NLP: Characters, words and sentences

Compute the first statistics of any NLP project for one line of text.

What to build

A program that counts the characters, letters, words and sentences in a line of text.

Requirements

  • Characters: every character of the line, spaces included. Letters: characters for which isalpha() is true.
  • Words: the pieces you get by splitting on whitespace.
  • A sentence ends at each ., ! or ? that is the last character of the line or is followed by a space (so 3.14 does not end a sentence, but Dr. does).
  • Print characters: N, letters: N, words: N and sentences: N on four lines.

Input

One line of text.

Output

Four lines.

Constraints

The line has at most 1,000 characters.

Example

Input:

Hi there. How are you?

Output:

characters: 22
letters: 16
words: 5
sentences: 2

Why it works

Counting units is the start of every text pipeline, and the sentence rule shows why splitting text is harder than it looks: a full stop inside a number is not a sentence end, but the simple rule still cannot tell an abbreviation from the end of a sentence.

Examples

Example 1 · Input
Hi there. How are you?
Expected output
characters: 22
letters: 16
words: 5
sentences: 2
Example 2 · Input
Wait... what?!
Expected output
characters: 14
letters: 8
words: 2
sentences: 2

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

`len(text)` counts characters, `text.split()` gives the words, and `ch.isalpha()` is true for letters.

Hint 2

A mark ends a sentence when it is the last character or the next character is a space: `i + 1 == len(text) or text[i + 1] == " "`.

Solution

Show one correct solution (try it yourself first)
text = input()
sentences = 0
for i, ch in enumerate(text):
    if ch in ".!?" and (i + 1 == len(text) or text[i + 1] == " "):
        sentences += 1
print("characters:", len(text))
print("letters:", sum(1 for ch in text if ch.isalpha()))
print("words:", len(text.split()))
print("sentences:", sentences)

Lesson: What natural language processing is →