NLP Program to Detect Language with Python

Use the Python library langdetect to detect language by calling detect(text) which returns the language code like 'en' for English.

Examples

InputHello, how are you?

Outputen

InputBonjour tout le monde

Outputfr

Inputこれは日本語の文章です。

Outputja

How to Think About It

To detect the language of a text, the program looks for patterns and clues in the words and characters that match known languages. It compares the input text against models built from many languages and picks the best match.

Algorithm

1

Get the input text from the user or source.

2

Use a language detection tool or library to analyze the text.

3

The tool compares text features to known language profiles.

4

Return the language code that best matches the text.

Code

python

from langdetect import detect

def detect_language(text: str) -> str:
    return detect(text)

sample_text = "Hello, how are you?"
language = detect_language(sample_text)
print(f"Detected language: {language}")

Output

Detected language: en

Dry Run

Let's trace detecting language for the input 'Hello, how are you?' through the code

1

Input text

text = 'Hello, how are you?'

2

Call detect function

detect('Hello, how are you?')

3

Analyze text

The function matches text patterns to English language profile

4

Return result

Returns 'en' as detected language code

Step	Action	Value
1	Input text	Hello, how are you?
2	Call detect()	detect('Hello, how are you?')
3	Match language profile	English (en)
4	Return language code	en

Why This Works

Step 1: Importing langdetect

The langdetect library provides a simple detect() function to identify language from text.

Step 2: Detecting language

The detect() function analyzes the input text's character and word patterns to guess the language.

Step 3: Returning language code

It returns a short code like en for English or fr for French, which is easy to use in programs.

Alternative Approaches

Using fasttext language identification

python

import fasttext
model = fasttext.load_model('lid.176.ftz')
def detect_language_fasttext(text: str) -> str:
    prediction = model.predict(text)
    return prediction[0][0].replace('__label__', '')

print(detect_language_fasttext('Hello, how are you?'))

FastText is faster and supports more languages but requires downloading a model file.

Using TextBlob library

python

from textblob import TextBlob

def detect_language_textblob(text: str) -> str:
    return TextBlob(text).detect_language()

print(detect_language_textblob('Bonjour tout le monde'))

TextBlob is easy to use but depends on internet connection for language detection.

Complexity: O(n) time, O(1) space

Time Complexity

The detection scans the input text once, so time grows linearly with text length.

Space Complexity

Uses constant extra space for internal language profiles and variables.

Which Approach is Fastest?

FastText is faster for large texts but needs model loading; langdetect is simple and good for small to medium texts.

Approach	Time	Space	Best For
langdetect	O(n)	O(1)	Simple scripts and small texts
fasttext	O(n)	O(m) model size	High speed and many languages
TextBlob	O(n)	O(1)	Easy use with internet access