macos-vision 0.1.4 → 0.2.0

package/CHANGELOG.md

@@ -1,3 +1,14 @@
 # Changelog
 
+## [0.2.0](https://github.com/woladi/macos-vision/compare/v0.1.4...v0.2.0) (2026-04-08)
+
+### Features
+
+* add confidence to VisionBlock and Barcode ([a87df27](https://github.com/woladi/macos-vision/commit/a87df275e51dec4b57fbff6e3bffc4220b96b4d7))
+
+### Bug Fixes
+
+* correct mkdirSync, CLI error on missing file, execFile timeout, README scope ([1cef2c7](https://github.com/woladi/macos-vision/commit/1cef2c7078430c9182fcd39792cf0c002833203f))
+* replace try? with do/catch in Swift helper — surface Vision errors properly ([f287065](https://github.com/woladi/macos-vision/commit/f2870655225806070be3db462ea15923201fecbf))
+
 ## 0.1.4 (2026-04-08)

package/README.md

@@ -22,6 +22,19 @@ npm install macos-vision
 
 The native Swift binary is compiled automatically on install.
 
+## What this is (and isn't)
+
+`macos-vision` gives you **raw Apple Vision results** — text, coordinates, bounding boxes, labels.
+
+It is **not** a document pipeline. It does not:
+- Convert PDFs or images to Markdown
+- Understand document structure (headings, tables, paragraphs)
+- Chain multiple detections into a final report
+
+For those use cases, use the raw output as input to an LLM or a post-processing layer of your own.
+
+---
+
 ## CLI
 
 ```bash

package/dist/cli.js

@@ -1,8 +1,6 @@
 #!/usr/bin/env node
-import { resolve, dirname } from 'path';
-import { fileURLToPath } from 'url';
+import { resolve } from 'path';
 import { ocr, detectFaces, detectBarcodes, detectRectangles, detectDocument, classify, } from './index.js';
-const __dirname = dirname(fileURLToPath(import.meta.url));
 const USAGE = `
 Usage: vision-cli [options] <image>
 
@@ -28,9 +26,14 @@ if (rawArgs.includes('--help') || rawArgs.length === 0) {
     console.log(USAGE);
     process.exit(0);
 }
-const flags = new Set(rawArgs.filter(a => a.startsWith('--')));
-const fileArgs = rawArgs.filter(a => !a.startsWith('--'));
-const imagePath = fileArgs[0] || resolve(__dirname, '../test/fixtures/sample.png');
+const flags = new Set(rawArgs.filter((a) => a.startsWith('--')));
+const fileArgs = rawArgs.filter((a) => !a.startsWith('--'));
+if (!fileArgs[0]) {
+    console.error('Error: no image path provided.\n');
+    console.log(USAGE);
+    process.exit(1);
+}
+const imagePath = resolve(fileArgs[0]);
 const runAll = flags.has('--all');
 const runOcr = runAll || flags.has('--ocr');
 const runBlocks = runAll || flags.has('--blocks');
@@ -40,9 +43,14 @@ const runRects = runAll || flags.has('--rectangles');
 const runDoc = runAll || flags.has('--document');
 const runClassify = runAll || flags.has('--classify');
 // Default: OCR text when no feature flag is given
-const anyFeatureFlag = runAll || flags.has('--ocr') || flags.has('--blocks') ||
-    flags.has('--faces') || flags.has('--barcodes') || flags.has('--rectangles') ||
-    flags.has('--document') || flags.has('--classify');
+const anyFeatureFlag = runAll ||
+    flags.has('--ocr') ||
+    flags.has('--blocks') ||
+    flags.has('--faces') ||
+    flags.has('--barcodes') ||
+    flags.has('--rectangles') ||
+    flags.has('--document') ||
+    flags.has('--classify');
 const useDefault = !anyFeatureFlag;
 async function main() {
     try {
@@ -51,27 +59,27 @@ async function main() {
             console.log(text);
         }
         if (runBlocks) {
-            const blocks = await ocr(imagePath, { format: 'blocks' });
+            const blocks = (await ocr(imagePath, { format: 'blocks' }));
             console.log(JSON.stringify(blocks, null, 2));
         }
         if (runFaces) {
-            const faces = await detectFaces(imagePath);
+            const faces = (await detectFaces(imagePath));
             console.log(JSON.stringify(faces, null, 2));
         }
         if (runBarcodes) {
-            const barcodes = await detectBarcodes(imagePath);
+            const barcodes = (await detectBarcodes(imagePath));
             console.log(JSON.stringify(barcodes, null, 2));
         }
         if (runRects) {
-            const rectangles = await detectRectangles(imagePath);
+            const rectangles = (await detectRectangles(imagePath));
             console.log(JSON.stringify(rectangles, null, 2));
         }
         if (runDoc) {
-            const doc = await detectDocument(imagePath);
+            const doc = (await detectDocument(imagePath));
             console.log(JSON.stringify(doc, null, 2));
         }
         if (runClassify) {
-            const labels = await classify(imagePath);
+            const labels = (await classify(imagePath));
             console.log(JSON.stringify(labels, null, 2));
         }
     }

package/dist/index.d.ts

@@ -9,6 +9,8 @@ export interface VisionBlock {
     width: number;
     /** Height, 0–1 relative to image */
     height: number;
+    /** OCR transcription confidence, 0–1 */
+    confidence: number;
 }
 export interface OcrOptions {
     /** Return plain text (default) or structured blocks with coordinates */
@@ -46,6 +48,8 @@ export interface Barcode {
     width: number;
     /** Height, 0–1 relative to image */
     height: number;
+    /** Detection confidence, 0–1 */
+    confidence: number;
 }
 export declare function detectBarcodes(imagePath: string): Promise<Barcode[]>;
 export interface Rectangle {

package/dist/index.js

@@ -5,19 +5,31 @@ import { fileURLToPath } from 'url';
 const execFileAsync = promisify(execFile);
 const __dirname = dirname(fileURLToPath(import.meta.url));
 const BIN_PATH = resolve(__dirname, '../bin/vision-helper');
+const BINARY_TIMEOUT_MS = 30_000;
 async function run(flag, imagePath) {
-    const { stdout } = await execFileAsync(BIN_PATH, [flag, resolve(imagePath)]);
+    const { stdout } = await execFileAsync(BIN_PATH, [flag, resolve(imagePath)], {
+        timeout: BINARY_TIMEOUT_MS,
+    });
     return stdout;
 }
 export async function ocr(imagePath, options = {}) {
     const absPath = resolve(imagePath);
     const { format = 'text' } = options;
     if (format === 'blocks') {
-        const { stdout } = await execFileAsync(BIN_PATH, ['--json', absPath]);
+        const { stdout } = await execFileAsync(BIN_PATH, ['--json', absPath], {
+            timeout: BINARY_TIMEOUT_MS,
+        });
         const raw = JSON.parse(stdout);
-        return raw.map((b) => ({ text: b.t, x: b.x, y: b.y, width: b.w, height: b.h }));
+        return raw.map((b) => ({
+            text: b.t,
+            x: b.x,
+            y: b.y,
+            width: b.w,
+            height: b.h,
+            confidence: b.confidence,
+        }));
     }
-    const { stdout } = await execFileAsync(BIN_PATH, [absPath]);
+    const { stdout } = await execFileAsync(BIN_PATH, [absPath], { timeout: BINARY_TIMEOUT_MS });
     return stdout.trim();
 }
 export async function detectFaces(imagePath) {
@@ -33,6 +45,7 @@ export async function detectBarcodes(imagePath) {
         y: b.y,
         width: b.w,
         height: b.h,
+        confidence: b.confidence,
     }));
 }
 export async function detectRectangles(imagePath) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "macos-vision",
-  "version": "0.1.4",
+  "version": "0.2.0",
   "description": "Apple Vision OCR & image analysis for Node.js — native, fast, offline, no API keys",
   "author": "Adrian Wolczuk",
   "license": "MIT",

package/scripts/build-native.js

@@ -13,9 +13,7 @@ if (existsSync(binPath)) {
   process.exit(0);
 }
 
-if (!mkdirSync(binDir, { recursive: true }) === false) {
-  // dir created
-}
+mkdirSync(binDir, { recursive: true });
 
 try {
   execSync(`swiftc -O "${swiftSrc}" -o "${binPath}"`, { stdio: 'inherit' });