Syntax, indentation, and comments

Indentation is the block structure

Most languages mark blocks with braces. Python uses indentation, so whitespace is part of the grammar.

Rules that matter:

• A colon : ends the line that opens a block (if, for, while, def, class, try, with).
• Everything in the block is indented by the same amount.
• Four spaces per level is the universal convention. Never mix tabs and spaces — Python 3 rejects it with TabError.
• The block ends when indentation returns to the outer level.

Two errors you will see:

• IndentationError: expected an indented block — you opened with a colon but did not indent.
• IndentationError: unindent does not match any outer indentation level — the levels are inconsistent, usually mixed tabs and spaces from copy-paste.

Configure your editor to insert spaces when Tab is pressed. That one setting removes an entire class of errors.

Statements and line length

One statement per line, no semicolons needed. Python accepts a = 1; b = 2 but nobody writes that.

Long lines break naturally inside brackets (), [], {} — no continuation character required. This is why function calls with many arguments are written one argument per line.

The backslash \ also continues a line, but it is fragile (a trailing space after it breaks the file). Prefer brackets.

PEP 8 suggests 79 characters; most teams settle on 88 or 100 and let a formatter such as ruff or black enforce it.

Comments and docstrings

# starts a comment to end of line. Python has no block comment syntax.

A docstring is a string literal as the first statement of a module, function, or class. Unlike a comment it is stored on the object and readable at runtime through help() or __doc__. Editors show it on hover, and API documentation is generated from it.

Comment the why, not the what. # add 1 to i is noise. # offset by 1 because the API is 1-indexed is worth keeping.

Naming conventions carry meaning in Python:

Copy-paste examples

Copy into your own editor, or load one into the compiler below and press Run.

score = 0.85

if score > 0.7:
    print("high confidence")
    print("still inside the if block")
print("outside the block — always runs")

for i in range(3):
    if i % 2 == 0:
        print(i, "even")
    else:
        print(i, "odd")

config = {
    "model": "gpt-4.1-mini",
    "temperature": 0.2,
    "max_tokens": 512,
}

def train(
    model_name,
    epochs=3,
    learning_rate=2e-5,
    batch_size=16,
):
    return f"{model_name}: {epochs} epochs at lr={learning_rate}"

print(config)
print(train("bert-base", epochs=5))

def cosine_similarity(a, b):
    """Return the cosine similarity between two equal-length vectors.

    Args:
        a: First vector as a list of floats.
        b: Second vector as a list of floats.

    Returns:
        A float between -1.0 and 1.0.
    """
    # guard against a zero vector, which would divide by zero
    dot = sum(x * y for x, y in zip(a, b))
    norm_a = sum(x * x for x in a) ** 0.5
    norm_b = sum(y * y for y in b) ** 0.5
    if norm_a == 0 or norm_b == 0:
        return 0.0
    return dot / (norm_a * norm_b)

print(round(cosine_similarity([1, 2, 3], [2, 4, 6]), 4))
print(cosine_similarity.__doc__.splitlines()[0])

MAX_TOKENS = 4096          # constant

class PromptBuilder:       # class: PascalCase
    def __init__(self, system_prompt):
        self.system_prompt = system_prompt   # public attribute
        self._call_count = 0                 # internal, leave alone

    def build(self, question):               # method: snake_case
        self._call_count += 1
        return f"{self.system_prompt}\n\nQ: {question}"

builder = PromptBuilder("Be concise.")
print(builder.build("What is Python?"))
print("MAX_TOKENS:", MAX_TOKENS)

Takeaways

• Indentation is syntax: four spaces per level, never mix tabs and spaces.
• Break long lines inside brackets rather than with a backslash.
• Docstrings are readable at runtime; comments explain why, not what.