DataArena
Week 1 • Day 4100% FREE

Day 4:Control Flow (`if/else`), Loops (`for`/`while`), Functions & Variable Scope

Battle-tested daily lesson note • 100% runnable code, interview Q&A, test suite & interactive quiz

Week 1 — Day 4: Control Flow (if/else), Loops (for/while), Functions & Variable Scope

Mentor's Note (10+ YOE Staff AI Engineer POV): Today you learn how to give your Python programs decision-making brains (if/elif/else), repetition power (for / while loops), and reusable building blocks (def functions).

Why Senior AI Engineers care deeply about Day 4:

  1. Every AI Agent Loop (while step < max_steps:) is just a while loop with a circuit-breaker condition so a confused LLM doesn't burn $500 calling tools infinitely!
  2. We cover the #1 most famous Python coding interview trap of all time: The Mutable Default Argument Bug (def add_message(msg, history=[]):)—a bug that has caused catastrophic data leaks between different users' chat sessions in real AI startups!

Part 1: Control Flow (if / elif / else) & Python "Truthiness"

1.1 Basic Decision Making

python
confidence_score = 0.88 if confidence_score >= 0.85: action = "RETURN_DIRECT_ANSWER" elif confidence_score >= 0.50: action = "ASK_CLARIFYING_QUESTION" else: action = "ESCALATE_TO_HUMAN_SRE"

1.2 "Truthiness" and "Falsiness" in Python

In Python, you don't need to write if len(messages) > 0: or if api_key != "":. Every object in Python is automatically either Truthy or Falsy:

  • Falsy values (evaluate to False in an if check):
    • False, None, 0, 0.0, "" (empty string), [] (empty list), {} (empty dict), set() (empty set).
  • Truthy values: Everything else!
python
# Clean, Pythonic check: if not retrieved_chunks: print("No relevant documents found in vector database!")

Part 2: Loops (for and while), enumerate(), and zip()

2.1 for Loops with enumerate() and zip()

Beginners often write clunky for i in range(len(chunks)):. Professional Python engineers use enumerate() and zip():

python
chunks = ["## Redis Cache", "## Postgres Failover"] scores = [0.94, 0.71] # 1. enumerate() gives you BOTH the index number and the value! for idx, chunk in enumerate(chunks, start=1): print(f"Chunk #{idx}: {chunk}") # 2. zip() pairs items from two lists together side-by-side! for chunk, score in zip(chunks, scores): print(f"{chunk} -> similarity={score}")

2.2 while Loops, break, and continue (The AI Agent Retry Loop!)

Whenever we call an LLM API or run an autonomous Agent, we use a while loop with a maximum retry/step cap (break):

python
attempt = 0 max_retries = 3 while attempt < max_retries: attempt += 1 # If API call succeeds, exit the loop immediately with `break`: if api_succeeded: break

Part 3: Functions (def), *args, **kwargs & Variable Scope (LEGB)

3.1 Anatomy of a Clean Function

python
def calculate_chunk_overlap( chunk_size: int, overlap_ratio: float = 0.10 ) -> int: """Calculates how many tokens should overlap between adjacent RAG chunks.""" return int(chunk_size * overlap_ratio)

3.2 What Are *args and **kwargs? (Used Everywhere in PyTorch & FastAPI!)

  • *args (Positional Arguments Tuple): Collects any number of extra positional arguments into a tuple.
  • **kwargs (Keyword Arguments Dictionary): Collects any number of named key=value arguments into a dict (e.g., passing temperature=0.0, top_p=0.9 into an LLM wrapper!).

Part 4: THE #1 PYTHON INTERVIEW TRAP — Mutable Default Arguments!

[!CAUTION] Real-World AI Startup Disaster Story: A junior developer writes this function to append a user's message to a chat history list:

python
# BUGGY CODE — NEVER DO THIS! def append_chat_message( user_msg: str, history: list[str] = [] ) -> list[str]: history.append(user_msg) return history

What goes wrong? In Python, default argument values (history=[]) are evaluated ONLY ONCE when the def statement is first loaded into memory—NOT each time the function is called! That means User A calls append_chat_message("My secret password is 123") →\rightarrow returns ["My secret password is 123"]. One second later, User B calls append_chat_message("Hello") →\rightarrow returns ["My secret password is 123", "Hello"]! User A's private chat just leaked to User B because they shared the exact same [] list in memory!

The Safe, Professional Pattern (None Sentinel):

python
def append_chat_message( user_msg: str, history: list[str] | None = None ) -> list[str]: if history is None: history = [] # Creates a brand-new list on EVERY call! history.append(user_msg) return history

Part 5: Hands-On Lab — Rule-Based SRE Incident Router & Safe Chat Session Manager (incident_router.py)

Check out src/incident_router.py (in Tab 2: Codes — click Run in Browser to execute) and run the script and tests:

powershell
python week01_computer_terminal_git_python_basics/day04_control_flow_loops_functions/src/incident_router.py python -m unittest discover -s week01_computer_terminal_git_python_basics/day04_control_flow_loops_functions/tests -v
Week 1 Day 4: Week 1 — Day 4: Control Flow (`if/else`), Loops (`for`/`while`), Functions & Variable Scope | 6-Month AI Engineer Roadmap · DataArena