fix: lazy-load NLTK resources to eliminate startup overhead

[Tech4Aditya]

📝 Description

Fixes the startup overhead caused by unconditional nltk.download() calls at module import time.

Replaced the two bare top-level calls:

nltk.download('punkt')
nltk.download('stopwords')

with a _ensure_nltk_resources() helper that uses nltk.data.find() to check if each resource is already present before downloading. The helper is called lazily inside analyze_resume(), so downloads only happen when the function is actually invoked — and only if the resource is missing.

Changes made:

• Removed nltk.download('punkt') and nltk.download('stopwords') from module level
• Added _ensure_nltk_resources() with LookupError guard for punkt, punkt_tab, and stopwords
• Called _ensure_nltk_resources() as the first line of analyze_resume()

🔗 Linked Issue

Closes #1000

---

📋 Contribution Checklist

• My code strictly adheres to the project guidelines (Standard library only, procedural code, clean spacing).
• I have verified that my files are placed in the correct directory.
• I have tested my changes thoroughly on my local machine.
• I have included interactive emojis and clean console/UI outputs.
• GSSoC 2026: I have been formally assigned to this issue and noted it above.

[vercel]

@Tech4Aditya is attempting to deploy a commit to the Anuj's projects Team on Vercel.

A member of the Team first needs to authorize it.

[Tech4Aditya]

The CI failure in modal.spec.js is unrelated to this PR . it tests
focus-trapping behaviour in the web-apps modal component, which I haven't
touched. This test appears to be failing on main independently of these changes.

[github-actions]

🎉 Thank you for your contribution!

Your Pull Request has been merged successfully.

We appreciate the time and effort you put into improving this project. Contributions like yours help the repository grow and stay useful for everyone.

If you'd like to contribute again, please check the open issues and make sure you are assigned before opening another Pull Request.

Thanks again for your support! 🙌