 As shown, the whole-chain meat safety knowledge graph retrieval device 100 includes a user input unit 10, a knowledge graph storage unit 20, a retrieval unit 30, and a display unit 40.

[0048] User input section 10 is used for users to input search statements containing specified relationships and specified nodes.

[0049] The knowledge graph storage unit 20 contains a database that stores data on the entire meat safety knowledge graph.

[0050] The whole-chain meat safety knowledge graph was constructed based on steps S1 to S10.

[0051] The database is Neo4j. The data storage method for the full-chain meat safety knowledge graph is as follows: Each node in the full-chain meat safety knowledge graph is numbered to obtain an entity number. Then, the node is transformed into a form represented by entity number, primary attribute, and secondary attribute, i.e., (entity number, primary attribute, secondary attribute). According to different meat safety standard documents, it is saved as different files. The relationship between two nodes is represented by the entity number of one node, the relationship, and the entity number of the other node, i.e., (node ​​A entity number, relationship, node B entity number). According to different relationships, it is saved as different files.

[0052] The retrieval unit 30 is used to obtain the nodes corresponding to the specified relationships and nodes from the whole-chain meat safety knowledge graph based on the retrieval statement as retrieval result nodes.

[0053] The content of the search result node includes the node's primary and secondary attributes.

[0054] Display unit 40 is used to display to the user the contents of the specified node and the search result node, as well as the specified relationship between the specified node and the search result node.

[0055] In this embodiment, the process of searching for pathogenic bacteria and contaminants in meat products related to meat product nodes using the full-chain meat safety knowledge graph retrieval device 100 is as follows:

[0056] First, the user input unit 10 inputs the search statement "MATCH p=(n:benti{name:'meat products'})-[r:rel2]->()RETURNp", where "meat products" is the specified node and "rel2" is the specified relationship, i.e., the first artificial relation related. Then, the retrieval unit 30 retrieves the corresponding node from the data of the whole chain meat safety knowledge graph stored in the knowledge graph storage unit 20 according to the search statement, and uses it as the search result node. Finally, the display unit 40 displays all the search result nodes, the specified node, and the relationships between the nodes.

[0057] Figure 4 This is a schematic diagram of the retrieval results of pathogenic bacteria and contaminants in meat products related to the meat product node in an embodiment of the present invention.

[0058] like Figure 4 As shown, the search results reveal that pathogens associated with meat products include Staphylococcus aureus, Salmonella, Listeria monocytogenes, and Escherichia coli, while contaminants associated with meat products include cadmium, lead, chromium, arsenic, N-nitrosodimethylamine, and benzo[a]pyrene. The relationships between the aforementioned pathogen nodes and contaminant nodes and the meat product nodes are all related. Therefore, by directly connecting the meat product nodes and the raw meat nodes corresponding to each meat product sub-category node with the corresponding pathogen nodes and contaminant nodes in the full-chain meat safety knowledge graph, the specific pathogens and contaminants associated with the product can be displayed intuitively, avoiding complex knowledge chain displays and improving search speed and accuracy.

[0059] In this embodiment, the process of retrieving specific standard requirements for pathogenic bacteria and contaminants in meat products related to meat product nodes using the full-chain meat safety knowledge graph retrieval device 100 is as follows:

[0060] First, the user input unit 10 inputs the search statement "MATCH p=(n:benti{name:'meat products'})-[r:rel1]->()RETURNp", where "meat products" is the specified node and "rel1" is the specified relationship, i.e., the second artificial relationship must. Then, the retrieval unit 30 retrieves the corresponding node as the search result node from the data of the whole chain meat safety knowledge graph stored in the knowledge graph storage unit 20 according to the search statement. Finally, the display unit 40 displays all the search result nodes, the specified node, and the relationships between the nodes. Furthermore, by clicking on the corresponding search result node, the user can view the secondary attributes of the node, i.e., the relevant specific standard requirements for pathogenic bacteria or contaminants in meat products, thus achieving specific and convenient retrieval.

[0061] In this embodiment, in order to display all the information of a node, the display unit 40 includes a left display block 401 and a right display block 402. The left display block 401 is used to display the content of the specified node and the search result node, as well as the specified relationship between the specified node and the search result node. The right display block 402 is used to display all the information of the specific node, including the node's id, primary attributes, and secondary attributes.

[0062] Figure 5 This is a schematic diagram showing the search results for specific standard requirements related to meat product pathogens and contaminants in an embodiment of the present invention.

[0063] like Figure 5 As shown, the left-hand display block 401 displays the meat product node, the specific standard requirements node for pathogenic bacteria in meat products, and the specific standard requirements node for contaminants in meat products obtained through retrieval. The relationship between the above specific standard requirements node and the meat product node is "must". After the user clicks to view the specific standard requirements node "n=5,c=0,m=0", the right-hand display block 402 displays the relevant information of the node, including the node's main attribute, i.e., name, which is "n=5,c=0,m=0", the secondary attribute, i.e., belong, which is "Salmonella", and the node's id, which is "45". Through the above display, users can know the standard testing object corresponding to the standard requirements when searching for standard requirements, and realize the quick access to key information on meat safety.

[0064] The role and effect of the embodiments

[0065] According to the method and retrieval device for constructing a full-chain meat safety knowledge graph involved in this embodiment, on the one hand, based on the HACCP principle, hazard analysis is performed on the raw meat and processing of each meat product subcategory to determine the hazards and corresponding control measures as key control points for constructing the full-chain meat safety knowledge graph. This combines the food itself with safety standards and HACCP, thereby satisfying the integration of meat safety information across the entire chain. On the other hand, pathogens and contaminants, along with their corresponding specific standards, are connected to the corresponding raw meat nodes of the meat products and meat product subcategories. This allows users to directly and clearly obtain the relevant regulations on pathogens and contaminants most relevant to food safety for meat products or raw meat during the retrieval process, facilitating user queries. In summary, this method can construct a more comprehensive meat safety knowledge graph and can retrieve and display the data of the meat safety knowledge graph more quickly, conveniently, and accurately.

[0066] The above embodiments are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention.

Claims

1. A method for constructing a whole-chain meat safety knowledge graph based on HACCP principles, for constructing a whole-chain meat safety knowledge graph according to a plurality of meat safety standard documents, characterized in that, Includes the following steps: Step S1: Extract entities from all the specific standards in the meat safety standard documents to obtain multiple standard testing objects and corresponding standard requirements; Step S2: Extract relations from all the meat safety standard documents to obtain multiple existing relations; Step S3: Construct a meat product node, and then construct a meat product contaminant limit node, a meat product pathogenic bacteria limit node, and each meat product subclass node as the lower-level nodes of the meat product node, and use the corresponding existing relationships as the relationships between nodes; Step S4: For each meat product sub-category node, construct the corresponding raw meat node, safety standard node, and CCP selection node as the lower-level nodes of the meat product sub-category node, and use the corresponding existing relationship as the relationship between nodes; Step S5: For each raw meat node, construct a corresponding raw meat contaminant limit node, raw meat pathogen limit node, and raw meat pesticide and veterinary drug residue node as the lower-level nodes of the raw meat node, and use the corresponding existing relationship as the relationship between nodes; Step S6: For each safety standard node, construct corresponding sensory requirement nodes, physicochemical requirement nodes, and microbiological requirement nodes as subordinate nodes of the safety standard node, and use the corresponding existing relationships as relationships between nodes; Step S7: Take each of the corresponding standard test objects as a lower-level node of the meat product contaminant limit node, the meat product pathogenic bacteria limit node, the raw meat contaminant limit node, the raw meat pathogenic bacteria limit node, the raw meat pesticide and veterinary drug residue node, the sensory requirement node, the physicochemical requirement node, and the microbiological requirement node, and then take the standard requirements as the lower-level node of the corresponding standard test object, and take the corresponding existing relationship as the relationship between nodes; Step S8: For each CCP selection node, perform hazard analysis on the raw meat and processing process of the meat product sub-category corresponding to the CCP selection node according to the HACCP principle, and obtain multiple critical control points as the lower-level nodes of the CCP selection node. The critical control points include specific hazards and corresponding control measures. Step S9: For each of the raw meat nodes and the meat product nodes, the node is directly connected to the corresponding related pathogenic bacteria and contaminant nodes through a first artificial relationship; Step S10: For each of the raw meat nodes and the meat product nodes, directly connect the node to the node corresponding to the relevant pathogens and contaminants according to the standard requirements through a second artificial relationship. All nodes and relationships between nodes constructed through steps S1 to S10 will be used as the whole-chain meat safety knowledge graph. Each node has a primary attribute and a secondary attribute. The primary attribute is the specific content of the node, and the secondary attribute is the specific content of the node's parent node.

2. The method for constructing a full-chain meat safety knowledge graph based on HACCP principles according to claim 1, characterized in that: wherein The meat safety standard document is a national meat safety standard document, which includes food safety standards, limits for contaminants, limits for pathogens, limits for pesticide and veterinary drug residues, and some local safety standards.

3. The method for constructing a full-chain meat safety knowledge graph based on HACCP principles according to claim 1, characterized in that: wherein In step S1, the specific standards include the specific content of the sensory requirements, technical requirements, microbial limits, pathogenic bacteria limits, and contaminant limits that meat products should meet.

4. The method for constructing a full-chain meat safety knowledge graph based on HACCP principles according to claim 1, characterized in that: wherein In step S3, the classification standard for the meat product subclass corresponding to the meat product subclass node is based on the food production license classification catalog.

5. A full-chain meat safety knowledge graph retrieval device constructed according to the full-chain meat safety knowledge graph construction method based on the HACCP principle according to any one of claims 1 to 4, characterized in that, include: The user input section is used for users to input search statements that include specified relationships and specified nodes; The knowledge graph storage unit contains a database that stores data on the entire meat safety knowledge graph. The retrieval unit is used to obtain the specified relationship and the node corresponding to the specified node from the whole-chain meat safety knowledge graph according to the retrieval statement as the retrieval result node; The display unit is used to display to the user the contents of the specified node and the search result node, as well as the specified relationship between the specified node and the search result node. The full-chain meat safety knowledge graph is constructed based on steps S1 to S10.

6. The whole-chain meat safety knowledge graph retrieval device according to claim 5, characterized in that: wherein, The content of the search result node includes the node's primary attributes and secondary attributes.

7. The whole-chain meat safety knowledge graph retrieval device according to claim 5, characterized in that: wherein The database is the Neo4j database, and the data of the full-chain meat safety knowledge graph is stored in the following way: Each node in the full-chain meat safety knowledge graph is numbered to obtain an entity number. The node is then converted into a form represented by the entity number, the main attribute, and the secondary attribute, and saved as different files according to different meat safety standard documents. The relationship between two nodes is represented by the entity number corresponding to one node, the relationship, and the entity number corresponding to the other node, and saved as different files according to different relationships.

Images