The Python She Runs On

# labs/hello_glados.py
def greet(subject_name: str) -> str:
    return f"Hello, {subject_name}. The facility is now online."

print(greet("Test Subject"))

$ uv run python labs/hello_glados.py
Hello, Test Subject. The facility is now online.

Look at where the printing happens: the last line, outside the function. greet builds the text and hands it back; the print is a caller deciding what to do with it. The type hints (: str in, -> str out) document the contract and let your editor catch a wrong argument before the file ever runs.

# labs/hello_glados.py
name = "GLaDOS"
status = "ACTIVE"
subject_count = 1

def greet(subject_name: str) -> str:
    return f"Hello, {subject_name}. The facility is now online."

def report_status() -> None:
    print(f"System: {name}")
    print(f"Status: {status}")
    print(f"Subjects logged: {subject_count}")

print(greet("Test Subject"))
report_status()

$ uv run python labs/hello_glados.py
Hello, Test Subject. The facility is now online.
System: GLaDOS
Status: ACTIVE
Subjects logged: 1

report_status prints on purpose, and its return type says so: -> None announces there is nothing to hand back, because printing the dashboard is the function's whole job. That contrast is the real lesson of this stage. Both kinds of function are valid; you just have to know which one you are writing, and the return type is where you say it.

# labs/hello_glados.py — full file
name = "GLaDOS"
status = "ACTIVE"
subject_count = 1

def greet(subject_name: str) -> str:
    return f"Hello, {subject_name}. The facility is now online."

def report_status() -> None:
    print(f"System: {name}")
    print(f"Status: {status}")
    print(f"Subjects logged: {subject_count}")

def build_message(speaker: str, text: str) -> str:
    return f"[{speaker}]: {text}"

def main() -> None:
    print(greet("Test Subject"))
    report_status()
    msg = build_message(name, "Interesting. You're still alive.")
    print(msg)

if __name__ == "__main__":
    main()

$ uv run python labs/hello_glados.py
Hello, Test Subject. The facility is now online.
System: GLaDOS
Status: ACTIVE
Subjects logged: 1
[GLaDOS]: Interesting. You're still alive.

build_message is the most reusable function in the file: two arguments in, one formatted string out, no state touched. A pattern you will write dozens of times. And the demo calls now live inside main() behind the guard, so the file has become two things at once — a runnable script, and a toolbox other files can import without triggering the demo.

# labs/config_demo.py
from dataclasses import dataclass

@dataclass
class VoiceConfig:
    sample_rate: int = 22050
    voice_model: str = "glados-v1"
    volume: float = 0.8

def describe(cfg: VoiceConfig) -> str:
    return f"{cfg.voice_model} at {cfg.sample_rate} Hz, volume {cfg.volume:.0%}"

if __name__ == "__main__":
    cfg = VoiceConfig()
    print(describe(cfg))
    loud = VoiceConfig(volume=1.0)
    print(describe(loud))

$ uv run python labs/config_demo.py
glados-v1 at 22050 Hz, volume 80%
glados-v1 at 22050 Hz, volume 100%

Every field has a name, a type and a default; overriding one at construction leaves the rest alone. When chapter 3 needs a sample rate and chapter 7 needs a voice model, they will read them from an object like this instead of from scattered globals, and when volume 2 moves settings into files on disk, the dataclass is what those files load into. The 22050 is not arbitrary either; you will find out what a sample rate is, physically, two chapters from now.