natural-language

Name: natural-language
Author: dpearson2699

2.5kinstalls45forks

How to install natural-language?

dpearson2699/swift-ios-skills/natural-language

$npx -y skills add dpearson2699/swift-ios-skills --skill natural-language

Installs into the current project.

Summary

Tokenize, tag, and analyze natural language text using Apple's NaturalLanguage framework and translate between languages with the Translation framework. Use when adding language identification, sentiment analysis, named entity recognition, part-of-speech tagging, text embeddings,

Files · 1

View on GitHub

SKILL.md

1	# NaturalLanguage + Translation
2
3	Analyze natural language text for tokenization, part-of-speech tagging, named
4	entity recognition, sentiment analysis, language identification, and word/sentence
5	embeddings. Translate text between languages with the Translation framework.
6
7	> This skill covers two related frameworks: NaturalLanguage (`NLTokenizer`, `NLTagger`, `NLEmbedding`) for on-device text analysis, and Translation (`TranslationSession`, `LanguageAvailability`) for language translation.
8
9	Scope boundary: Use this skill after you already have text. It owns
10	tokenization, language identification, POS/NER tagging, sentiment, embeddings,
11	custom `NLModel` classifiers/taggers, and in-app translation. Hand off OCR to
12	`vision-framework`, speech-to-text to `speech-recognition`, UI strings and
13	locale formatting to `ios-localization`, and generative summarization or Apple
14	Intelligence workflows to `apple-on-device-ai`.
15
16	## Contents
17
18	- [Setup](#setup)
19	- [Tokenization](#tokenization)
20	- [Language Identification](#language-identification)
21	- [Part-of-Speech Tagging](#part-of-speech-tagging)
22	- [Named Entity Recognition](#named-entity-recognition)
23	- [Sentiment Analysis](#sentiment-analysis)
24	- [Text Embeddings](#text-embeddings)
25	- [Translation](#translation)
26	- [Common Mistakes](#common-mistakes)
27	- [Review Checklist](#review-checklist)
28	- [References](#references)
29
30	## Setup
31
32	Import `NaturalLanguage` for text analysis and `Translation` for language
33	translation. No special entitlements or capabilities are required for
34	NaturalLanguage. Translation has split availability: system translation
35	presentation is iOS 17.4+ / macOS 14.4+, while `TranslationSession`,
36	`.translationTask()`, `LanguageAvailability`, and batch translation require
37	iOS 18+ / macOS 15+.
38	Direct `TranslationSession(installedSource:target:)` is the non-UI option, but
39	only when the source and target languages are already installed on device.
40
41	```swift
42	import NaturalLanguage
43	import Translation
44	```
45
46	NaturalLanguage classes (`NLTokenizer`, `NLTagger`) are not thread-safe.
47	Use each instance from one thread or dispatch queue at a time.
48
49	## Tokenization
50
51	Segment text into words, sentences, or paragraphs with `NLTokenizer`.
52
53	```swift
54	import NaturalLanguage
55
56	func tokenizeWords(in text: String) -> [String] {
57	let tokenizer = NLTokenizer(unit: .word)
58	tokenizer.string = text
59
60	let range = text.startIndex..<text.endIndex
61	return tokenizer.tokens(for: range).map { String(text[$0]) }
62	}
63	```
64
65	### Token Units
66
67	\| Unit \| Description \|
68	\|---\|---\|
69	\| `.word` \| Individual words \|
70	\| `.sentence` \| Sentences \|
71	\| `.paragraph` \| Paragraphs \|
72	\| `.document` \| Entire document \|
73
74	### Enumerating with Attributes
75
76	Use `enumerateTokens(in:using:)` to detect numeric or emoji tokens.
77
78	```swift
79	let tokenizer = NLTokenizer(unit: .word)
80	tokenizer.string = text
81
82	tokenizer.enumerateTokens(in: text.startIndex..<text.endIndex) { range, attributes in
83	if attributes.contains(.numeric) {
84	print("Number: \(text[range])")
85	}
86	return true // continue enumeration
87	}
88	```
89
90	## Language Identification
91
92	Detect the dominant language of a string with `NLLanguageRecognizer`.
93
94	```swift
95	func detectLanguage(for text: String) -> NLLanguage? {
96	NLLanguageRecognizer.dominantLanguage(for: text)
97	}
98
99	// Multiple hypotheses with confidence scores
100	func languageHypotheses(for text: String, max: Int = 5) -> [NLLanguage: Double] {
101	let recognizer = NLLanguageRecognizer()
102	recognizer.processString(text)
103	return recognizer.languageHypotheses(withMaximum: max)
104	}
105	```
106
107	Constrain the recognizer to expected languages for better accuracy on short text.
108
109	```swift
110	let recognizer = NLLanguageRecognizer()
111	recognizer.languageConstraints = [.english, .french, .spanish]
112	recognizer.processString(text)
113	let detected = recognizer.dominantLanguage
114	```
115
116	## Part-of-Speech Tagging
117
118	Identify nouns, verbs, adjectives, and other lexical classes with `NLTagger`.
119
120	```swift
121	func tagPartsOfSpeech(in text: String) -> [(String, NLTag)] {
122	let tagger = NLTagger(tagSchemes: [.lexicalClass])
123	tagger.string = text
124
125	var results: [(String, NLTag)] = []
126	let range = text.startIndex..<text.endIndex
127	let options: NLTagger.Options = [.omitPunctuation, .omitWhitespace]
128
129	tagger.enumerateTags(in: range, unit: .word, scheme: .lexicalClass, options: options) { tag, tokenRange in
130	if let tag {
131	results.append((String(text[tokenRange]), tag))
132	}
133	return true
134	}
135	return results
136	}
137	```
138
139	### Common Tag Schemes
140
141	\| Scheme \| Output \|
142	\|---\|---\|
143	\| `.lexicalClass` \| Part of speech (noun, verb, adjective) \|
144	\| `.nameType

People who liked this also viewed

6 picks

Type

natural-language

2.5kinstalls45forks

View on GitHub

How to install natural-language?

dpearson2699/swift-ios-skills/natural-language

$npx -y skills add dpearson2699/swift-ios-skills --skill natural-language

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# NaturalLanguage + Translation
2
3	Analyze natural language text for tokenization, part-of-speech tagging, named
4	entity recognition, sentiment analysis, language identification, and word/sentence
5	embeddings. Translate text between languages with the Translation framework.
6
7	> This skill covers two related frameworks: NaturalLanguage (`NLTokenizer`, `NLTagger`, `NLEmbedding`) for on-device text analysis, and Translation (`TranslationSession`, `LanguageAvailability`) for language translation.
8
9	Scope boundary: Use this skill after you already have text. It owns
10	tokenization, language identification, POS/NER tagging, sentiment, embeddings,
11	custom `NLModel` classifiers/taggers, and in-app translation. Hand off OCR to
12	`vision-framework`, speech-to-text to `speech-recognition`, UI strings and
13	locale formatting to `ios-localization`, and generative summarization or Apple
14	Intelligence workflows to `apple-on-device-ai`.
15
16	## Contents
17
18	- [Setup](#setup)
19	- [Tokenization](#tokenization)
20	- [Language Identification](#language-identification)
21	- [Part-of-Speech Tagging](#part-of-speech-tagging)
22	- [Named Entity Recognition](#named-entity-recognition)
23	- [Sentiment Analysis](#sentiment-analysis)
24	- [Text Embeddings](#text-embeddings)
25	- [Translation](#translation)
26	- [Common Mistakes](#common-mistakes)
27	- [Review Checklist](#review-checklist)
28	- [References](#references)
29
30	## Setup
31
32	Import `NaturalLanguage` for text analysis and `Translation` for language
33	translation. No special entitlements or capabilities are required for
34	NaturalLanguage. Translation has split availability: system translation
35	presentation is iOS 17.4+ / macOS 14.4+, while `TranslationSession`,
36	`.translationTask()`, `LanguageAvailability`, and batch translation require
37	iOS 18+ / macOS 15+.
38	Direct `TranslationSession(installedSource:target:)` is the non-UI option, but
39	only when the source and target languages are already installed on device.
40
41	```swift
42	import NaturalLanguage
43	import Translation
44	```
45
46	NaturalLanguage classes (`NLTokenizer`, `NLTagger`) are not thread-safe.
47	Use each instance from one thread or dispatch queue at a time.
48
49	## Tokenization
50
51	Segment text into words, sentences, or paragraphs with `NLTokenizer`.
52
53	```swift
54	import NaturalLanguage
55
56	func tokenizeWords(in text: String) -> [String] {
57	let tokenizer = NLTokenizer(unit: .word)
58	tokenizer.string = text
59
60	let range = text.startIndex..<text.endIndex
61	return tokenizer.tokens(for: range).map { String(text[$0]) }
62	}
63	```
64
65	### Token Units
66
67	\| Unit \| Description \|
68	\|---\|---\|
69	\| `.word` \| Individual words \|
70	\| `.sentence` \| Sentences \|
71	\| `.paragraph` \| Paragraphs \|
72	\| `.document` \| Entire document \|
73
74	### Enumerating with Attributes
75
76	Use `enumerateTokens(in:using:)` to detect numeric or emoji tokens.
77
78	```swift
79	let tokenizer = NLTokenizer(unit: .word)
80	tokenizer.string = text
81
82	tokenizer.enumerateTokens(in: text.startIndex..<text.endIndex) { range, attributes in
83	if attributes.contains(.numeric) {
84	print("Number: \(text[range])")
85	}
86	return true // continue enumeration
87	}
88	```
89
90	## Language Identification
91
92	Detect the dominant language of a string with `NLLanguageRecognizer`.
93
94	```swift
95	func detectLanguage(for text: String) -> NLLanguage? {
96	NLLanguageRecognizer.dominantLanguage(for: text)
97	}
98
99	// Multiple hypotheses with confidence scores
100	func languageHypotheses(for text: String, max: Int = 5) -> [NLLanguage: Double] {
101	let recognizer = NLLanguageRecognizer()
102	recognizer.processString(text)
103	return recognizer.languageHypotheses(withMaximum: max)
104	}
105	```
106
107	Constrain the recognizer to expected languages for better accuracy on short text.
108
109	```swift
110	let recognizer = NLLanguageRecognizer()
111	recognizer.languageConstraints = [.english, .french, .spanish]
112	recognizer.processString(text)
113	let detected = recognizer.dominantLanguage
114	```
115
116	## Part-of-Speech Tagging
117
118	Identify nouns, verbs, adjectives, and other lexical classes with `NLTagger`.
119
120	```swift
121	func tagPartsOfSpeech(in text: String) -> [(String, NLTag)] {
122	let tagger = NLTagger(tagSchemes: [.lexicalClass])
123	tagger.string = text
124
125	var results: [(String, NLTag)] = []
126	let range = text.startIndex..<text.endIndex
127	let options: NLTagger.Options = [.omitPunctuation, .omitWhitespace]
128
129	tagger.enumerateTags(in: range, unit: .word, scheme: .lexicalClass, options: options) { tag, tokenRange in
130	if let tag {
131	results.append((String(text[tokenRange]), tag))
132	}
133	return true
134	}
135	return results
136	}
137	```
138
139	### Common Tag Schemes
140
141	\| Scheme \| Output \|
142	\|---\|---\|
143	\| `.lexicalClass` \| Part of speech (noun, verb, adjective) \|
144	\| `.nameType

People who liked this also viewed

6 picks

Type