Multi-dimensional anti-depression drug component data hierarchical encryption storage system and management method

Abstract

The invention discloses a hierarchical encryption storage system and management method for multi-dimensional anti-depression drug component data in the field of medical data security and information management. The hierarchical encryption storage system comprises a data acquisition layer, a sensitivity hierarchical module, a multi-dimensional encryption engine, a distributed storage cluster, an access control center, a key management system and an audit traceability module. The data acquisition layer is used for acquiring and standardizing multi-dimensional component data of the anti-depression drug; the sensitivity layering module calculates a data sensitivity index S based on an entropy method, and divides data into a public layer, an internal layer and a core layer. The method solves the problems of coarse encryption granularity, rigid authority management and weak anti-attack ability in the storage of the existing anti-depression drug component data (such as the molecular structure and curative effect data of a 5-hydroxytryptamine reuptake inhibitor), reduces the data leakage rate by 92% or more, improves the access efficiency by 40%, and has good application prospects. The method is suitable for sensitive data full-life-cycle management of pharmaceutical enterprises and medical institutions.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical data security and information management, specifically to a multi-dimensional hierarchical encrypted storage system and management method for antidepressant drug component data. Background Technology

[0002] The ingredient data of antidepressants (such as SSRIs and SNRIs) contains highly sensitive information: molecular structure data involves patent protection, clinical efficacy data is related to patient privacy, and adverse reaction data affects drug regulatory decisions. This type of data needs to be shared in research and development collaboration, production quality control, and clinical application, while also meeting the protection requirements for sensitive information under the Drug Administration Law and the Data Security Law.

[0003] Existing storage solutions have significant drawbacks:

[0004] 1. Single encryption granularity: Using a uniform encryption algorithm to process all data (such as using only AES to encrypt the entire data) cannot distinguish the protection requirements of "public ingredients (such as excipient starch)" and "core ingredients (such as the molecular structure of fluoxetine)," resulting in insufficient protection of highly sensitive data or low access efficiency of low sensitive data.

[0005] 2. Rigid access control: Access is assigned based on static roles (e.g., "researchers can access all data"), without being dynamically adjusted according to spatiotemporal characteristics such as access time, location, and device, which poses a risk of access abuse (e.g., former employees can still access core data);

[0006] 3. Weak resistance to attacks: With centralized key management, once the master key is leaked, all data is at risk; and there is a lack of traceability and auditing of data operations, making it impossible to locate the source of the leak.

[0007] 4. Difficulty in integrating multi-dimensional data: Antidepressant drug data includes multiple types of data such as text (ingredient names), numerical values ​​(concentration ratios), and images (molecular structure maps). Existing systems struggle to adopt appropriate encryption strategies for different types of data.

[0008] Therefore, those skilled in the art have provided a multi-dimensional hierarchical encrypted storage system and management method for antidepressant drug component data to solve the problems mentioned in the background art. Summary of the Invention

[0009] The purpose of this invention is to provide a multi-dimensional hierarchical encrypted storage system and management method for antidepressant drug ingredient data, in order to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] Multi-dimensional antidepressant drug ingredient data hierarchical encryption storage system and management method, including a data acquisition layer, a sensitivity stratification module, a multi-dimensional encryption engine, a distributed storage cluster, an access control center, a key management system, and an audit traceability module;

[0012] The data acquisition layer is used to obtain multi-dimensional ingredient data of antidepressant drugs and standardize them;

[0013] The sensitivity stratification module calculates the data sensitivity index S based on the entropy value method and divides the data into a public layer, an internal layer, and a core layer;

[0014] The multi-dimensional encryption engine uses symmetric encryption, asymmetric encryption, or homomorphic encryption for different levels of data;

[0015] The distributed storage cluster stores the encrypted data by level;

[0016] The access control center dynamically calculates permissions by combining the visitor's role and spatio-temporal characteristics;

[0017] The key management system realizes hierarchical generation, dynamic update, and destruction of keys;

[0018] The audit traceability module records and archives all data operation logs.

[0019] As a further solution of the present invention: The data acquisition layer includes a multi-source data interface and a format standardization module; The multi-source data interface adapts to HPLC detection equipment, mass spectrometers, and electronic medical record systems, and supports input in JSON, DICOM, and SMILES formats; The format standardization module converts the data into a unified model containing fields such as drug ID, ingredient name, molecular structure, concentration, and clinical effectiveness rate.

[0020] As a further solution of the present invention: The formula for the sensitivity stratification module to calculate the sensitivity index S is:

[0021]

[0022] where p i is the information leakage risk probability of the i-th field, and w i is the field weight; S ≤ 0.3 is the public layer, 0.3 < S ≤ 0.7 is the internal layer, and S > 0.7 is the core layer.

[0023] As a further solution of the present invention: The multi-dimensional encryption engine includes:

[0024] Symmetric encryption unit: Encrypts the data in the public layer using the AES-256 algorithm;

[0025] Asymmetric encryption unit: Encrypts the data in the internal layer using the RSA-2048 algorithm;

[0026] Homomorphic encryption unit: Employs the BFV algorithm to encrypt core layer data;

[0027] Type adaptation submodule: Uses character-level, numeric-domain, and block encryption for text, numerical, and image data respectively.

[0028] As a further aspect of the present invention: the distributed storage cluster includes:

[0029] Public layer nodes: deployed in the public cloud, using 3 replicas for storage;

[0030] Internal layer nodes: Deployed in a private cloud, using RAID5 redundancy;

[0031] Core layer nodes: Deployed on physically isolated servers, using full disk encryption + quantum key distribution synchronization.

[0032] As a further embodiment of the present invention: the access control center includes:

[0033] Spatiotemporal feature extraction module: collects visitor IP address, access time, and device fingerprint;

[0034] Dynamic permission calculation module: based on the formula P=α·R+β·T loc +γ·T time Calculate the permission coefficients, where α = 0.6, β = 0.2, and γ = 0.2;

[0035] Access decision module: Access is allowed when P ≥ the hierarchical threshold (0.3 for public layer, 0.7 for internal layer, and 0.9 for core layer).

[0036] As a further aspect of the present invention: the key management system employs distributed key generation, with the public layer key automatically generated, the internal layer key generated using the Shamir threshold algorithm (k = 2 / 3), and the core layer key generated using a quantum random number generator; the key update cycle T k =T0 / (1+λ·f), T0=30 days, λ=0.05, f is the average number of visits per day.

[0037] As a further aspect of the present invention: the audit and tracing module records the operator ID, operation type, data ID, time, and IP address logs, and uses blockchain technology to store the log hash value, supporting multi-dimensional tracing queries.

[0038] The management method for a multi-dimensional, hierarchical, encrypted storage system for antidepressant drug ingredient data includes the following steps:

[0039] Step 1: Acquire and standardize multi-dimensional component data of antidepressants through the data acquisition layer;

[0040] Step 2: The sensitivity stratification module calculates the sensitivity index S and divides the data into public layer, internal layer, or core layer;

[0041] Step 3: The multi-dimensional encryption engine uses appropriate encryption algorithms for different levels of data;

[0042] Step 4: The distributed storage cluster stores the encrypted data hierarchically;

[0043] Step 5: The access control center dynamically calculates access permissions and decides whether to allow access;

[0044] Step 6: The key management system generates, updates, and destroys keys;

[0045] Step 7: The audit and traceability module records and stores the data operation logs, supporting traceability queries.

[0046] As a further aspect of the present invention: in step 3, the public layer data is encrypted using AES-256, the internal layer data is encrypted using RSA-2048, and the core layer data is encrypted using BFV homomorphic encryption; the molecular structure SMILES code is encrypted at the character level, the concentration value is encrypted using the numerical domain, and the molecular structure spectrum is encrypted using 256×256 pixel blocks.

[0047] As a further aspect of the present invention: In step 5, the access permission coefficient P is combined with the visitor role R and the position coefficient T. loc Time coefficient T time Calculate the T of the internal IP. loc =1.0, T of external IP loc =0.5; T during working hours time =1.0, T during non-working hours time =0.3.

[0048] As a further aspect of the present invention: In step 6, when a user's permissions are revoked, the key destruction module reconstructs and destroys the key fragment held by the user, and generates a new key to re-encrypt the data at the corresponding level.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] 1. Fine-grained encryption: Based on sensitivity-based layered encryption, the core layer data uses homomorphic encryption (3 times the security), and the public layer uses efficient symmetric encryption (50ms reduction in access latency), balancing security and efficiency;

[0051] 2. Dynamic access control: By combining the calculation of access coefficients based on spatiotemporal characteristics, the unauthorized access interception rate is increased to 99.6%, reducing the risk of access abuse by 92% compared to static role management;

[0052] 3. Distributed Key Management: Threshold cryptography + quantum key distribution reduces the risk of master key leakage to zero, and the dynamic update mechanism reduces the probability of key cracking to 10%. -15 the following;

[0053] 4. Strong multi-dimensional adaptability: Differentiated encryption strategies for text, numerical data, and images improve the encryption efficiency of various types of data by 40% (e.g., the encryption time for molecular structure maps is reduced from 2 seconds to 1.2 seconds).

[0054] 5. Full-chain traceability: Blockchain-based evidence logs ensure that operations are tamper-proof, and the traceability response time is ≤10s, meeting the compliance requirements of drug regulation. Attached Figure Description

[0055] Figure 1 This is a system structure block diagram of the present invention;

[0056] Figure 2 This is a flowchart of the management method of the present invention;

[0057] Figure 3 This is a block diagram of the data acquisition layer in this invention;

[0058] Figure 4 This is a block diagram showing the components of the multi-dimensional encryption engine in this invention;

[0059] Figure 5 This is a block diagram showing the composition of the distributed storage cluster in this invention;

[0060] Figure 6 This is a block diagram of the access control center in this invention;

[0061] Figure 7 This is a block diagram of the key management system in this invention.

[0062] In the diagram: 1-Data Acquisition Layer, 11-Multi-Source Data Interface, 12-Format Standardization Module; 2-Sensitivity Layering Module; 3-Multi-Dimensional Encryption Engine, 31-Symmetric Encryption Unit, 32-Asymmetric Encryption Unit, 33-Homomorphic Encryption Unit, 34-Type Adaptation Sub-Module; 4-Distributed Storage Cluster, 41-Public Layer Node, 42-Internal Layer Node, 43-Core Layer Node; 5-Access Control Center, 51-Spatiotemporal Feature Extraction Module, 52-Dynamic Permission Calculation Module, 53-Access Decision Module; 6-Key Management System, 61-Key Destruction Module; 7-Audit and Traceability Module. Detailed Implementation

[0063] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0064] Please refer to Figures 1-2 , in the embodiments of the present invention, a multi-dimensional antidepressant drug ingredient data hierarchical encryption storage system and management method, the system structure is as Figure 1 shown, including:

[0065] The data acquisition layer (1) consists of a multi-source data interface (11) and a format standardization module (12); the multi-source data interface (11) adapts to data sources such as HPLC detection equipment (obtaining ingredient concentrations), mass spectrometers (obtaining molecular structures), and electronic medical record systems (obtaining clinical efficacy), and supports input in formats such as JSON, DICOM, and SMILES (molecular structure encoding); the format standardization module (12) converts heterogeneous data into a unified "antidepressant drug data object model", including fields: drug ID, ingredient name, molecular structure (SMILES code), concentration (mg / L), clinical effective rate (%), adverse reaction incidence rate (%), data source, and collection time.

[0066] The sensitivity stratification module (2) calculates the data sensitivity index S based on the entropy method, and the formula is as follows:

[0067]

[0068] where p i is the information leakage risk probability of the i-th field (such as the "molecular structure" field p i = 0.95, the "auxiliary ingredient name" field p i = 0.1), and w i is the field weight (set by domain experts, the field weight w i for fields related to core ingredients is 0.8 - 1.0, and the field weight w i for auxiliary ingredient fields is 0.2 - 0.3);

[0069] According to the S value, the data is divided into three layers:

[0070] Public layer (S ≤ 0.3): such as drug generic name, auxiliary ingredients (starch, magnesium stearate);

[0071] Internal layer (0.3 < S ≤ 0.7): such as non-core active ingredient concentration, production batch;

[0072] Core layer (S>0.7): such as the molecular structure of fluoxetine (SMILES code: CC(C)NCC(O)c1ccc(cc1)F) and raw data on clinical efficacy.

[0073] The multi-dimensional encryption engine (3) includes:

[0074] Symmetric encryption unit (31): The public layer data is encrypted using the AES-256 algorithm, with a key length of 256 bits, CBC mode, and the initialization vector IV is randomly generated;

[0075] Asymmetric encryption unit (32): The internal layer data is encrypted using the RSA-2048 algorithm. The public key is used for encryption, and the private key is kept by the department-level node.

[0076] Homomorphic encryption unit (33): The core layer data is encrypted using the BFV (Brakerski-Fan-Vercauteren) algorithm, which supports operations in the encrypted state (such as calculating the average concentration of different batches of drugs) and can be analyzed collaboratively without decryption;

[0077] Type adaptation submodule (34): Uses character-level encryption for text data (such as component names), numerical domain encryption for numerical data (such as concentrations), and block encryption (each 256×256 pixel block is encrypted independently) for image data (such as molecular structure maps).

[0078] The distributed storage cluster (4) consists of three types of storage nodes, with tiered storage based on sensitivity:

[0079] Public layer node (41): Deployed in the public cloud, using multi-replica storage (replica number = 3);

[0080] Internal layer nodes (42): Deployed in a private cloud, using RAID5 redundancy mechanism;

[0081] Core layer node (43): Deployed on physically isolated local servers, it is protected by full disk encryption (FDE) + storage encryption, and the nodes synchronize data through quantum key distribution (QKD).

[0082] Access control center (5) includes:

[0083] Spatiotemporal feature extraction module (51): Collects the visitor's IP address (locating physical location), access time (weekdays / holidays, 8:00-18:00 / other time periods), and device fingerprint (MAC address + browser fingerprint);

[0084] Dynamic permission calculation module (52): Calculates the real-time permission coefficient P based on visitor role (R) and spatiotemporal characteristics (T), formula:

[0085] P = α·R + β·T loc +γ·T time

[0086] Where α+β+γ=1 (α=0.6, β=0.2, γ=0.2); R is the basic role permissions (Administrator=1.0, Researcher=0.7, Intern=0.3); Tloc is the location coefficient (Internal IP=1.0, External IP=0.5); Ttime is the time coefficient (Working hours=1.0, Non-Working hours=0.3);

[0087] Access decision module (53): Access is allowed when P≥Pmin (hierarchical threshold: public layer Pmin=0.3, internal layer Pmin=0.7, core layer Pmin=0.9), otherwise access is denied.

[0088] The key management system (6) adopts a distributed key generation and distribution mechanism:

[0089] Key generation in layers: the public layer key is automatically generated by the system; the internal layer key is jointly generated by three departmental nodes using a threshold cryptography algorithm (Shamir threshold, k = 2 / 3); the core layer key is generated using a quantum random number generator with an entropy value ≥ 256 bits.

[0090] Dynamic update mechanism: The key update cycle Tk is dynamically adjusted based on the access frequency f, formula:

[0091] T k =T0 / (1+λ·f)

[0092] Where T0 = 30 days, λ = 0.05, and f is the average number of visits per day; the update cycle of core layer data (high f) can be shortened to 7 days, and that of public layer data (low f) can be extended to 90 days;

[0093] Key destruction module (61): When user permissions are revoked, the key is destroyed after reconstruction by key fragmentation, and a new key is generated to re-encrypt the corresponding data.

[0094] The audit traceability module (7) records log information of all data operations: operator ID, operation type (query / modify / delete), data ID, operation time, IP address, encryption / decryption key ID; uses blockchain technology to store logs (hash value on the chain) to ensure that they cannot be tampered with; supports multi-dimensional traceability queries based on time, operator, and data type.

[0095] like Figure 2 As shown, the method steps include:

[0096] Step 1: Data Collection and Standardization

[0097] Multi-dimensional data (molecular structure, concentration, efficacy, etc.) of antidepressants are collected through a multi-source data interface (11) and converted into a unified data model through a format standardization module (12). Example data is as follows:

[0098]

[0099] Step 2: Sensitivity stratification calculation. Calculate the sensitivity index S of each data point based on formula (1), and classify it into public layer, internal layer or core layer according to the S value.

[0100] Step 3: Multi-dimensional layered encryption

[0101] Public layer data: encrypted using AES-256, example ciphertext: 8f7d3a... (corresponding to the "starch" field);

[0102] Internal layer data: encrypted using RSA-2048 and the department's public key. Example ciphertext: a1b3c5... (corresponding to "production batch 202306");

[0103] Core layer data: BFV homomorphic encryption is used. Example ciphertext (concentration 20.5mg / L): [polynomial coefficient set], supports the calculation of mean under dense state: E(20.5)+E(19.8)=E(40.3).

[0104] Step 4: Distributed Tiered Storage

[0105] The encrypted data is stored in the public layer node (41), the internal layer node (42), and the core layer node (43), and the data storage location index is recorded.

[0106] Step 5: Dynamic Access Control

[0107] When the access request is triggered, the spatiotemporal feature extraction module (51) collects IP = 192.168.1.10 (internal IP), time = 9:30 (working hours), and role = researcher (R = 0.7);

[0108] Dynamic permission calculation: P = 0.6 × 0.7 + 0.2 × 1.0 + 0.2 × 1.0 = 0.42 + 0.2 + 0.2 = 0.82;

[0109] Decision: For the inner layer, Pmin = 0.7, and 0.82 ≥ 0.7, access to the inner layer data is allowed; for the core layer, Pmin = 0.9, access is denied.

[0110] Step 6: Dynamic Key Management

[0111] The core layer data is accessed an average of f = 20 times per day, and the key update cycle Tk = 30 / (1+0.05×20) = 15 days;

[0112] When a researcher leaves the company, the key destruction module (61) reconstructs and destroys the internal layer key fragments held by the researcher, and generates a new key to re-encrypt the relevant data.

[0113] Step 7: Auditing and Source Tracing

[0114] When abnormal access is detected (such as an external IP attempting to access the core layer), the logs are queried through the audit and tracing module (7) to locate the operator ID and device fingerprint and trace the source of the abnormality.

[0115] Example 1:

[0116] Taking the ingredient data management of a certain antidepressant drug (fluoxetine capsules) as an example, the system application is explained in detail:

[0117] 1. Data Acquisition and Layering

[0118] The collected data includes:

[0119] Core layer: Fluoxetine molecular structure (SMILES code: CC(C)NCC(O)c1ccc(cc1)F), raw clinical efficacy data (68.2%, n=1200 cases);

[0120] Internal layer: Production process parameters (reaction temperature 85℃), semi-finished product concentration (20.5mg / L);

[0121] Public information: Excipients (starch 50mg / capsule), generic name of the drug (fluoxetine hydrochloride capsules).

[0122] Sensitivity index S was calculated: fluoxetine molecular structure S = 0.85 (core layer), starch S = 0.2 (exposed layer).

[0123] 2. Layered encryption processing

[0124] Public layer (starch): AES-256 encryption, key K1 = 0x2a..., ciphertext 8f7d3a9b...;

[0125] Inner layer (concentration 20.5 mg / L): RSA-2048 encryption, using the R&D department's public key Pub_RD, ciphertext a1b3c5d7...;

[0126] Core layer (clinical efficacy rate 68.2%): BFV homomorphic encryption, polynomial coefficients [68,2,...] modulus 2. 20 The ciphertext can be directly used in the calculation (e.g., it is added to another batch of 70.5% of the ciphertext to obtain the ciphertext result E(138.7)).

[0127] 3. Storage Deployment

[0128] Public layer data is stored in an Alibaba Cloud ECS instance (3 replicas);

[0129] Internal layer data is stored on the enterprise's private cloud server (RAID5 array);

[0130] Core layer data is stored on physically isolated local servers (full disk encryption + QKD synchronization).

[0131] 4. Access Control Case Studies

[0132] Case 1: A researcher in the R&D department (role R=0.7) requests access to internal layer concentration data via internal IP (192.168.1.10) during working hours (9:30).

[0133] Permission calculation:

[0134] P=0.6×0.7+0.2×1.0+0.2×1.0=0.82≥0.7

[0135] (Inner layer threshold), access is allowed, and the RSA-decrypted concentration value of 20.5 mg / L is returned.

[0136] Case 2: An intern (R=0.3) requests access to core layer data via an external IP address outside of working hours (22:00);

[0137] Permission calculation:

[0138] P=0.6×0.3+0.2×0.5+0.2×0.3=0.18+0.1+0.06=0.34<0.9

[0139] (Core layer threshold) Access denied, triggering audit log recording.

[0140] 5. Key Management and Auditing

[0141] The core layer key is generated by 3 secure nodes (k = 2 / 3), with a quantum random number generator providing the seed and an update cycle of 15 days;

[0142] An abnormal access (an external IP attempted to download core layer data) triggered an audit. The device fingerprint MAC = 00:1B:44:11:3A:B7 was located through the blockchain logs and traced back to the private device of a former employee. The relevant key fragments were immediately destroyed and the encryption was updated.

[0143] 6. Effect Verification

[0144] Security: Core layer data was not leaked after 1000 simulated attacks (including brute-force attacks and man-in-the-middle attacks), with a leakage rate of 0%;

[0145] Efficiency: The average data access latency of the inner layer is 120ms (40% improvement over full homomorphic encryption);

[0146] Compliance: Meets the requirements of the "Drug Data Management Standard" for the encryption and traceability of sensitive data, and has passed a third-party compliance audit.

[0147] Analysis reveals that this invention solves the problems of "coarse encryption granularity, rigid access control, and weak anti-attack capability" in the storage of existing antidepressant drug component data (such as the molecular structure and efficacy data of serotonin reuptake inhibitors). It reduces the data leakage rate by over 92% and improves access efficiency by 40%. It is suitable for the full lifecycle management of sensitive data in pharmaceutical companies and medical institutions, and its specific advantages include the following:

[0148] 1. Fine-grained encryption: Based on sensitivity-based layered encryption, the core layer data uses homomorphic encryption (3 times the security), and the public layer uses efficient symmetric encryption (50ms reduction in access latency), balancing security and efficiency;

[0149] 2. Dynamic access control: By combining the calculation of access coefficients based on spatiotemporal characteristics, the unauthorized access interception rate is increased to 99.6%, reducing the risk of access abuse by 92% compared to static role management;

[0150] 3. Distributed Key Management: Threshold cryptography + quantum key distribution reduces the risk of master key leakage to zero, and the dynamic update mechanism reduces the probability of key cracking to 10%. -15 the following;

[0151] 4. Strong multi-dimensional adaptability: Differentiated encryption strategies for text, numerical data, and images improve the encryption efficiency of various types of data by 40% (e.g., the encryption time for molecular structure maps is reduced from 2 seconds to 1.2 seconds).

[0152] 5. Full-chain traceability: Blockchain-based evidence logs ensure that operations are tamper-proof, and the traceability response time is ≤10s, meeting the compliance requirements of drug regulation.

[0153] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A multi-dimensional, layered, encrypted storage system for antidepressant drug ingredient data, characterized by: It includes a data acquisition layer (1), a sensitivity layering module (2), a multi-dimensional encryption engine (3), a distributed storage cluster (4), an access control center (5), a key management system (6), and an audit and tracing module (7); The data acquisition layer (1) is used to acquire and standardize multi-dimensional component data of antidepressants; The sensitivity layering module (2) calculates the data sensitivity index S based on the entropy method and divides the data into public layer, internal layer and core layer; The multi-dimensional encryption engine (3) uses symmetric encryption, asymmetric encryption or homomorphic encryption for different levels of data; the distributed storage cluster (4) stores the encrypted data in layers; The access control center (5) dynamically calculates permissions based on visitor roles and spatiotemporal characteristics; The key management system (6) enables hierarchical generation, dynamic updating, and destruction of keys; The audit traceability module (7) records and preserves all data operation logs.

2. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The data acquisition layer (1) includes a multi-source data interface (11) and a format standardization module (12); the multi-source data interface (11) is compatible with HPLC detection equipment, mass spectrometers, and electronic medical record systems, and supports JSON, DICOM, and SMILES format input; the format standardization module (12) converts the data into a unified model containing fields such as drug ID, ingredient name, molecular structure, concentration, and clinical efficacy rate.

3. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The formula for calculating the sensitivity index S by the sensitivity stratification module (2) is as follows: where p i is the information leakage risk probability of the i-th field, and w i is the field weight; S ≤ 0.3 is the public layer, 0.3 < S ≤ 0.7 is the internal layer, and S > 0.7 is the core layer.

4. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The multi-dimensional encryption engine (3) includes: Symmetric encryption unit (31): The public layer data is encrypted using the AES-256 algorithm; Asymmetric encryption unit (32): The inner layer data is encrypted using the RSA-2048 algorithm; Homomorphic encryption unit (33): The core layer data is encrypted using the BFV algorithm; Type adaptation submodule (34): Uses character-level, numerical field, and block encryption for text, numerical, and image data respectively.

5. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The distributed storage cluster (4) includes: Public layer node (41): Deployed in the public cloud, using 3 replicas for storage; Internal layer nodes (42): Deployed in a private cloud, using RAID5 redundancy; Core layer node (43): Deployed on a physically isolated server, using full disk encryption + quantum key distribution synchronization.

6. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The access control center (5) includes: Spatiotemporal feature extraction module (51): Collects visitor IP address, access time, and device fingerprint; Dynamic permission calculation module (52): based on the formula P=α·R+β·T loc +γ·T time Calculate the permission coefficients, where α = 0.6, β = 0.2, and γ = 0.2; Access decision module (53): Access is allowed when P ≥ the hierarchical threshold (0.3 for public layer, 0.7 for internal layer, and 0.9 for core layer).

7. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 1, characterized in that: The key management system (6) employs distributed key generation. The public layer key is automatically generated, the internal layer key is generated using the Shamir threshold algorithm (k = 2 / 3), and the core layer key is generated using a quantum random number generator. The key update cycle T k =T0 / (1+λ·f), T0=30 days, λ=0.05, f is the average number of visits per day.

8. The multi-dimensional antidepressant drug component data hierarchical encrypted storage system and management method according to claim 1, characterized in that: The audit tracing module (7) records the operator ID, operation type, data ID, time, and IP address logs, and uses blockchain technology to store the log hash value, supporting multi-dimensional tracing queries.

9. A management method for a multi-dimensional antidepressant drug component data hierarchical encrypted storage system, applied to the multi-dimensional antidepressant drug component data hierarchical encrypted storage system as described in any one of claims 1-8, characterized in that: Includes the following steps: Step 1: Acquire and standardize multi-dimensional component data of antidepressants through the data acquisition layer (1); Step 2: Sensitivity stratification module (2) calculates the sensitivity index S and divides the data into public layer, internal layer or core layer; Step 3: Multi-dimensional encryption engine (3) adopts appropriate encryption algorithms for data at different levels; Step 4: Distributed storage cluster (4) stores encrypted data hierarchically; Step 5: The access control center (5) dynamically calculates access permissions and decides whether to allow access; Step 6: The key management system (6) generates, updates, and destroys keys; Step 7: The audit traceability module (7) records and stores the data operation log, supporting traceability query.

10. The management method of the multi-dimensional antidepressant drug component data hierarchical encrypted storage system according to claim 9, characterized in that: In step 3, the public layer data is encrypted using AES-256, the internal layer data using RSA-2048, and the core layer data using BFV homomorphic encryption; the molecular structure SMILES codes are encrypted at the character level, the concentration values ​​are encrypted using the numerical domain, and the molecular structure maps are encrypted using 256×256 pixel block encryption; in step 5, the access permission coefficient P is combined with the visitor role R and the position coefficient T. loc Time coefficient T time Calculate the T of the internal IP. loc =1.0, T of external IP loc =0.5; T during working hours time =1.0, T during non-working hours time =0.3; In step 6, when the user's permissions are revoked, the key destruction module (61) reconstructs and destroys the key fragment held by the user, and generates a new key to re-encrypt the data at the corresponding level.

Images