CS50 Final Project - PythonicGrammarChecker

Video Demo: https://youtu.be/5-No6oaCl6Y

Description: I always use Grammarly to figure out my errors while writing my school assignments or any such event that might show up. After learning Python, I thought I could implement a similar program in Python.

---

Libraries Used

• spaCy (open-source software library for advanced natural language processing)
• rich (writing colored text to the terminal)
• pyspellchecker (library responsible for finding misspelled words)
• language-tool-python (responsible for pointing out potential grammatical errors)
• diff-match-patch (finds the difference between two given strings)

---

How Does It Work?

def main()

The main function prompts the user for a sentence or a paragraph (depending on the use case).

• Other functionalities of the main function:
  • Strips leading and trailing whitespace from the user-provided input.
  • Calls out the user if only digits are provided.
  • Once the input is cleaned, it calls a series of functions that process the text for spelling and grammar corrections.

---

def spell_()

Expects a sentence or paragraph and corrects the spelling mistakes in the input.

• How it works:

  • A variable called spelling_e keeps track of spelling errors.
  • The function uses spaCy to tokenize the input for better accuracy.
  • The input is broken down into words, and each word is checked against a dictionary.
  • If a word is incorrect, it is replaced with the closest correct match.

  correct_text = []
  for word in words_list:
      correct_text.append(spell.correction(word))
      if word != spell.correction(word):
          spelling_e += 1

  • Finally, the function returns the corrected text with spelling errors fixed.

---

def grammar_()

Expects a sentence or paragraph and corrects grammatical errors.

• How it works:

  • A variable called grammar_e keeps track of grammatical errors.

  • Uses language-tool-python to analyze and correct grammar mistakes.

    grammar_e += len(tool.check(para))
    return tool.correct(para)

---

def highlight()

Uses the rich library to highlight corrections and display them in a neat, color-coded format.

• How changes are detected:

  • The diff-match-patch library compares the original and corrected text.

  • Semantic comparison is used to determine differences.

    diff = dmp.diff_main(old, new)
    dmp.diff_cleanupSemantic(diff)

  • The differences are then formatted:

    for op, data in diff:
        if op == diff_match_patch.DIFF_EQUAL:
            colored_diff.append(data)
        if op == diff_match_patch.DIFF_INSERT:
            colored_diff.append(f"[green]{data}[/]")

  • Words that were changed are highlighted in green.

---

Implementation of the Rich Library

• Used to display corrections with formatting and colors.

• Highlights spelling and grammar mistakes:

  console.print(Align.center(colored_diff))

  • Aligns output in the center.
  • Recognizes [green]...[/] formatting to color the corrected text.

• Displays mistake counts:

  spell_p = Panel(Text(f'Spelling Mistake(s): {spelling_e}', justify="center", style="red"))
  grammar_p = Panel(Text(f'Grammar Mistake(s): {grammar_e}', justify="center", style="red"))

---

Downsides

• The corrections may not always be 100% accurate.
• Capitalization changes can sometimes be flagged as errors.
• The program takes a longer runtime due to the use of multiple libraries.

---

Possible Improvements

• Implement multithreading to improve performance.
• Train a better NLP model for improved grammar correction.
• Allow users to choose between different correction modes.

---