Leetcode Solution Automation For Github

A python script to generate solution files for LeetCode problems to be pushed to GitHub.

Continuation...

As I've already mentioned in my previous dump, I had a habit of posting solutions to LeetCode problems that I solve on my GitHub repository.

In that dump I solved the 1st part of the problem, which was to scrap the question data from the LeetCode website. Now, I need to automate the 2nd part of the problem i.e. the process of generating the solution file. And as I maintain a README file (as a log of all the solutions), I've to automate the process of updating that file as well.

The Problem To Solve

I've a specific template for the solution files which I maintain in my repository. It is like a design standard which I've imposed on myself 😁.

// <Question ID> : <Question Title>
 
/*
<Problem Statement>
*/
 
/*
<Example Test Cases>
*/
 
/*
<Constraints>
*/
 
/*
<My Approached Explained>
*/
 
class Solution {
public:
    function_name() {
        // Code goes here
    }
};

So, I just need to perform these things in order:

1. Extract the question data from the LeetCode website.
2. Create a file with my solution for the problem.
3. Either create a new file for the explanation of the approach or write it in the same file.
4. Generate a new CPP file with the extracted data & the solution.
5. Update the README file with the new solution.

Pretty simple, right?

I guess it isn't so simple as it sounds. I've another habit of overcomplicating things 😓, the logs are supposed to be maintained in ascending order of Question ID. So, I can't just simply append the new entry to the end of the README file. Instead, I've to find the correct position to insert the new entry.

No worries, it's not that hard to do but I'm too sleepy 😴.

Let's Code It

The Input File

I'll be using a simple text file to store data such as URL, My Approach & Solution, which I'll be using to generate the solution file.

https://leetcode.com/problems/<problem-name>/description/
 
/*
<Approach Explained Here>
*/
 
class Solution {
public:
    function_name() {
        // code here
    }
};

Time For Some REGEX Magic

Now, we need to separately take them from the input file. There are many ways to do this, but I'll be using REGEX (re module in Python) to extract the data. I'll be using the patterns in detail in the script, but here's a brief overview:

def extract_from_input(content : str):    
    urlPattern = r'https://leetcode.com/problems/[\w-]+/description/'
    commentPattern = r'/\*([^*]|[\r\n]|(\*([^/]|[\r\n])))*\*/'
 
    urlMatch = re.search(urlPattern, content, re.MULTILINE)
    commentMatch = re.search(commentPattern, content, re.MULTILINE)
    
    url = urlMatch.group(0) if urlMatch else None
    comment = commentMatch.group(0) if commentMatch else None
    
    index = commentMatch.end() if commentMatch else 0
    code = content[index:].rstrip()
 
    return url, comment, code

The URL Pattern

urlPattern = r'https://leetcode.com/problems/[\w-]+/description/'

[\w-]+ matches any number of instances of characters (equal to [a-zA-Z0-9_]) & -.

The Comment Pattern

commentPattern = r'/\*([^*]|[\r\n]|(\*([^/]|[\r\n])))*\*/'

1. /\* matches the starting of the comment.
2. ( opens a group.
   • [^*] negates the * character i.e. it consumes any * which is potentially not the end of the comment in */.
   • [\r\n] matches any return or newline character.
   • | is the OR operator.
   • (\*([^/]|[\r\n])) matches any nested sequences of comments.
3. )* marks the end of the group & matches any number of instances of the group.
4. \*/ matches the end of the comment.

Generate The Solution File

...
url, comment, code = extract_data(content)
questionId, title, question, difficulty, example, constraint = scrape_data(url)
 
fileName = f'{questionId}_{title.replace(' ', '_')}.cpp'
 
with open(fileName, 'w') as f:
    f.write(f'// {questionId} : {title}\n\n')
    f.write(f'/*\n{question}\n*/\n\n')
    f.write(f'/*\n{example}\n*/\n\n')
    f.write(f'/*\n{constraint}\n*/\n\n')
    f.write(comment)
    f.write(code)

No rocket science here, just writing the scraped data & the extracted data into a file with the required format. Note that the file name is generated using the questionId & title of the question as I required.

Update The README File

readme_path = '../README.md'
if not os.path.exists(readme_path):
    with open(readme_path, 'w') as f:
        # Create the README file if it doesn't exist
        
with open(readme_path, 'r') as f:
    lines = f.readlines()
    
table_start = lines.index("| # | Title | Solution |\n")
insert_index = table_start + 2
while insert_index < len(lines) and lines[insert_index].split("|")[1] < f"{int(questionId):04d}":
    insert_index += 1
    
new_entry = f"|{int(questionId):04d}|[{title}]({url})|[cpp](Solutions_CPP/{fileName})|\n"
lines.insert(insert_index, new_entry)
 
with open(readme_path, 'w') as f:
    f.writelines(lines)

First of all we need to extract the content of the README file & find the correct position to insert the new entry. For that I'll simply iterate over the lines of the README file, match the # of the question ID from the table with the current question ID.

When I get the index for the new entry, I'll insert the new entry using insert(self, index: SupportsIndex, object: _T, /) method. Finally, I'll write the updated content back to the README file.

Automate GitHub Push?

In my opinion this is just an overkill. Let's not be too lazy 😅.

But here is the script for those who want to automate the push to GitHub. We can use the gitpython module to automate the push to GitHub.

Note that you need to install the gitpython module to use this script.

from git import Repo
 
repo = Repo('<path-to-repo>')
origin = repo.remote("origin")  
assert origin.exists()
 
msg = "<your-commit-message>"
 
repo.git.add('.')
repo.git.commit('-m', msg)
repo.git.push("--set-upstream", origin, repo.head.ref)

Read more about the module & advanced usage from GitPython docs.

For Some Other Night

Hey... A little secret about me. I tend to forget things I've got to do. Using wall notes has helped me a lot but I had this idea of creating a Task Management System. To add an element of interest, I'm considering developing it with both CLI and GUI interfaces. And for an extra challenge, WHY NOT MAKE IT IN C ??? 😂😱