Odor compound database and rapid screening method development and software system establishment thereof

Abstract

The invention discloses an odor compound database, development of a rapid screening method and establishment of a software system of the odor compound database, and belongs to the technical field of character image recognition and establishment and retrieval of compound databases. The method comprises the following steps: constructing a smell database, extracting a CAS number and a corresponding smell description in a smell query file by preloading a smell substance database in an Excel format and a PDF analyzer, and accelerating initialization by adopting a cache mechanism; analyzing the volatile compound detection report, and extracting peak area and compound key information based on a PDF analyzer; matching is performed through a double-database matching mechanism, so that rapid screening of odor compounds and dynamic updating of databases are realized; and finally, outputting a retrieval result through the Web interactive interface. The analysis efficiency of the volatile compound detection report and the screening accuracy of the odor compounds are remarkably improved, sustainable expansion of a database is supported, operation is easy and convenient, and the method is suitable for rapid identification and retrieval of the volatile odor compounds.

Description

Technical Field

[0001] This invention relates to the technical fields of text and image recognition, compound database establishment, and retrieval methods. Specifically, it relates to the development of a software system for establishing an odor compound database, identifying and screening detection results, and related software. Background Technology

[0002] In recent years, gas chromatography-mass spectrometry (GC-MS) has been widely used in the detection of volatile compounds in many fields such as food, environment, and medicine due to its advantages such as accurate qualitative analysis, high sensitivity, low detection limit, good reproducibility, and strong anti-interference ability. It plays an irreplaceable and crucial role, especially in the identification and analysis of odor compounds. However, traditional GC-MS results are obtained by manually extracting content from PDF reports and conducting manual searches. This process is complex, lacks sufficient reference materials, and cannot quickly and effectively screen out odor compounds from the test results. Therefore, establishing an odor compound database that complements the detection technology and developing methods for rapid screening of odor substances have become urgent problems to be solved.

[0003] Currently, the application and development of computers have made combining technology and data to achieve rapid detection a major future development direction. Among various programming languages, Python's syntax is concise and clear, with a rich standard library and third-party libraries covering various programming needs, making it possible to compile odor compound databases and screening systems compatible with detection technologies. Within the Flask framework, different databases can be used to store application data, and combined with tools such as NumPy and Pandas, detection results can be effectively parsed, thus laying a solid foundation for establishing odor compound databases and developing rapid odor compound screening methods. Summary of the Invention

[0004] The present invention aims to provide an odor compound database and a rapid screening method that are compatible with detection technology. This method can quickly identify and screen odor compounds in volatile compound detection results reports and provides an expandable odor compound database to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] Step S100: Construction and initialization of the multi-source odor database. A structured odor database (Excel format) is preloaded, containing fields such as CAS number, compound name, odor threshold, and odor characteristics. The CAS number and odor description correspondence in the unstructured document "odor query.pdf" are automatically extracted using a PDF parsing engine. The parsing results are stored using a caching mechanism to accelerate the system initialization process; the cache file is "odor query.cache".

[0007] Step S200: Intelligent parsing and feature extraction of the GC-MS report, employing a three-stage text processing workflow:

[0008] Stage S210: Peak Area Data Extraction. In the "Peak Report TIC" area of ​​the location report, extract all values ​​greater than 10000 as candidate peak areas. Filter valid peak area data according to the odd / even index rule and establish a mapping relationship between peak area and compound records.

[0009] Phase S220: Compound Feature Information Capture. Key fields are matched using regular expression patterns. These include the Formula field: "Formula:(.*?)CAS"; the CAS number field: "CAS:(.*?)MolWeight"; and the RetIndex field: "RetIndex:(.*?)CompName". Only high-confidence matches with Hit# of 1 or 2 are extracted to construct a standardized compound information table.

[0010] Stage S230: Data Alignment and Integration. Peak area data and compound characteristic information are correlated according to the detection order, row index information is supplemented, and a complete data input table is established.

[0011] Step S300: Dual-cascade database matching mechanism:

[0012] Main matching path S310: Performs inner join matching between the input data and the enhanced odor database, and outputs results for known odor compounds;

[0013] Auxiliary matching path S320: Connects unmatched data with the extended odor database to identify new odor compounds;

[0014] Database update path S330: Automatically appends newly identified compound information to "Resin Odor Substances Database Retrieval_Writable Version.xlsx" to achieve dynamic database expansion.

[0015] Step S400: Web Interaction and Result Visualization. A web service interface is provided through the Flask framework, supporting PDF file uploads and parsing request submissions. Matching results are converted into a tabular data structure and rendered using an HTML template to display complete odor compound information in an interactive tabular format.

[0016] The database retrieves the relevant systems, including:

[0017] Data acquisition module: Used to receive GC-MS report PDF files uploaded by users.

[0018] PDF parsing engine: Implements report text extraction and feature recognition based on the Fitz library.

[0019] Cache Management Unit: Responsible for caching and loading scent query results. Database Management Module: Manages the main database, extended databases, and performs dynamic updates.

[0020] Core of matching processing: Execution of the double-cascade matching algorithm

[0021] Web service interface: Provides user interaction and data display functions.

[0022] The preferred embodiment of this invention is an odor compound database system, which includes odor compound thresholds and odor descriptions of various raw materials accumulated from laboratory research, information compiled from books such as "Threshold Values ​​in Air, Water and Other Media" and "Fenaroli's Handbook of Flavor Ingredients", and professional chemical websites, to screen and identify odor compounds among volatile compounds, and to update unknown odor compounds.

[0023] Preferably, the database system of this invention includes a PDF report parsing function, a compound information extraction and identification function, a database update function, and a web-based upload and output function. The PDF report parsing function is used to identify uploaded volatile compound detection reports. The compound information extraction and identification function is used to extract and identify information such as peak area, odor description, and odor threshold of compounds. The database update function is used to supplement the odor data with information on newly detected odor compounds. The web-based system can upload PDF reports and output the odor compound screening results and detailed information in a table.

[0024] The beneficial effects of this invention are:

[0025] 1. Improved report parsing efficiency: From input to output of odor compound results, only PDF parsing technology is needed, and an automatic alignment algorithm for area data and compound records is implemented, which greatly reduces the manual processing time of reports.

[0026] 2. High accuracy in matching odor compound characteristics; cross-referencing search results from public and manually managed databases ensures the accuracy of odor compound screening.

[0027] 3. Database data sustainability: Information on newly added compounds selected from the test results is appended to the exported odor substance database retrieval writable version .xlsx, enabling dynamic database updates.

[0028] 4. Simple and convenient operation: The web-based interactive interface allows for the screening and searching of odor compounds. The operation is easy and the output results are simple and clear. Attached Figure Description

[0029] AppendixFigure 1 Schematic diagram of the odor compound database and retrieval system architecture of the present invention

[0030] Appendix Figure 2 Workflow diagram of the odor compound database and retrieval system of the present invention

[0031] Appendix Figure 3 Initialization loading and operation diagram of the odor compound database and retrieval system of the present invention

[0032] Appendix Figure 4 Example 1: Odor Compound Database and Retrieval System Web-based Search Input Image

[0033] Appendix Figure 5 Analysis process of odor compound database and retrieval system in Example 1

[0034] Appendix Figure 6 Output image of the web-based results of the odor compound database and retrieval system in Example 1.

[0035] Appendix Figure 7 Output image of the web-based results from the odor compound database and retrieval system in Example 2.

[0036] Appendix Figure 8 Analysis process of odor compound database and retrieval system in Example 3

[0037] Appendix Figure 9 Input and output diagrams of the web-based input and output of the odor compound database and retrieval system in Example 3.

[0038] Appendix Figure 10 Example 4: Output image of the odor compound database and retrieval system web page. Detailed Implementation

[0039] Example 1

[0040] Taking target sample No. 1 as an example, raw data is collected, its volatile compounds are detected, the results are integrated, and a test report is output. A PDF test report containing information such as compound name, peak number, molecular formula, retention time, retention index, peak area, and CAS number is obtained and stored in the data terminal.

[0041] Run the odor database and its screening and retrieval system to find odor compounds in sample 1-1, as shown in the figure:

[0042] Step 1: Build an odor database. The system automatically extracts the CAS numbers and corresponding odor descriptions from the "odor query.pdf" file by preloading an odor substance database (Excel format) and using a PDF parser. A caching mechanism is then employed to accelerate database initialization (the data is stored as "odor query.cache").

[0043] Step Two: Parsing the Volatile Compound Detection Report. The user-uploaded PDF report is parsed using a PDF parser, employing a three-stage text processing workflow. Peak area data is extracted from the "Peak Report TIC" area, with integers of 10,000 considered valid peak areas and filtered according to rules. Key compound fields (Formula / CAS number / RetIndex) are captured using regular expressions. Finally, a data input table is created and associated with the peak areas.

[0044] Step 3: Retrieve Results Using a Dual-Database Matching Mechanism. An input data table is constructed by storing information such as molecular formulas, CAS numbers, and retention indices extracted from the PDF results report. Peak areas and row indexes are added to the input table to form standardized input data. The primary matching joins the input data within the odor database, outputting odor compound results. The secondary matching joins unmatched data with the extended odor database, appending new compound information to the extended odor database. The result is exported as a writable version of the odor substance database search (.xlsx), enabling dynamic database updates.

[0045] Step 4: Output results via web interaction. Search for odor compound results via a webpage, supporting PDF file uploads and result table visualization. Data retrieval results, including CAS number, chemical formula, compound name, RetIndex, threshold, area, and odor characteristics, are converted into a list format and transmitted to the webpage for tabular display.

[0046] The specific operation process is as follows: Place the test report in the specified location within the Python package folder, i.e., the storage location for the volatile compound test report. Run the software written with the PyCharm compiler, access the local search platform web interface, and enter the original data name: 1.pdf (see attachment). Figure 4 Click "Start Search" to run the program, parse the test report, and filter and match it against the odor compound database, as shown in the attached file. Figure 5 As shown; obtain the odor substance information from the test report and output it from the webpage, see attachment. Figure 6 .

[0047] In the operation of Example 1, the operator only needs to import the search report into the designated storage location, run the program, enter the name of the target report through the web interface, and click "Start Search" to obtain the search results. The estimated time for this process is shown in the attached figure. Figure 5 As shown, it only takes 207.94 milliseconds. Using this odor compound database and retrieval system to screen odor compounds in volatile compound test reports is time-efficient, simple to operate, improves retrieval efficiency, and significantly saves manual labor costs.

[0048] Example 2

[0049] Taking target sample No. 2 as an example, the specific operation steps and processes are the same as in Example 1 above. The odor substance information in the detection report is obtained and output from the webpage, as shown in the appendix. Figure 7 In Example 2, as shown in the attached... Figure 7 As shown, two unidentifiable odor compounds were found in the output search results. After recording the CAS number feature value, the detailed information of the odor compound was obtained through manual search and added to the odor compound database to achieve continuous database updates.

[0050] Example 3

[0051] Taking target sample No. 3 as an example, the specific operation steps and process are the same as in Example 1 above. Enter the name of the target report on the webpage and click "Start Query." The odorant information in the detection report will be output from the webpage. (See attached image.) Figure 8 The estimated time for this process is shown in the attached figure. Figure 9 As shown, it only takes 722.52 milliseconds. Using this odor compound database and retrieval system to screen odor compounds in volatile compound test reports is time-efficient, simple to operate, improves retrieval efficiency, and significantly saves manual labor costs.

[0052] Example 4

[0053] Taking target sample No. 4 as an example, the specific operation steps and processes are the same as in Example 1 above. The odor substance information in the detection report is obtained and output from the webpage, as shown in the appendix. Figure 10 In Example 4, as shown in the attached document... Figure 10 As shown, three unidentifiable odor compounds were found in the output search results. After recording the CAS number feature value, detailed information of the odor compound was obtained through manual search and added to the odor compound database to achieve continuous database updates.

[0054] Obviously, the above embodiments are merely examples for clear illustration and are not intended to limit the implementation. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A database of odor compounds and a rapid screening method developed therein, along with the establishment of its software system, characterized in that, Includes the following steps: Step 1: Build an odor database. Extract the CAS number and corresponding odor description from the odor query file by preloading the odor substance database and PDF parser, and use a caching mechanism to speed up database initialization. Step 2: Analyze the volatile compound test report, extract peak area data and key compound information, and establish a data input table; Step 3: Perform primary and secondary matching through a dual-database matching mechanism, output odor compound results, and dynamically update the database; Step 4: Output the search results through the web interface, supporting PDF file upload and result visualization.

2. The method according to claim 1, characterized in that, The odor substance database mentioned in step one is in Excel format, and the PDF parser is used to parse the "odor query.pdf" file.

3. The method according to claim 1, characterized in that, In step two, the PyMuPDF engine is used to parse the PDF report, extract the compound molecular formula, CAS number, retention index information through regular expressions, and associate it with peak area data.

4. The method according to claim 1, characterized in that, The dual-database matching in step three includes primary matching, which connects the input data to the odor database and outputs odor compounds; and secondary matching, which connects unmatched data to the extended odor database, supplements new compound information, and updates the primary database.

5. The method according to claim 1, characterized in that, The search results in step four are displayed in tabular form on the webpage, including CAS number, chemical formula, compound name, retention index, threshold, peak area, and odor characteristics.

6. An odor compound database and rapid screening system for implementing the method according to any one of claims 1 to 5, characterized in that, include: The database building module is used to preload and initialize the odor database; The report parsing module is used to extract compound and peak area information from volatile compound detection reports; The matching and retrieval module is used to perform dual database matching and output odor compound results; the web interaction module is used to receive user-uploaded files and display search results.

7. The system according to claim 6, characterized in that, The matching and retrieval module supports dynamic updates and export of the database.

8. The system according to claim 6, characterized in that, The web interaction module is implemented based on the Flask framework and supports front-end table visualization and user interaction.

Images