Method and system for automatically extracting chemical data from literature

Abstract

The invention discloses a method and a system for automatically extracting chemical data from literatures. The method comprises the following steps of: converting the input literatures into a plain text format; establishing a paragraph identifier for the target article according to a predefined paragraph tag; establishing a content identifier for the target paragraph according to a predefined content label; performing chemical data identification positioning and marking on the target paragraph content by using a preset regular expression; extracting and outputting chemical data according to an identification result; the system comprises an input module, a paragraph marking module, an internal marking module, a chemical data identification positioning and marking module and an extraction and output module. According to the method, the chemical data can be automatically identified and extracted from the literature, the extracted chemical data is comprehensive, the extracted data can form a corresponding chemical database to realize electronic service of the chemical literature, comprehensive information support can be provided for chemical research and development, and the method is of great significance to chemical research and development.

Description

Technical Field

[0001] The invention relates to a method and system for automatically extracting chemical data from documents, belonging to the technical field of information processing. Background Art

[0002] The role of scientific experiments in the development of natural sciences is becoming more and more obvious. As one of the main outputs of scientific experiments, scientific data has become an important resource that has received more and more attention. Efficient and structured data extraction has always attracted much attention. As one of the three basic disciplines of natural sciences, the research and application of chemistry has penetrated into all aspects of scientific research, national economy, social development, and national security, forming many cross-fields and accumulating a large amount of valuable information and data. Constructing a chemical database can provide comprehensive information services for chemical and chemical research and development. Therefore, the demand for chemical data has prompted the study of chemical data extraction technology in literature.

[0003] Although the collection and processing of chemical data has a history of more than 200 years in the world, before 2010, the collection and processing of chemical data was still mainly done manually. Since 2011, Jessop, Hawizy and D. M. Lowe of Cambridge University have successively published chemical text mining systems for automatic identification of compound names, which has opened up a new direction for the use of artificial intelligence methods in chemical data processing. However, at present, the main method is to automatically extract chemical data from literature using machine learning methods, and the use of machine learning methods requires manual indexing of a large number of similar corpora to train the model, which has high requirements for manpower and hardware; in addition, the current research on the automatic identification and extraction methods of various types of data in journal literature is mainly concentrated on English literature, and the current research on the extraction of Chinese literature is still limited to the research on chemical entities (compound names and structures) in the literature. In addition to chemical entities (compound names and structures), chemical data in chemical literature usually also includes product characteristics, synthesis reaction parameters, physical property data, spectral data and other scientific data; therefore, how to achieve low-cost and high-accuracy extraction of comprehensive chemical data in chemical literature has become a technical problem that is currently urgently needed in this field. Summary of the invention

[0004] In view of the above problems and needs in the prior art, the purpose of the present invention is to provide a method and system for automatically extracting chemical data from documents, so as to achieve low-cost and high-accuracy extraction of comprehensive chemical data in documents.

[0005] In order to achieve the above-mentioned invention object, the present invention adopts the following technical scheme:

[0006] A method for automatically extracting chemical data from literature comprises the following steps:

[0007] S1) Convert the input document into plain text format;

[0008] S2) establishing paragraph tags for the target article according to predefined paragraph tags;

[0009] S3) establishing a content tag for the target paragraph according to a predefined content tag;

[0010] S4) using a preset regular expression to identify, locate and mark the chemical data of the target paragraph content;

[0011] S5) Extracting and outputting chemical data according to the identification result of step S4).

[0012] In one embodiment, the document mentioned in step S1) may be a Chinese document or / and an English document.

[0013] In one embodiment, the document in step S1) is input in PDF or Word format.

[0014] In one implementation scheme, the paragraph tags in step S2) are predefined according to XML tag design rules.

[0015] In one implementation scheme, the content of the paragraph tag in step S2) is composed of the sc start element and Chinese characters describing the paragraph type.

[0016] According to a further implementation scheme, the paragraph types are divided into general paragraph types and chemical paragraph types. The general paragraph type refers to whether a paragraph is a title, abstract, subheading, main text, figure or reference. The chemical paragraph type is used to further classify subheadings and main text paragraphs according to whether they involve target chemical content. Specifically, it can be defined as experimental part, instruments and reagents, synthesis methods, products and properties, results discussion part, etc.

[0017] In one implementation scheme, the content tags in step S3) are predefined according to XML tag design rules.

[0018] In one implementation scheme, the content of the content tag in step S3) is composed of a cc start element and Chinese characters describing the content type.

[0019] In a further implementation scheme, the content type refers to the content classification of a certain paragraph of text, which can be specifically defined as instrument content, reagent content, synthesis operation content, product property content, etc.

[0020] In one implementation scheme, the regular expression in step S4) is manually constructed based on the language expression pattern of chemical data in chemical literature.

[0021] In one implementation scheme, in step S4), data tags for chemical data tagging are predefined according to XML tag design rules.

[0022] In a further implementation scheme, the data tag content is composed of a DC starting element and Chinese characters describing the chemical data type.

[0023] In a further embodiment, the types of chemical data are divided into product characteristics, synthesis reaction parameters, physical property data and spectral data.

[0024] In a further implementation plan, the product characteristics are divided into product name, product code, CAS number, molecular formula, molecular weight and elemental analysis, the synthesis reaction parameters are divided into reaction temperature, reaction time, purity, yield and separation method, the physical property data are divided into melting point, boiling point, properties, solubility, optical rotation, crystal parameters and magnetic susceptibility, and the spectral data are divided into infrared spectrum, ultraviolet spectrum, fluorescence spectrum, hydrogen nuclear magnetic spectrum, mass spectrum, carbon 13 nuclear magnetic spectrum, phosphorus 31 nuclear magnetic spectrum, fluorine 19 nuclear magnetic spectrum, boron 11 nuclear magnetic spectrum, silicon 29 nuclear magnetic spectrum, nitrogen 15 nuclear magnetic spectrum, hydrogen heteronuclear polycarbon correlation spectrum, Mössbauer spectrum and circular dichroism spectrum.

[0025] In one implementation, step S5) uses natural language processing (NLP) methods to extract chemical data.

[0026] A system for automatically extracting chemical data from literature, comprising:

[0027] An input module for inputting documents and converting the input documents into a plain text format;

[0028] A paragraph marking module is used to create paragraph tags for the target article according to predefined paragraph tags;

[0029] A content tagging module is used to create a content tag for a target paragraph according to a predefined content tag;

[0030] A chemical data identification, positioning and marking module is used to identify, position and mark chemical data of a target paragraph using a preset regular expression;

[0031] The extraction and output module is used to extract chemical data and output the extraction results.

[0032] In addition, the present invention also provides a storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by electronic devices (including but not limited to computers, servers, or network devices, etc.) to execute the above-mentioned method of automatically extracting chemical data from literature of the present invention.

[0033] The present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned method of automatically extracting chemical data from literature of the present invention.

[0034] The present invention also provides a computer program product, which includes a computer program stored on a storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer is enabled to execute the above-mentioned method of automatically extracting chemical data from literature of the present invention.

[0035] Compared with the prior art, the beneficial technical effects of the present invention are:

[0036] The present invention is based on language expression patterns and natural language processing (NLP) methods, which not only realizes automatic recognition and extraction of chemical data from literature, but also extracts comprehensive chemical data, including more than 30 kinds of chemical data such as product characteristics, synthesis reaction parameters, physical property data, spectral data, etc., so that the extracted data can be built into a corresponding chemical database to realize the electronic service of chemical literature, and can provide comprehensive information support for chemical research and development; therefore, the present invention has significant application value and is of great significance to chemical research and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A flow chart of a method for automatically extracting chemical data from literature provided by the present invention.

[0038] Figure 2 A system structure diagram for automatically extracting chemical data from literature provided by the present invention. DETAILED DESCRIPTION

[0039] The technical solution of the present invention is further described in detail and completely below in conjunction with the embodiments.

[0040] like Figure 1 As shown, the present invention provides a method for automatically extracting chemical data from documents, comprising the following steps:

[0041] S1) Convert the input document into plain text format;

[0042] S2) establishing paragraph tags for the target article according to predefined paragraph tags;

[0043] S3) establishing a content tag for the target paragraph according to a predefined content tag;

[0044] S4) Identify, locate, and tag chemical data in the target paragraph content using a preset regular expression;

[0045] S5) Extract and output chemical data based on the identification result in step S4).

[0046] The literature described in step S1) can be Chinese literature or / and English literature. The literature described in step S1) is input in PDF or Word format, and the input PDF or Word format literature can be converted into plain text format through an XML editor.

[0047] The paragraph tags described in step S2) are predefined according to XML tag design rules. The content of the paragraph tags consists of the sc start element and Chinese describing the paragraph type. The paragraph types are divided into general paragraph types and chemical paragraph types. The general paragraph types refer to whether a paragraph is a title, abstract, subheading, body text, chart, or reference. The chemical paragraph types are used to further classify subheadings and body text paragraphs according to whether they involve the target chemical content, and can be specifically defined as experimental part, instruments and reagents, synthesis methods, products and properties, result discussion part, etc. For example: <sc synthesis method title> < / sc synthesis method title>; <sc product and properties> < / sc product and properties>.

[0048] The content tags described in step S3) are predefined according to XML tag design rules. The content of the content tags consists of the cc start element and Chinese describing the content type. The content type refers to the content classification of a certain piece of text, and can be specifically defined as instrument content, reagent content, synthesis operation content, product property content, etc. For example: <cc synthesis operation content> < / cc synthesis operation content>; <cc product property> < / cc product property>.

[0049] The regular expression described in step S4) is constructed manually based on the language expression patterns of chemical data in chemical literature.

[0050] The inventors of this application summarized the language expression characteristics in chemical literature and generalized their language expression patterns into the following 3 general formulas:

[0051] General formula 1: <ep numerical range / ep> <ep physical unit 1 / ep> ( / <ep physical unit 2 / ep>|·<ep physical unit 2 / ep>)?

[0052] General formula 2: <ep start identifier / ep> <ep data part / ep> <ep end identifier

[0053] General formula 3: (.*<ep keyword / ep>.*)+;

[0054] In General Formula 1, | represents "or optional",? represents that this part can appear 0 or 1 time, and Physical Unit 1 and Physical Unit 2 therein represent specific physical units; in General Formula 3, represents any character, * represents 0 to n times, and + represents 1 to n times.

[0055] The data meaning of General Formula 1 can be determined by the scenario or defined by the author in words. For example:

[0056] Regular expression for the numerical range of scientific data <ep numerical range / ep>:

[0057] [<><>≤≥]?[+-]?\(?(\d+(\.\d+)?) (±(\d+(\.\d+)?) \)? | ([-~、和](\d+(\.\d+)?) )+)?) \s?;

[0058] Language expression pattern for temperature: <ep numerical range / ep> \s? <ep temperature unit / ep>

[0059] For example: m.p. 280 °C, the melt-cast explosive carrier itself has a suitable melting point of 80 - 100 °C;

[0060] Language expression pattern for density: <ep numerical range A / ep> \s? <ep mass unit / ep> ( / <ep length unit / ep>3 | ·]<ep length unit / ep>-3)

[0061] For example: density 1.844 g / cm 3 and the densities of Compounds 2 and 3 are 1.87 g·cm -3 and 1.75 g·cm -3 .

[0062] The data meaning of General Formula 2 has no scenario restrictions and is completely determined by the starting identifier. All scientific data of the General Formula 2 type have their specific starting identifiers. If the ending part of the remaining part after removing the starting identifier can also be used as the ending identifier of this data. For example:

[0063] Language expression pattern for optical rotation:

[0064] [α](\d* \s)? D? <ep data part / ep> <ep test condition / ep>?

[0065] For example: [α] 25 D-17.4 (c 1.0, CHCl 3 );

[0066] Language expression pattern for infrared spectrum:

[0067] (\bATR-FTIR\b|\bIR\b|\bFTIR\b|\bF[Tt]-?IR\b).*?\b[vν](max)?:?\s<ep data part / ep>+\scm\-1

[0068] For example: IR(KBr) ν: 3331, 3280, 1679, 1635, 1552 cm -1 。

[0069] The data of General Formula 3 are text data, which usually need to be defined by multiple keywords (groups). The meaning of the data is completely determined by the scenario. These keywords can be divided into three types according to different matching requirements: required keywords, alternative keywords, and optional keywords. Among them, required keywords are the keywords that must be included in the hit results. Alternative keywords mean that one of multiple keywords must be selected, while optional keywords mean that these keywords may or may not be in the hit results. Required keywords and alternative keywords are the core of a language expression pattern and also a necessary condition for a string to become a hit result. The role of optional keywords is to help determine the front and back boundaries of the language expression pattern. It should be noted that some words without independent meaning in alternative keywords, as well as the combinations of optional keywords and words without independent meaning in alternative keywords, may become noises interfering with automatic recognition and need to be excluded in subsequent operations. For example:

[0070] 1) Colorless oily liquid, yellow solid;

[0071] The Chinese regular expression for the above morphology is:

[0072] ([\u4e00-\u9fa5])*(color|transparent|color transparent))? Keyword 1 / Optional

[0073] (\u4e00-\u9fa5])*(crystal|foam|crystal [shape]|acicular [crystal]|crystalline [shape]|liquid|solid [state]|oil|substance|powder|viscous|paste|syrup|powdered|flocculent|flaky|waxy|liquid-solid mixture|suspension|mucus|serum|fluid) shape? Keyword 2 / Alternative

[0074] ([\u4e00-\u9fa5]*(crystal|acicular crystal|crystal|liquid|solid|oil|substance|powder|serum|fluid)(substance)?)? Keyword 3 / Optional

[0075] 2) Colorless needle crystal;

[0076] The English regular expression for the above morphology is:

[0077] [\w\s]*(\s[Crystals?\b|\s[Crystalline\b|\s[Foam\b|\s[Gel\b|\s[Gum\b|\s[Granular\b|\s[Liquid\b|\s[Mucilaginous\b|\s[Needle\b|\s[Oil\b|\s[Powder\b|\s[Semi -?solid\b|\s[Solid\b|\s[Syrup\b) Keyword 1 / or optional

[0078] (\s[Crystals?\b|\s[Crystalline\b|\s[Gel\b|\s[Gum\b|\s[Granular\b|\s[Liquid\b|\s[Needle\b|\s[Oil\b|\s[Powder\b|\s[Solid\b)? Keyword 2 / optional

[0079] (with\s.*?([;,\.]?(\s|$))? Keyword 3 / optional

[0080] 3) Regarding the language expression pattern of Chinese compounds:

[0081] (unsaturated)?(\[[^\]]+\]-)?((([\{\[\(]*(<ep substitution position part / ep>|<ep multiple prefix / ep>|<ep Chinese substitution position / ep>|<ep English substitution position / ep>))+\-)*(unsaturated)?([\{\[\(<]*(<ep substitution group part / ep>| reagent)[\)\]\}>]*)+(\-((?!DdTt]\s)[DdTt)?|·)?)+(<ep multiple prefix / ep>?<ep chemical property word / ep>)?(\s?\(\d+(:\d+)+\))?

[0082] Among them:

[0083] <ep substitution position part / ep>:

[0084] <ep substitution position / ep>?(<ep Greek letter / ep>|<ep element substituent / ep>[‘’~]*?|..

[0085] <ep multiple prefix / ep>:

[0086] [one two three four five six seven eight nine ten half times single whole double geminal oligomeric]|even(?!con)

[0087] <ep Chinese replacement position / ep>:

[0088] [Primary, secondary, tertiary, quaternary, adjacent, opposite, average, partial, complete, terminal]|meta (?! body)

[0089] <ep Substituent group part / ep>:

[0090] <ep Multiple numeral prefix / ep>(\[[^\]+\]){0,3}|(<ep Chinese stereochemical symbol / ep>|<ep Chinese replacement position / ep>|<ep Chinese group prefix / ep>)\-?|...

[0091] <ep Chemical property word / ep>:

[0092] (Complex (compound|ligand)|Complex [salt]|([Hydration, mixing, complexation, polymerization, conjugation] compound|Substituted|Derivative)|[Anion, cation, onium]? ion|…

[0093] Examples of matches are: 3,3 - bis(azidomethyl) - oxetane, 3,6,7 - triamino - 7H - [1,2,4] - triazolo[4,3 - b][1,2,4]triazole, 3,3 - bis(azidomethyl)oxetane - tetrahydrofuran copolymer, hexanitrohexaazaisowurtzitane, 3,4 - dihydropyrimidin - 2 - one derivative;

[0094] It should be noted here that: the sub - pattern <ep Chemical property word / ep> is an optional keyword in the Chinese - language expression pattern of compounds. Among them, the optional word "derivative" moves the end boundary of the name from ketone to the following "compound" in the recognition of the compound name in the subtitle "Synthesis of 3,4 - dihydropyrimidin - 2 - one derivatives".

[0095] In addition, taking the English - language expression pattern of product data as an example, according to the different voices of keywords, the English - language expression pattern of product data can be divided into two types: active mode and passive mode. For example:

[0096] English - language expression pattern of product data (active mode):

[0097] <ep Obtain - like verb / ep><ep Field without excluded words / ep><ep Paragraph or sentence end symbol / ep>

[0098] Among them:

[0099] <ep Field without excluded words / ep>:

[0100] ((?!<ep Excluded entity and action - type nouns / ep>|<ep Obtain - like verb / ep>).)*?

[0101] Regular expression corresponding to the sub - expression pattern:

[0102] <ep Verbs related to obtaining / ep>:

[0103] ((\bgives?\b|\bafford(ed)?\b|\bfurnish(ed)?\b|\bgave\b|\bprovided?\b|\bobtain\b|\bto\sget\b|\bto\sproduce\b|\bresult(ing|ed)\sin\b|\bsupply\b|\byielded\b)\s

[0104] <ep Field without excluded words / ep>:

[0105] ((?!\bno\sproducts?|\sm\.p\.\s?|\sIR\s|\bNMR\s|\b(precipitates?|cakes?|solutions?|added)\b|<ep Verbs related to obtaining / ep>.)*?

[0106] <ep Paragraph or sentence end symbol / ep>:

[0107] (\.\s|\.$)

[0108] For example:

[0109] ...to give colorless oil methyl (2E,6R)-3-phenyl-7-methyl-6,7-dihydroxy-octa-2-enoate (16a) 382 mg, yield 78%.......to afford 32.9 mg (79%)of 4,5-dimethoxy-2-nitrobenzonitrile (2a) as a yellow solid, m.p. 165~167℃。

[0110] Language expression pattern of English product data (passive pattern):

[0111] <ep Sentence start symbol or semantic start word / ep><ep Field 1 without excluded words / ep>\s(waslwere|be)\s([a-z]+\s)?\b<ep Passive form of verb related to obtaining / ep><ep Field 2 without excluded words / ep><ep Paragraph or sentence end symbol / ep>

[0112] The regular expression corresponding to the sub-expression pattern is as follows:

[0113] <ep Sentence start symbol or semantic start word / ep>:

[0114] (\. \s|^|According\sto\s|\sthen\s)

[0115] <ep field 1 without excluded words / ep>:

[0116] ((?![\.:]\s|\bthen\b|\b[Ss]olutions?\b).)*?

[0117] <ep passive form of the word "obtained" / ep>:

[0118] \b(afforded|furnished|gotlobtained|prepared|provided|resulted\sin|synthesized|yielded)\b

[0119] <ep field 2 without excluded words / ep>:

[0120] (?!\b[Ss]olutions?|added)\b|<ep passive form of the word "obtained" / ep>).)*?

[0121] <ep paragraph or sentence end symbol / ep>:

[0122] (\. \s|$)

[0123] For example:

[0124] .... After drying over anhydrous sodium sulfate, purification byrecrystallized with ethanol, white solid 4-(4-(diethylamino)phenyl)-7,7-dimethyl-5,6,7,8-tetrahy-dro-6,8-methanoquinazolin-2-amine (3) (6.67 g, yield78.5%) was obtained.

[0125] ....By subjecting 5 to the reaction system of 2a and catalytic amount of CuBr, in CH, CN, expectedly, the desired 3,3'-biindole oxindole 6 was afforded in 54% yield, highlighting the formation of cation intermediate (Scheme 5).

[0126] Step S4) Pre-define the data tags for chemical data tagging according to the XML tag design rules. The content of the data tag consists of the starting element "dc" and Chinese characters describing the type of chemical data. The types of the chemical data are divided into product characteristics, synthesis reaction parameters, physical property data, and spectroscopic data. The product characteristics are divided into product name, product code, CAS number, molecular formula, molecular weight, and elemental analysis. The synthesis reaction parameters are divided into reaction temperature, reaction time, purity, yield, and separation method. The physical property data are divided into melting point, boiling point, property, solubility, optical rotation, crystal parameters, and magnetic susceptibility. The spectroscopic data are divided into infrared spectrum, ultraviolet spectrum, fluorescence spectrum, hydrogen nuclear magnetic spectrum, mass spectrum, carbon-13 nuclear magnetic spectrum, phosphorus-31 nuclear magnetic spectrum, fluorine-19 nuclear magnetic spectrum, boron-11 nuclear magnetic spectrum, silicon-29 nuclear magnetic spectrum, nitrogen-15 nuclear magnetic spectrum, heteronuclear multiple carbon correlation spectrum of hydrogen, Mössbauer spectrum, and circular dichroism spectrum; for example:

[0127] <dc:product name> < / dc:product name>; <dc:product code> < / product code>; <dc:yield> < / dc:yield>; <dc:hydrogen spectrum> < / dc:hydrogen spectrum>; <dc:carbon spectrum> < / dc:carbon spectrum>.

[0128] Step S5) Use the natural language processing (NLP) method to extract chemical data. The extraction program can extract data according to the data identifier. Utilizing the feature that the entities and their property data in the experimental part and the result discussion part of a chemical research article are in the same article paragraph, it is considered that all the property data extracted from this paragraph belong to the entity identified as the substance name in this paragraph. When there is no entity identified as the substance name in this paragraph, the data extraction program can extract the entity identified as the substance name from the adjacent subheading paragraph according to the corresponding rules as the entity of these property data.

[0129] Example 1

[0130] Suppose the original text is as follows:

[0131] 3.2.2 Synthesis of 5-hydroxyamyl 4-methylbenzenesulfonate (4)

[0132] 1,5 - Pentanediol (41.70 g, 0.4 mol), DMAP (1.47 g, 12 mmol) were dissolved in 300 mL of methyl tert - butyl ether. Triethylamine (31 mL, 0.22 mol) was added, and a dichloromethane solution of 4 - methylbenzenesulfonyl chloride (38.13 g, 0.2 mol, dissolved in 150 mL of dichloromethane) was slowly added dropwise under an ice bath. After the addition was complete, the temperature was raised to 20 °C, and the mixture was stirred for 18 h. Then 50 mL of water was added to quench the reaction, and it was neutralized with 1 mol / L dilute hydrochloric acid. The organic phase was washed with water, saturated sodium chloride solution, dried over anhydrous sodium sulfate, and the solvent was removed to obtain 40 g (155 mmol) of a colorless oily liquid with a yield of 77.4%. 1 H NMR(400 MHz, CDCl 3 ) δ : 7.78(d,J = 8.3 Hz, 1H), 7.35(d, J = 8.2 Hz, 1H), 4.03(t, J = 6.4 Hz, 1H), 3.61(t, J = 6.4Hz, 1H), 2.90(s, 1H), 2.45(s, 1H), 1.72~1.63(m, 1H), 1.51(dd, J = 14.4, 6.6Hz, 1H), 1.40~1.37(m, 1H); 13 C NMR (100 MHz, CDCl 3 )δ:144.86, 133.14, 129.94,127.95, 70.56, 62.56, 31.84, 28.67, 21.76; ESI - HRMS calcd for C 12 H 19 O 4 S [M + H] + 259.0999, found 259.1005.

[0133] The identification results after chemical data recognition, location, and tagging using the method of the present invention are as follows:

[0134] <sc Synthesis method title><cc Synthesis operation title>3.2.2<dc Synthesis method title><dc Substance name>5 - Hydroxypentyl 4 - methylbenzenesulfonate< / dc Substance name>(<dc Substance code>4< / dc Substance code>) Synthesis< / dc Synthesis method title>< / cc Synthesis operation title>< / sc Synthesis method title>

[0135] <sc Product and properties><cc Synthesis operation content><dc Synthesis method>1,5-Pentanediol (41.70 g, 0.4 mol), DMAP (1.47 g, 12 mmol) were dissolved in 300 mL of methyl tert-butyl ether. Triethylamine (31 mL, 0.22 mol) was added, and a dichloromethane solution of 4-methylbenzenesulfonyl chloride (38.13 g, 0.2 mol, dissolved in 150 mL of dichloromethane) was slowly added dropwise under an ice bath. After addition, the temperature was raised to 20 °C, and after stirring for 18 h, 50 mL of water was added to quench the reaction. It was neutralized with 1 mol / L dilute hydrochloric acid, the organic phase was washed with water, washed with saturated sodium chloride, dried over anhydrous sodium sulfate, and the solvent was removed. <dc Product data>40 g (155 mmol) of a <dc Form>colorless oily liquid< / dc Form> was obtained, and the <dc Yield>yield was 77.4%< / dc Yield>.< / dc Product data>< / dc Synthesis method>< / cc Synthesis operation content><cc Product property content><dc 1H NMR>1H NMR (400 MHz, CDCl3) δ: 7.78 (d, J = 8.3 Hz, 1H), 7.35 (d, J = 8.2 Hz, 1H), 4.03 (t, J = 6.4 Hz, 1H), 3.61 (t, J = 6.4 Hz, 1H), 2.90 (s, 1H), 2.45 (s, 1H), 1.72~1.63 (m, 1H), 1.51 (dd, J = 14.4, 6.6 Hz, 1H), 1.40~1.37 (m, 1H); < / dc 1H NMR> <dc 13C NMR>13C NMR (100 MHz, CDCl3) δ: 144.86, 133.14, 129.94, 127.95, 70.56, 62.56, 31.84, 28.67, 21.76; < / dc 13C NMR><dc ESI-HRMS>ESl-HRMS calcd for C12H19O4S [M + H]+ 259.0999, found 259.1005.< / dc ESI-HRMS>< / cc Product property content>< / sc Product and properties>

[0136] Example 2

[0137] Suppose the original text is as follows:

[0138] Synthesis of 1,4-bis(-2-(2-benzimidazolyl)vinyl)benzene (10): A mixture containing 20 mmol (2.68 g) of terephthalaldehyde, 45 mmol (5.94 g) of 2-methylbenzimidazole and 40 mL of acetic anhydride was heated under reflux for 6 h, then poured into 200 mL of water, allowed to stand, and filtered to obtain a crude yellow solid product. The crude product was purified by washing with ethanol multiple times to obtain the product

[39] , yield 25.3%, m.p. > 300 °C. MS-ESI m / z: 363.2[M+H] + .

[0139] The identification results after paragraph tagging are as follows:

[0140] <sc synthesis method>Synthesis of 1,4-bis(-2-(2-benzimidazolyl)vinyl)benzene (10): A mixture containing 20 mmol (2.68 g) of terephthalaldehyde, 45 mmol (5.94 g) of 2-methylbenzimidazole and 40 mL of acetic anhydride was heated under reflux for 6 h, then poured into 200 mL of water, allowed to stand, and filtered to obtain a crude yellow solid product. The crude product was purified by washing with ethanol several times to obtain the product

[39] , yield 25.3%, m.p. > 300 °C. MS-ESI m / z: 363.2[M+H]+.< / sc synthesis method>

[0141] The identification results after content tagging are as follows:

[0142] <sc synthesis method><cc synthesis operation content>Synthesis of 1,4-bis(-2-(2-benzimidazolyl)vinyl)benzene (10): A mixture containing 20 mmol (2.68 g) of terephthalaldehyde, 45 mmol (5.94 g) of 2-methylbenzimidazole and 40 mL of acetic anhydride was heated under reflux for 6 h, then poured into 200 mL of water, allowed to stand, and filtered to obtain a crude yellow solid product. The crude product was purified by washing with ethanol several times to obtain the product.< / cc synthesis operation content><cc product property content>MS-ESI m / z: 363.2[M+H]+.< / cc product property content>< / sc synthesis method>

[0143] The identification results after data identification, location and tagging are as follows:

[0144] <sc Synthesis method><cc Synthesis operation content><dc Substance name>The synthesis of 1,4-bis(-2-(2-benzimidazolyl)vinyl)benzene (<dc Substance code>10< / dc Substance code>): A mixture containing 20 mmol (2.68 g) of terephthalaldehyde, 45 mmol (5.94 g) of 2-methylbenzimidazole and 40 mL of acetic anhydride was heated under reflux for 6 h and then poured into 200 mL of water. After standing and filtering, a yellow solid crude product was obtained. The crude product was purified by washing with ethanol several times< / dc Synthesis method><dc Product data>The product

[39] was obtained, with a yield of <dc Yield>25.3%< / dc Yield> and a melting point of <dc Melting point>>300 °C< / dc Melting point>.< / dc Product data>< / cc Synthesis operation content><cc Product property content><dc Mass spectrum>MS-ESI m / z: 363.2 [M+H]+.< / dc Mass spectrum>< / cc Product property content>< / sc Synthesis method>

[0145] Example 3

[0146] Suppose the original text is as follows:

[0147] 3-(4-Methylphenyl)-imidazo[1, 2-α]pyridine (3a): White solid (yield 90%); m.p. 79~80 °C; 1 H NMR (400 MHz, CDCl 3 ) δ: 8.28 (d, J = 6.8 Hz, 1H), 7.66~7.63 (m, 2H), 7.43 (dd, J = 6.8, 1.6 Hz, 2H), 7.31 (d, J = 7.6 Hz, 2H), 7.18~7.14 (m, 1H), 6.77 (td, J = 6.8, 1.2 Hz, 1H), 2.42 (s, 3H); 13 C NMR (100 MHz, CDCl 3 ) δ: 146.0, 138.2, 132.3, 129.9, 128.0, 126.4, 125.8, 124.0, 123.4, 118.2, 112.4, 21.3; MS m / z: 209.1 [M+H] + .

[0148] The identification results after paragraph tagging are as follows:

[0149] <sc Product and properties>3-(4-Methylphenyl)-imidazo[1, 2-α pyridine( 3a ):White solid (yield 90%); m.p.79~80℃; 1 H NMR (400 MHz, CDCl 3 ) d

[0150] :8.28(d. J =6.8Hz, 1H), 7.66~7.63(m, 2H), 7.43(dd, J =6.8, 1.6Hz, 2H), 7.31(d, J =7.6Hz, 2H), 7.18~7.14(m, 1H), 6.77(td, J )=6.8, 1.2Hz, 1H), 2.42(s, 3H); 13 C NMR (100 MHz, CDCl 3 ) δ :146.0, 138.2, 132.3, 129.9, 128.0, 126.4, 125.8, 124.0,123.4, 118.2, 112.4, 21.3; MS m / z :209.1 [M+H] + .< / sc Product and properties>

[0151] The identification results after content tagging are as follows:

[0152] <sc Product and properties><cc Product property content>3-(4-Methylphenyl)-imidazo[1, 2- α pyridine( 3a ):White solid (yield 90%); m.p.79~80 ℃; 1 HNMR (400 MHz, CDCl 3 ) δ :8.28(d. J =6.8Hz, 1H), 7.66~7.63(m, 2H), 7.43(dd, J =6.8, 1.6Hz, 2H), 7.31(d, J =7.6Hz, 2H),7.18~7.14(m, 1H), 6.77(td, J )=6.8, 1.2Hz, 1H), 2.42(s, 3H); 1313C NMR (100 MHz, CDCl 3 ) δ : 146.0, 138.2, 132.3, 129.9, 128.0, 126.4, 125.8, 124.0, 123.4, 118.2, 112.4, 21.3; MS m / z : 209.1 [M+H] + .< / cc Product properties content>< / sc Product and properties>

[0153] The identification results after data identification, location, and tagging are as follows:

[0154] <sc Product and properties><cc Product properties content><dc Substance name>3-(4-Methylphenyl)-imidazo[1, 2- α pyridine< / dc Substance name>( <dc Substance Code>3a< / dc Substance Code> ): <dc Product data><dc Form>White solid< / dc Form>(yield <dc Yield>90%< / dc Yield>); m.p. <dc Melting point>79 - 80 °C< / dc Melting point>; <dc 1H NMR spectrum> 1 1H NMR (400 MHz, CDCl 3 ) δ : 8.28 (d, J J = 6.8 Hz, 1H), 7.66 - 7.63 (m, 2H), 7.43 (dd, J J = 6.8, 1.6 Hz, 2H), 7.31 (d, J J = 7.6 Hz, 2H), 7.18 - 7.14 (m, 1H), 6.77 (td, J J) = 6.8, 1.2 Hz, 1H), 2.42 (s, 3H); < / dc 1H NMR spectrum><dc 13C NMR spectrum> 13 13C NMR (100 MHz, CDCl 3 ) δ : 146.0, 138.2, 132.3, 129.9, 128.0, 126.4, 125.8, 124.0, 123.4, 118.2, 112.4, 21.3; < / dc 13C NMR spectrum><dc Mass spectrum>MS m / z : 209.1 [M+H] + .< / dc Mass spectrum>< / dc Product data>< / cc Product properties content>< / sc Product and properties>

[0155] In addition, if Figure 2 As shown, the present invention provides a system for automatically extracting chemical data from documents, comprising:

[0156] An input module for inputting documents and converting the input documents into a plain text format;

[0157] A paragraph marking module is used to create paragraph tags for the target article according to predefined paragraph tags;

[0158] A content tagging module is used to create a content tag for a target paragraph according to a predefined content tag;

[0159] A chemical data identification, positioning and marking module is used to identify, position and mark chemical data of a target paragraph using a preset regular expression;

[0160] The extraction and output module is used to extract chemical data and output the extraction results.

[0161] The above-mentioned program modules generally include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types.

[0162] The present invention also provides a storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by electronic devices (including but not limited to computers, servers, or network devices, etc.) to execute the above-mentioned method of automatically extracting chemical data from literature of the present invention.

[0163] The present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned method of automatically extracting chemical data from literature of the present invention.

[0164] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of a system for automatically extracting chemical data from documents, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the system for automatically extracting chemical data from documents via a network. The network includes, but is not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0165] The electronic device may exist in various forms, including but not limited to:

[0166] (1) Mobile communication devices: These devices are characterized by having mobile communication functions and providing voice and data communications as their main purpose. These terminals include: smart phones (such as iPhone), multimedia phones, feature phones, and low-end phones;

[0167] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers and have computing and processing capabilities. They also generally have mobile Internet access features. These terminals include PDAs, MIDs, and UMPC devices, such as the iPad.

[0168] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players (such as iPods), handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0169] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server has a similar architecture to general computers, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0170] (5) Other electronic devices with data interaction functions.

[0171] The present invention also provides a computer program product, the computer program product includes a computer program stored on a storage medium, the computer program includes program instructions, when the program instructions are executed by a computer, the computer can perform the method of automatically extracting chemical data from documents of the present invention. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments or some parts of the embodiments.

[0172] The inventor conducted a performance test on the method and system of the present invention using all 422 articles in the journal "Organic Chemistry" in 2022 as an example. The total time required was no more than 22 minutes on a ThinkPad laptop with a dual-core single CPU 1.6 GHz, 2 GB of memory, and running Windows XP. All 4,665 extracted product names were manually checked, and 58 problematic product names were found. The product name error rate was 1.24%, but the recognition accuracy of data in the text that conforms to the language expression pattern was close to 100%. Moreover, the method of the present invention does not require high computing power, and ordinary laptops and desktop computers can meet the computing power requirements. When encountering exceptions, you only need to modify the language expression pattern and the corresponding regular expression to complete the modification and improvement for the exceptions.

[0173] In summary, it can be seen that the present invention, based on language expression patterns and natural language processing (NLP) methods, not only realizes the automatic identification and extraction of chemical data from literature, but also the extracted chemical data is comprehensive, including more than 30 kinds of chemical data such as product characteristics, synthesis reaction parameters, physical property data, spectral data, etc., so that the extracted data can be built into a corresponding chemical database to realize the electronic service of chemical literature, which can provide comprehensive information support for chemical research and development, and has important significance and significant application value for chemical research and development.

[0174] Finally, it is necessary to point out here that the above is only a preferred specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. In the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other. In addition, for the sake of simplicity of description, the aforementioned method embodiments are all expressed as a combination of a series of actions, but those skilled in the art should be aware that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously; secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

Claims

1. A method for automatically extracting chemical data from literature, It is characterized in that The steps include: S1) Convert the input document into plain text format; S2) establishing paragraph tags for the target article according to predefined paragraph tags; S3) establishing a content tag for the target paragraph according to a predefined content tag; S4) using a preset regular expression to identify, locate and mark the chemical data of the target paragraph content; S5) Extracting and outputting chemical data according to the identification result of step S4).

2. The method according to claim 1, Features: The documents described in step S1) are Chinese documents and / or English documents, and the documents are input in PDF or Word format.

3. The method according to claim 1, Features: The paragraph tag in step S2) is predefined according to XML tag design rules, and the content of the paragraph tag is composed of the sc start element and Chinese characters describing the paragraph type.

4. The method according to claim 3, Features: The paragraph types are divided into general paragraph types and chemical paragraph types. The general paragraph type refers to whether a paragraph is a title, abstract, subheading, text, figure or reference. The chemical paragraph type is used to further classify and define subheadings and text paragraphs according to whether they involve target chemical content.

5. The method according to claim 1, Features: The content tag in step S3) is predefined according to XML tag design rules. The content of the content tag is composed of a cc start element and Chinese characters describing the content type. The content type refers to the content classification of a certain text.

6. The method according to claim 1, Features: The regular expression described in step S4) is constructed manually based on the language expression pattern of chemical data in chemical literature.

7. The method according to claim 1, Features: Step S4) predefines the data tags for chemical data according to XML tag design rules. The data tag content consists of a dc start element and Chinese characters describing the type of chemical data. The types of chemical data are divided into product characteristics, synthesis reaction parameters, physical property data and spectral data.

8. A system for automatically extracting chemical data from literature, It is characterized in that The system comprises: An input module for inputting documents and converting the input documents into a plain text format; A paragraph marking module is used to create paragraph tags for the target article according to predefined paragraph tags; A content tagging module is used to create a content tag for a target paragraph according to a predefined content tag; A chemical data identification, positioning and marking module is used to identify, position and mark chemical data of a target paragraph using a preset regular expression; The extraction and output module is used to extract chemical data and output the extraction results.

9. A storage medium having a computer program stored thereon, Features: When the program is executed by a processor, the method described in any one of claims 1 to 7 can be implemented.

10. An electronic device, wherein include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor; characterized in that the at least one processor is capable of executing the method described in any one of claims 1 to 7.

Images