🌀 Core Python Projects :

A collection of small, focused Python projects — each one built to go deep on a specific set of core language concepts, the way production code is written: typed, error-handled, and runnable from the command line.

This isn't a "100 Python exercises" repo. Each project is developed by myself to learn and implemnt python concepts properly. Every line earns its place, no boilerplate, no unused imports, no half-finished functions.

"Beautiful is better than ugly. Explicit is better than implicit. Simple is better than complex." — The Zen of Python

---

Projects

1. Log Analyzer — 01_log_analyzer.py

A command-line tool that streams a log file and summarizes it: severity-level counts, the time range covered, and the top recurring messages.

Why it's interesting:

• Generators — read_lines() uses yield to process the file one line at a time. A 5-line sample log and a 5 GB production log run with the same memory footprint.
• Decorators — @timer wraps analyse() and reports how long it took, using functools.wraps so the wrapped function keeps its original name and docstring.
• Named regex groups — (?P<level>\w+) lets the code read match["level"] instead of match.group(2), which stays readable even if the pattern grows.
• Counter.most_common() — aggregates severity levels and repeated messages without manual dictionary bookkeeping.

Run it:

python 01_log_analyzer.py data1_sample.log

Sample output:

---

2. CSV Data Analyzer — 02_csv_data_Analyzer.py

Reads a CSV and computes summary statistics without pandas or numpy — to show the statistics actually understood, not just imported.

Why it's interesting:

• csv.DictReader — each row becomes a dict keyed by header, so columns are addressed by name (row["Groceries"]) instead of brittle index numbers.
• for...else (an underused Python feature) — get_numeric_columns() tries to convert every value in a column to float. If any value fails, break skips the else, and the whole column is excluded from numeric stats. One bad value correctly disqualifies the column — no partial/garbage averages.
• Mode via Counter, not statistics.mode — for continuous data (prices, spending amounts) every value is often unique, so statistics.mode would return a misleading "first value seen." Here, mode is only reported if a value actually repeats — otherwise it prints n/a.
• Categorical detection — any non-numeric column with a small number of unique values (≤ 20) gets a frequency breakdown automatically.

Run it:

python 02_csv_data_Analyzer.py data2_customer_spending.csv

Sample output:

---

3. OOP Bank System — 03_oop_bank_system.py

A small banking system that's actually about object-oriented design, not just "classes that hold data."

Why it's interesting:

• Abstract Base Class (ABC + @abstractmethod) — Account.withdraw() is declared but not implemented. Python refuses to let you instantiate Account directly — you're forced to define a subclass that decides what "withdraw" means.
• Polymorphism, not just inheritance — SavingsAccount enforces a ₹1,000 minimum balance; CurrentAccount allows overdraft up to ₹5,000. Same method name (withdraw), same call site (account.withdraw(amount)), completely different rules — the calling code never needs to know which type it has.
• Custom exception hierarchy — InsufficientFundsError is separated from Python's built-in ValueError. This means "you typed a negative number" (ValueError) and "your account rules don't allow this" (InsufficientFundsError) can be caught and handled differently.
• Encapsulation — Bank owns a private _accounts dict. Account lookup, creation, and listing all go through Bank's methods; nothing outside the class touches the dict directly.
• Dunder methods — Transaction.__str__ and Account.__str__ give every object a sensible human-readable representation, used automatically by print().

Run it:

python 03_oop_bank_system.py

Then use the interactive menu to open an account, deposit, withdraw, transfer between accounts, and view statements. Try withdrawing below the ₹1,000 minimum on a savings account, or beyond the ₹5,000 overdraft on a current account, to see InsufficientFundsError caught and reported cleanly.

Sample Output:

---

4. Text File Search Engine — 04_text_file_search_Engine.py

A miniature search engine: builds an inverted index over .txt files in a folder, then ranks files by how many query words they contain — the same core idea behind real search systems, just without the scale.

Why it's interesting:

• Inverted index with defaultdict(set) — instead of "file → words it contains," the index is "word → set of files containing it." This is the data structure that makes search fast: looking up a word is O(1), not a scan of every file.
• Stopword filtering — common words ("the", "is", "and") are stripped during tokenizing. Without this, every file would score highly on every query just from shared filler words, making the ranking meaningless.
• Set comprehensions + regex — tokenise() does lowercasing, word extraction, and stopword filtering in a single readable line.
• Relevance scoring — results show score/total query words matched, giving a sense of how relevant a match is, not just a raw count.

Run it:

python 04_text_file_search_Engine.py .

This indexes data3_AI_Ethics.txt in the repo root. Try searching for terms like ethics regulation or data privacy and you'll see the file and its match score.

Sample Output:

---

5. Web Scraper — 05_web_scrapper.py

Fetches book listings from books.toscrape.com — a sandbox site built specifically for scraping practice — parses them with BeautifulSoup, and saves the results as JSON.

Why it's interesting:

• Defensive HTTP — fetch() sets a User-Agent, a timeout, and calls raise_for_status() so a 404/500 response raises immediately instead of silently returning a useless page. All of it wrapped in try/except requests.RequestException.
• Compound CSS selectors — item.select_one("p.instock.availability") matches an element with both CSS classes. (find(class_="instock availability") — the more common but incorrect approach — doesn't reliably match multi-class attributes in BeautifulSoup.)
• urljoin — converts relative links (catalogue/a-book/index.html) into full, usable URLs relative to the page they were found on.
• Dataclass → JSON — Book is a @dataclass; asdict() converts each instance to a dict for json.dump, with ensure_ascii=False so currency symbols like £ are written as real characters, not \u00a3 escapes.

Run it:

pip install requests beautifulsoup4
python 05_web_scrapper.py

Prints the first 5 books with price and stock status, then saves all 20 to report.json.

Sample Output:

---

Fun Project: Dream Catcher — DreamCatcher.py

A "just for fun" script — built with matplotlib and numpy. It draws a dream catcher, and asks you for your dreams — placing each one as text inside the web, using a collision-avoidance algorithm so they don't overlap, falling back to a golden-angle spiral layout if the circle gets crowded. Type in a few dreams, hit done, and watch them appear inside the web.

python DreamCatcher.py

Output:

---

Setup

All projects use Python 3.10+ and the standard library, except the web scraper and the bonus Dream Catcher:

pip install requests beautifulsoup4 matplotlib numpy

---

It's a small reminder that this repo isn't just about proving proficiency — it's about a genuine love for what Python lets you build, from log parsers to dream catchers. Every project above started as "let me see if I can build this," and that curiosity is the thing worth carrying into whatever comes next.