sdf-parser 7.0.2 → 7.0.4

package/lib/index.js

@@ -108,6 +108,7 @@ function getMolecule$1(sdfPart, labels, currentLabels, options) {
  * @param {object} [options.modifiers] - Object containing callbacks to apply on some specific fields
  * @param {boolean} [options.mixedEOL=false] - Set to true if you know there is a mixture between \r\n and \n
  * @param {string} [options.eol] - Specify the end of line character. Default will be the one found in the file
+ * @returns {object} - Object containing the molecules, the labels and the statistics
  */
 function parse(sdf, options = {}) {
   options = { ...options };
@@ -148,7 +149,7 @@ function parse(sdf, options = {}) {
 
   for (let i = 0; i < entriesBoundaries.length; i++) {
     let sdfPart = sdf.slice(...entriesBoundaries[i]);
-
+    if (sdfPart.length < 40) continue;
     let currentLabels = [];
     const molecule = getMolecule$1(sdfPart, labels, currentLabels, options);
     if (!molecule) continue;
@@ -214,18 +215,26 @@ class MolfileStream extends TransformStream {
         this.#buffer += chunk;
         let begin = 0;
         let index = 0;
-        while ((index = this.#buffer.indexOf('\n$$$$', index)) !== -1) {
-          controller.enqueue(this.#buffer.slice(begin, index));
-          index += 5;
-          if (this.#buffer[index] === '\r') {
-            index++;
+        while ((index = this.#buffer.indexOf('$$$$', index)) !== -1) {
+          // we need to check if the delimiter '\n' is in the current buffer
+          // if it is not we need to wait for the next chunk
+          const endOfDelimiter = this.#buffer.indexOf('\n', index);
+          if (endOfDelimiter === -1) {
+            index = begin;
+            break;
+          }
+          const eolLength = this.#buffer[endOfDelimiter - 1] === '\r' ? 2 : 1;
+          // need to remove the last eol because we will split on eol+'>' in getMolecule
+          if (index - eolLength - begin > 40) {
+            controller.enqueue(this.#buffer.slice(begin, index - eolLength));
           }
+          index = endOfDelimiter + eolLength;
           begin = index;
         }
         this.#buffer = this.#buffer.slice(begin);
       },
       flush: (controller) => {
-        if (this.#buffer) {
+        if (this.#buffer && this.#buffer.length > 40) {
           controller.enqueue(this.#buffer);
         }
       },
@@ -234,7 +243,7 @@ class MolfileStream extends TransformStream {
 }
 
 /**
- *  Parse a SDF file
+ *  Parse a SDF file as an iterator
  * @param {ReadableStream} readStream - SDF file to parse
  * @param {object} [options={}] - iterator options
  * @param {Function} [options.filter] - Callback allowing to filter the molecules
@@ -245,10 +254,9 @@ class MolfileStream extends TransformStream {
 async function* iterator(readStream, options = {}) {
   const { eol = '\n', dynamicTyping = true } = options;
 
-  const moleculeStream = readStream.pipeThrough(new MolfileStream());
-  for await (const molfile of moleculeStream) {
-    if (molfile.length < 20) continue;
-    const molecule = getMolecule(molfile, {
+  const moleculeStream = readStream.pipeThrough(new MolfileStream({ eol }));
+  for await (const entry of moleculeStream) {
+    const molecule = getMolecule(entry, {
       eol,
       dynamicTyping,
     });

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "sdf-parser",
-  "version": "7.0.2",
+  "version": "7.0.4",
   "description": "SDF parser",
   "main": "lib/index.js",
   "module": "src/index.js",

package/src/MolfileStream.js

@@ -7,18 +7,26 @@ export class MolfileStream extends TransformStream {
         this.#buffer += chunk;
         let begin = 0;
         let index = 0;
-        while ((index = this.#buffer.indexOf('\n$$$$', index)) !== -1) {
-          controller.enqueue(this.#buffer.slice(begin, index));
-          index += 5;
-          if (this.#buffer[index] === '\r') {
-            index++;
+        while ((index = this.#buffer.indexOf('$$$$', index)) !== -1) {
+          // we need to check if the delimiter '\n' is in the current buffer
+          // if it is not we need to wait for the next chunk
+          const endOfDelimiter = this.#buffer.indexOf('\n', index);
+          if (endOfDelimiter === -1) {
+            index = begin;
+            break;
           }
+          const eolLength = this.#buffer[endOfDelimiter - 1] === '\r' ? 2 : 1;
+          // need to remove the last eol because we will split on eol+'>' in getMolecule
+          if (index - eolLength - begin > 40) {
+            controller.enqueue(this.#buffer.slice(begin, index - eolLength));
+          }
+          index = endOfDelimiter + eolLength;
           begin = index;
         }
         this.#buffer = this.#buffer.slice(begin);
       },
       flush: (controller) => {
-        if (this.#buffer) {
+        if (this.#buffer && this.#buffer.length > 40) {
           controller.enqueue(this.#buffer);
         }
       },

package/src/iterator.js

@@ -3,7 +3,7 @@ import { parseString } from 'dynamic-typing';
 import { MolfileStream } from './MolfileStream.js';
 
 /**
- *  Parse a SDF file
+ *  Parse a SDF file as an iterator
  * @param {ReadableStream} readStream - SDF file to parse
  * @param {object} [options={}] - iterator options
  * @param {Function} [options.filter] - Callback allowing to filter the molecules
@@ -14,10 +14,9 @@ import { MolfileStream } from './MolfileStream.js';
 export async function* iterator(readStream, options = {}) {
   const { eol = '\n', dynamicTyping = true } = options;
 
-  const moleculeStream = readStream.pipeThrough(new MolfileStream());
-  for await (const molfile of moleculeStream) {
-    if (molfile.length < 20) continue;
-    const molecule = getMolecule(molfile, {
+  const moleculeStream = readStream.pipeThrough(new MolfileStream({ eol }));
+  for await (const entry of moleculeStream) {
+    const molecule = getMolecule(entry, {
       eol,
       dynamicTyping,
     });

package/src/parse.js

@@ -13,6 +13,7 @@ import { getMolecule } from './util/getMolecule';
  * @param {object} [options.modifiers] - Object containing callbacks to apply on some specific fields
  * @param {boolean} [options.mixedEOL=false] - Set to true if you know there is a mixture between \r\n and \n
  * @param {string} [options.eol] - Specify the end of line character. Default will be the one found in the file
+ * @returns {object} - Object containing the molecules, the labels and the statistics
  */
 export function parse(sdf, options = {}) {
   options = { ...options };
@@ -53,7 +54,7 @@ export function parse(sdf, options = {}) {
 
   for (let i = 0; i < entriesBoundaries.length; i++) {
     let sdfPart = sdf.slice(...entriesBoundaries[i]);
-
+    if (sdfPart.length < 40) continue;
     let currentLabels = [];
     const molecule = getMolecule(sdfPart, labels, currentLabels, options);
     if (!molecule) continue;