openchemlib-utils - npm Package Compare versions

Comparing version 2.1.3 to 2.2.0

package.json

 {
   "name": "openchemlib-utils",
-  "version": "2.1.3",
+  "version": "2.2.0",
   "description": "Various utilities that extends openchemlib-js like HOSE codes or diastereotopic IDs",
@@ -33,13 +33,13 @@ "main": "lib/index.js",
   "devDependencies": {
-    "@babel/plugin-transform-modules-commonjs": "^7.19.6",
-    "@types/jest": "^29.2.4",
+    "@babel/plugin-transform-modules-commonjs": "^7.20.11",
+    "@types/jest": "^29.2.5",
     "cheminfo-build": "^1.1.11",
-    "eslint": "^8.29.0",
+    "eslint": "^8.32.0",
     "eslint-config-cheminfo": "^8.1.3",
-    "eslint-plugin-import": "^2.26.0",
+    "eslint-plugin-import": "^2.27.5",
     "jest": "^29.3.1",
     "jest-matcher-deep-close-to": "^3.0.2",
     "openchemlib": "8.0.0",
-    "prettier": "^2.8.1",
-    "rollup": "^3.7.4"
+    "prettier": "^2.8.3",
+    "rollup": "^3.10.0"
   },
@@ -46,0 +46,0 @@ "dependencies": {

src/db/MoleculesDB.js

@@ -7,3 +7,3 @@ import appendCSV from './utils/appendCSV';
 import pushMoleculeInfo from './utils/pushMoleculeInfo';
-import search from './utils/search';
+import { search, searchAsync } from './utils/search';
 /*
@@ -21,3 +21,3 @@ this.db is an object with properties 'oclID' that has as value
    *
-   * @param {OCL} [OCL] The openchemlib library
+   * @param {import('openchemlib')} OCL - openchemlib library
    * @param {object} [options={}]
@@ -120,2 +120,21 @@ * @param {boolean} [options.computeProperties=false]
   /**
+   * Search in a MoleculesDB
+   * Inside the database all the same molecules are group together
+   * @param {string|OCL.Molecule} [query] smiles, molfile, oclCode or instance of Molecule to look for
+   * @param {object} [options={}]
+   * @param {string} [options.format='idCode'] - query is in the format 'smiles', 'oclid' or 'molfile'
+   * @param {string} [options.mode='substructure'] - search by 'substructure', 'exact' or 'similarity'
+   * @param {boolean} [options.flattenResult=true] - The database group the data for the same product. This allows to flatten the result
+   * @param {boolean} [options.keepMolecule=false] - keep the OCL.Molecule object in the result
+   * @param {number} [options.limit=Number.MAX_SAFE_INTEGER] - maximal number of result
+   * @param {number} [options.interval=100] - interval in ms to call the onStep callback
+   * @param {function} [options.onStep] - callback to execute after each interval
+   * @param {AbortController} [options.controler] - callback to execute to check if the search should be aborted
+   * @return {Promise<Array>} array of object of the type {(molecule), idCode, data, properties}
+   */
+  searchAsync(query, options) {
+    return searchAsync(this, query, options);
+  }
+
+  /**
    * Returns an array with the current database
@@ -122,0 +141,0 @@ * @returns

src/db/utils/search.js

@@ -0,12 +1,13 @@
+import { noWait } from '../../util/noWait.js';
+
 import getMoleculeCreators from './getMoleculeCreators';
 
-export default function search(moleculesDB, query = '', options = {}) {
-  const {
-    format = 'idCode',
-    mode = 'substructure',
-    flattenResult = true,
-    keepMolecule = false,
-    limit = Number.MAX_SAFE_INTEGER,
-  } = options;
+class AbortError extends Error {
+  name = 'AbortError';
+  code = 20;
+}
 
+function getQuery(moleculesDB, query, options) {
+  const { format = 'idCode' } = options;
+
   if (typeof query === 'string') {
@@ -18,3 +19,10 @@ const moleculeCreators = getMoleculeCreators(moleculesDB.OCL.Molecule);
   }
+  return query;
+}
 
+export function search(moleculesDB, query = '', options = {}) {
+  const { mode = 'substructure' } = options;
+
+  query = getQuery(moleculesDB, query, options);
+
   let result;
@@ -34,5 +42,27 @@ switch (mode.toLowerCase()) {
   }
-  return processResult(result, { flattenResult, keepMolecule, limit });
+  return processResult(result, options);
 }
 
+export async function searchAsync(moleculesDB, query = '', options = {}) {
+  const { mode = 'substructure' } = options;
+
+  query = getQuery(moleculesDB, query, options);
+
+  let result;
+  switch (mode.toLowerCase()) {
+    case 'exact':
+      result = exactSearch(moleculesDB, query);
+      break;
+    case 'substructure':
+      result = await subStructureSearchAsync(moleculesDB, query, options);
+      break;
+    case 'similarity':
+      result = similaritySearch(moleculesDB, query);
+      break;
+    default:
+      throw new Error(`unknown search mode: ${options.mode}`);
+  }
+  return processResult(result, options);
+}
+
 function exactSearch(moleculesDB, query) {
@@ -46,3 +76,3 @@ const queryIDCode = query.getIDCode();
 
-function subStructureSearch(moleculesDB, query) {
+function substructureSearchBegin(moleculesDB, query) {
   let resetFragment = false;
@@ -60,6 +90,29 @@ if (!query.isFragment()) {
     }
-  } else {
+  }
+  return { resetFragment, queryMW, searchResult };
+}
+
+function substructureSearchEnd(searchResult, queryMW, resetFragment, query) {
+  searchResult.sort((a, b) => {
+    return (
+      Math.abs(queryMW - a.properties.mw) - Math.abs(queryMW - b.properties.mw)
+    );
+  });
+
+  if (resetFragment) {
+    query.setFragment(false);
+  }
+
+  return searchResult;
+}
+
+function subStructureSearch(moleculesDB, query) {
+  const { resetFragment, queryMW, searchResult } = substructureSearchBegin(
+    moleculesDB,
+    query,
+  );
+
+  if (searchResult.length === 0) {
     const queryIndex = query.getIndex();
     const searcher = moleculesDB.searcher;
-
     searcher.setFragment(query, queryIndex);
@@ -75,13 +128,51 @@ for (let idCode in moleculesDB.db) {
 
-  searchResult.sort((a, b) => {
-    return (
-      Math.abs(queryMW - a.properties.mw) - Math.abs(queryMW - b.properties.mw)
-    );
-  });
+  return substructureSearchEnd(searchResult, queryMW, resetFragment, query);
+}
 
-  if (resetFragment) {
-    query.setFragment(false);
+async function subStructureSearchAsync(moleculesDB, query, options = {}) {
+  const { interval = 100, onStep, controller } = options;
+  let shouldAbort = false;
+
+  if (controller) {
+    const abortEventListener = () => {
+      shouldAbort = true;
+    };
+    controller.signal.addEventListener('abort', abortEventListener);
   }
 
-  return searchResult;
+  const { resetFragment, queryMW, searchResult } = substructureSearchBegin(
+    moleculesDB,
+    query,
+  );
+
+  let begin = performance.now();
+
+  if (searchResult.length === 0) {
+    const queryIndex = query.getIndex();
+    const searcher = moleculesDB.searcher;
+    searcher.setFragment(query, queryIndex);
+    let index = 0;
+    let length = Object.keys(moleculesDB.db).length;
+    for (let idCode in moleculesDB.db) {
+      if (shouldAbort) {
+        throw new AbortError('Query aborted');
+      }
+      let entry = moleculesDB.db[idCode];
+      searcher.setMolecule(entry.molecule, entry.index);
+      if (searcher.isFragmentInMolecule()) {
+        searchResult.push(entry);
+      }
+      if ((onStep || controller) && performance.now() - begin >= interval) {
+        begin = performance.now();
+        if (onStep) {
+          onStep(index, length);
+        }
+        if (controller && !onStep) {
+          await noWait();
+        }
+      }
+      index++;
+    }
+  }
+  return substructureSearchEnd(searchResult, queryMW, resetFragment, query);
 }
@@ -88,0 +179,0 @@

src/diastereotopic/getDiastereotopicAtomIDsFromMolfile.js

@@ -7,3 +7,3 @@ import { getDiastereotopicAtomIDsAndH } from './getDiastereotopicAtomIDsAndH.js';
  * Pelase take care than numbering of atoms starts at 0 !
- * @param {object} OCL - openchemlib library
+ * @param {import('openchemlib')} OCL - openchemlib library
  * @param {string} molfile
@@ -10,0 +10,0 @@ * @returns

src/util/combineSmiles.js

@@ -7,3 +7,3 @@ const MAX_R = 10;
  * @param {array} [fragments] Array of {smiles,R1,R2,...}
- * @param {OCL} [OCL] The openchemlib library
+ * @param {import('openchemlib')} OCL - openchemlib library
  * @param {object} [options={}]
@@ -10,0 +10,0 @@ * @param {function} [options.onStep] method to execute each new molecules

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

153056

increased by5.6%

Number of package files

52

increased by1.96%

Lines of code

4303

increased by5.11%

No dependency changes