progress on collect line numbers

Author: canbax

src/trie.ts

@@ -3,9 +3,10 @@ import {
   readGzippedJSON,
   TRIE_FILE,
   TSV_DB_FILE,
+  uniteSets,
   writeGzippedJSON,
 } from "./util.js";
-import { readFileSync, writeFileSync } from "fs";
+import { readFileSync } from "fs";
 
 interface EncodedTrieNode {
   c?: Record<string, EncodedTrieNode>;
@@ -70,6 +71,64 @@ export class Trie {
     return currentNode;
   }
 
+  findLastNodeMatch(query: string) {
+    // Initialize the currentNode pointer with the root node
+    let currentNode = this.root;
+
+    let i = 0;
+    // Iterate across the length of the string
+    for (i = 0; i < query.length; i++) {
+      const c = query[i] as string;
+      // Check if the node exist for the current character in the Trie.
+      if (!currentNode.children[c]) {
+        // Given word as a prefix does not exist in Trie
+        return { currentNode, i };
+      }
+
+      // Move the currentNode pointer to the already existing node for current character.
+      currentNode = currentNode.children[c];
+    }
+
+    // Prefix exist in the Trie
+    return { currentNode, i };
+  }
+
+  autocomplete(query: string, maxResultCount = 10) {
+    const resultSet = new Set<number>();
+
+    const { i, currentNode } = this.findLastNodeMatch(query);
+    this.collectLineNumbersWithBFS(
+      currentNode,
+      query.length,
+      i,
+      resultSet,
+      maxResultCount,
+    );
+    return resultSet;
+  }
+
+  collectLineNumbersWithBFS(
+    currentNode: TrieNode,
+    querySize: number,
+    queryIndex: number,
+    resultSet: Set<number>,
+    maxResultCount = 10,
+  ) {
+    if (resultSet.size > maxResultCount) return;
+    uniteSets(resultSet, currentNode.lineNumbers);
+    if (queryIndex >= querySize) return;
+
+    for (const [_, childNode] of Object.entries(currentNode.children)) {
+      this.collectLineNumbersWithBFS(
+        childNode,
+        querySize,
+        queryIndex++,
+        resultSet,
+        maxResultCount,
+      );
+    }
+  }
+
   searchPrefix(query: string): number[] | null {
     const searchTerm = query.trim().toLowerCase();
     if (searchTerm.length < 1) return null;

src/util.ts

@@ -189,3 +189,17 @@ export async function readLinesFromTSV(
 
   return linesRead;
 }
+
+/**
+ * add all elements of `setB` to `setA`
+ *
+ * @export
+ * @template T
+ * @param {Set<T>} setA
+ * @param {Set<T>} setB
+ */
+export function uniteSets<T>(setA: Set<T>, setB: Set<T>) {
+  for (const elem of setB) {
+    setA.add(elem);
+  }
+}

test/util.spec.ts

@@ -32,12 +32,12 @@ describe("util functions", () => {
       ]);
     });
 
-    it("should 1000 lines in maximum 50 ms", async () => {
+    it("should 100 lines in maximum 50 ms", async () => {
       function createArray(n: number) {
         return Array.from({ length: n }, (v, k) => k + 1);
       }
       const t1 = new Date().getTime();
-      await readLinesFromTSV(TSV_DB_FILE, createArray(1000));
+      await readLinesFromTSV(TSV_DB_FILE, createArray(100));
       const deltaTime = new Date().getTime() - t1;
       expect(deltaTime).toBeLessThan(50);
     });