Performance Benchmarks

urdu-text-utils is designed to be fast, lightweight, and zero-dependency. All functions are pure, synchronous, and side-effect free.

Benchmark Results

Measured on Node.js 22 (AMD Ryzen 5, 16GB RAM). Higher ops/sec is better.

Normalization

Function	Input	Ops/sec	Time/op
normalizeUrdu	Short (10 chars)	81,000+	0.012 ms
normalizeUrdu	Medium (100 chars)	21,000+	0.046 ms
normalizeUrdu	Long (500 chars)	4,400+	0.225 ms
removeDiacritics	Short	1,870,000+	0.001 ms
removeDiacritics	With diacritics	365,000+	0.003 ms
foldUrdu	Short	81,000+	0.012 ms

Transliteration

Function	Input	Ops/sec	Time/op
romanize	Short (10 chars)	60,000+	0.017 ms
romanize	Medium (100 chars)	7,800+	0.128 ms
romanToUrdu	Single word	436,000+	0.002 ms
romanToUrdu	Medium phrase	43,000+	0.023 ms

Search

Function	Corpus Size	Ops/sec	Time/op
searchUrdu (exact)	100 items	600+	1.667 ms
searchUrdu (fuzzy)	100 items	355+	2.815 ms

Sorting

Function	Items	Ops/sec	Time/op
sortUrdu	20 words	2,700+	0.365 ms

Statistics

Function	Input	Ops/sec	Time/op
countWords	Medium	108,000+	0.009 ms
countSentences	Medium	111,000+	0.009 ms
analyzeUrdu	Medium	17,600+	0.057 ms
splitWords	Medium	109,000+	0.009 ms
splitSentences	Medium	137,000+	0.007 ms

Detection

Function	Input	Ops/sec	Time/op
isUrdu	Medium	39,000+	0.025 ms
urduRatio	Medium	41,000+	0.024 ms

Numbers & Stemming

Function	Input	Ops/sec	Time/op
convertNumbers	Short	537,000+	0.002 ms
stemUrdu	Single word	132,000+	0.008 ms
stemUrduText	Medium	9,500+	0.105 ms

Date & Time

Function	Input	Ops/sec	Time/op
formatUrduDate	Current date	152,000+	0.007 ms
timeAgoUrdu	2 hours ago	376,000+	0.003 ms

Key Takeaways

1. Fastest functions — removeDiacritics, convertNumbers, and romanToUrdu run at 400K–1.8M ops/sec. These are essentially regex-based and have negligible overhead.

2. Normalization — Handles 80K+ short strings/sec, which is well above typical UI requirements (even for real-time search as the user types).

3. Search — Exact search over 100 items completes in ~1.7ms; fuzzy search adds ~1ms for edit distance calculation. For larger corpora, consider building an index with foldUrdu keys.

4. Transliteration — Dictionary lookups are near-instant (436K ops/sec). Rule-based fallback adds ~0.1ms per word, which is imperceptible in UI contexts.

5. Zero dependencies — No external libraries means smaller bundle size and no transitive dependency overhead.

Bundle Size

Format	Size (minified)	Size (min + gzip)
ESM	162 KB	~11 KB
CJS	163 KB	~11 KB
IIFE	128 KB	~11 KB

Running Benchmarks

To run benchmarks yourself:

npx tsx bench.mjs

The benchmark script is in the repository root. Results will vary based on your hardware and Node.js version.