Python Fundamentals

The Python a backend engineer actually gets asked about — written out, explained, and runnable.

---

Not a tutorial series and not a cheat sheet. These are working notes: the Python language features that keep showing up in real backend code and real interviews, each one written down until it was genuinely understood — with the gotchas that bite people in production kept in, not smoothed over.

Most files are deliberately heavy on commented-out experiments. That's the point: the reasoning is the content, not just the final answer.

What's inside

Core Python — the language itself

File	Covers
python_fundamentals.py	Types and built-in collections, comprehensions with conditionals, **kwargs and building URLs from them, a small APIClient class with custom exceptions and __repr__
decorator.py	Writing decorators from scratch — *args/**kwargs forwarding, why functools.wraps matters, and a working @timer
generators_and_terators.py	yield, batch processing a stream, and eager vs lazy — the list comprehension that costs 80MB next to the generator that costs 120 bytes
context_managers.py	__enter__ / __exit__, why guaranteed cleanup matters for DB connections and API sessions, and what with is really doing
collections_modules.py	defaultdict, Counter, OrderedDict, deque, namedtuple — and how they're built on dict's hash table, down to __missing__
advanced_oops.py	Dunder methods (__eq__, __call__, __repr__), @property for controlled access, frozen dataclasses, __slots__, read-only attributes, and the mutable-default-argument trap

A few of the traps covered

• The mutable default argument — one {} created at class-definition time, silently shared by every instance
• functools.wraps, without which your decorated function loses its own __name__
• Why a (x**2 for x in ...) generator and a [x**2 for x in ...] list differ by six orders of magnitude in memory
• Why defaultdict doesn't raise KeyError — it overrides exactly one method

DSA — patterns, not puzzle collections

Each solution is written for the technique, with the problem statement and the reasoning kept inline.

File	Problem	Pattern
two_sum.py	Two Sum	Hash map for complement lookup
valid_paranthese.py	Valid Parentheses	Stack matching
anagram.py	Valid Anagram	Frequency table comparison
group_anagram.py	Group Anagrams	defaultdict keyed by a canonical signature
self_product.py	Product of Array Except Self	Prefix and suffix products, no division
347_top_k_elements.py	Top K Frequent Elements	Counting plus a heap

Running it

No dependencies — standard library only.

git clone https://github.com/DeepakAtariya/python-fundamentals.git
cd python-fundamentals

Both folders use an index.py as a scratch entry point: uncomment the import for the module you want and run it.

python -m core_python.index    # core_python/index.py selects the module
cd dsa && python index.py      # dsa/index.py does the same

Or just run a file directly:

python core_python/decorator.py

Much of the code is intentionally commented out so you can uncomment one block at a time and watch what changes. Reading top to bottom works better than running it all.

Repo map

core_python/
  python_fundamentals.py       types, comprehensions, kwargs, a small API client
  decorator.py                 decorators from scratch, @timer
  generators_and_terators.py   yield, laziness, memory
  context_managers.py          __enter__ / __exit__
  collections_modules.py       defaultdict, Counter, deque, namedtuple + internals
  advanced_oops.py             dunders, @property, dataclasses, __slots__
  index.py                     entry point — uncomment what you want to run
dsa/
  two_sum.py  valid_paranthese.py  anagram.py
  group_anagram.py  self_product.py  347_top_k_elements.py
  index.py                     entry point

Related

LLM engineering notes — tokenization, embeddings, the chat completion API and a provider-agnostic wrapper — live in a separate repo:

→ Learn LLM with Atariya

Contributing

If an explanation here is wrong or unclear, that's worth an issue. Corrections welcome.

License

MIT © Deepak Atariya