From 560f427c430107de9a2785960667dddfa7785fbe Mon Sep 17 00:00:00 2001
From: Michael <michael.hladky@push-based.io>
Date: Tue, 11 Mar 2025 02:41:10 +0100
Subject: [PATCH] feat(utils): add find in file logic

---
 packages/utils/src/lib/file-system.ts         | 105 ++++++++++++++++++
 .../utils/src/lib/file-system.unit.test.ts    |  41 +++++++
 2 files changed, 146 insertions(+)

diff --git a/packages/utils/src/lib/file-system.ts b/packages/utils/src/lib/file-system.ts
index 966dbc536..da5e47ea0 100644
--- a/packages/utils/src/lib/file-system.ts
+++ b/packages/utils/src/lib/file-system.ts
@@ -1,7 +1,10 @@
 import { bold, gray } from 'ansis';
 import { type Options, bundleRequire } from 'bundle-require';
+import * as fs from 'node:fs';
 import { mkdir, readFile, readdir, rm, stat } from 'node:fs/promises';
 import path from 'node:path';
+import * as readline from 'node:readline';
+import type { SourceFileLocation } from '@code-pushup/models';
 import { formatBytes } from './formatting.js';
 import { logMultipleResults } from './log-results.js';
 import { ui } from './logging.js';
@@ -93,6 +96,7 @@ export type CrawlFileSystemOptions<T> = {
   pattern?: string | RegExp;
   fileTransform?: (filePath: string) => Promise<T> | T;
 };
+
 export async function crawlFileSystem<T = string>(
   options: CrawlFileSystemOptions<T>,
 ): Promise<T[]> {
@@ -159,3 +163,104 @@ export function filePathToCliArg(filePath: string): string {
 export function projectToFilename(project: string): string {
   return project.replace(/[/\\\s]+/g, '-').replace(/@/g, '');
 }
+
+export type LineHit = {
+  startColumn: number;
+  endColumn: number;
+};
+
+export type FileHit = Pick<SourceFileLocation, 'file'> &
+  Exclude<SourceFileLocation['position'], undefined>;
+
+const escapeRegExp = (str: string): string =>
+  str.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+const ensureGlobalRegex = (pattern: RegExp): RegExp =>
+  new RegExp(
+    pattern.source,
+    pattern.flags.includes('g') ? pattern.flags : `${pattern.flags}g`,
+  );
+
+const findAllMatches = (
+  line: string,
+  searchPattern: string | RegExp | ((line: string) => LineHit[] | null),
+): LineHit[] => {
+  if (typeof searchPattern === 'string') {
+    return [...line.matchAll(new RegExp(escapeRegExp(searchPattern), 'g'))].map(
+      ({ index = 0 }) => ({
+        startColumn: index,
+        endColumn: index + searchPattern.length,
+      }),
+    );
+  }
+
+  if (searchPattern instanceof RegExp) {
+    return [...line.matchAll(ensureGlobalRegex(searchPattern))].map(
+      ({ index = 0, 0: match }) => ({
+        startColumn: index,
+        endColumn: index + match.length,
+      }),
+    );
+  }
+
+  return searchPattern(line) || [];
+};
+
+/**
+ * Reads a file line-by-line and checks if it contains the search pattern.
+ * @param file - The file path to check.
+ * @param searchPattern - The pattern to match.
+ * @param options - Additional options. If true, the search will stop after the first hit.
+ * @returns Promise<FileHit[]> - List of hits with matching details.
+ */
+export async function findInFile(
+  file: string,
+  searchPattern: string | RegExp | ((line: string) => LineHit[] | null),
+  options?: { bail?: boolean },
+): Promise<FileHit[]> {
+  const { bail = false } = options || {};
+  const hits: FileHit[] = [];
+
+  return new Promise((resolve, reject) => {
+    const stream = fs.createReadStream(file, { encoding: 'utf8' });
+    const rl = readline.createInterface({ input: stream });
+    // eslint-disable-next-line functional/no-let
+    let lineNumber = 0;
+    // eslint-disable-next-line functional/no-let
+    let isResolved = false;
+
+    rl.on('line', line => {
+      lineNumber++;
+      const matches = findAllMatches(line, searchPattern);
+
+      matches.forEach(({ startColumn, endColumn }) => {
+        // eslint-disable-next-line functional/immutable-data
+        hits.push({
+          file,
+          startLine: lineNumber,
+          startColumn,
+          endLine: lineNumber,
+          endColumn,
+        });
+
+        if (bail && !isResolved) {
+          isResolved = true;
+          stream.destroy();
+          resolve(hits);
+        }
+      });
+    });
+    rl.once('close', () => {
+      if (!isResolved) {
+        isResolved = true;
+      }
+      resolve(hits); // Resolve only once after closure
+    });
+
+    rl.once('error', error => {
+      if (!isResolved) {
+        isResolved = true;
+        reject(error);
+      }
+    });
+  });
+}
diff --git a/packages/utils/src/lib/file-system.unit.test.ts b/packages/utils/src/lib/file-system.unit.test.ts
index dfb76ee06..957dcb2ef 100644
--- a/packages/utils/src/lib/file-system.unit.test.ts
+++ b/packages/utils/src/lib/file-system.unit.test.ts
@@ -8,6 +8,7 @@ import {
   crawlFileSystem,
   ensureDirectoryExists,
   filePathToCliArg,
+  findInFile,
   findLineNumberInText,
   findNearestFile,
   logMultipleFileResults,
@@ -263,3 +264,43 @@ describe('projectToFilename', () => {
     expect(projectToFilename(project)).toBe(file);
   });
 });
+
+describe('findInFile', () => {
+  const file = 'file.txt';
+  const content =
+    'line 1 - even:false\nline 2 - even:true\nline 3 - even:false\nline 4 - even:true\nline 5 - even:false\n';
+  const filePath = path.join(MEMFS_VOLUME, file);
+
+  beforeEach(() => {
+    vol.reset();
+    vol.fromJSON({ [file]: content }, MEMFS_VOLUME);
+  });
+
+  it('should find pattern in a file if a string is given', async () => {
+    const result = await findInFile(filePath, 'line 3');
+    expect(result).toStrictEqual([
+      {
+        file: filePath,
+        endColumn: 6,
+        endLine: 3,
+        startColumn: 0,
+        startLine: 3,
+      },
+    ]);
+  });
+
+  // @TODO any second test will fail
+  // Error: EBADF: bad file descriptor, close
+  it.todo('should find pattern in a file if a RegEx is given', async () => {
+    const result = await findInFile('file.txt', new RegExp('line 3', 'g'));
+    expect(result).toStrictEqual([
+      {
+        file: 'file.txt',
+        endColumn: 6,
+        endLine: 3,
+        startColumn: 0,
+        startLine: 3,
+      },
+    ]);
+  });
+});