10 Making Decisions

10.1 The Wall

AI generated a chain of if, elif, and else statements. The logic was wrong. It matched the first condition when it should have matched the third. You could not trace through the conditions to find the bug because you did not understand how Python evaluates them. You rearranged the order and broke something else.

Conditional logic is where AI makes its most subtle mistakes. The code runs, the syntax is valid, but the logic is backwards. If you cannot trace through conditions yourself, you will never catch these bugs.

This chapter fixes that.

10.2 Thinking Session

10.2.1 Getting Oriented

Thinking Session Prompt

Explain Python’s if, elif, and else statements. I want to understand: how does Python decide which branch to take? What happens if multiple conditions are true? And what is the difference between separate if statements and an if/elif chain?

Your AI should explain that elif chains are mutually exclusive, only the first true condition executes. Separate if statements each evaluate independently. This distinction is critical: AI sometimes uses separate if statements when it should use elif, causing multiple branches to fire.

10.2.2 Go Deeper

Thinking Session Prompt

What are common patterns for conditional logic in Python? I see AI using nested if statements, ternary expressions (x if condition else y), and guard clauses. When is each pattern appropriate?

What to Look For

Guard clauses (early returns for special cases) flatten deeply nested code. Ternary expressions are good for simple either/or assignments but become unreadable when nested. If your AI shows deeply nested if/else, ask it to refactor with guard clauses.

Thinking Session Prompt

What values are “truthy” and “falsy” in Python? AI writes things like if items: instead of if len(items) > 0:. Are these the same? When does this shorthand cause problems?

10.2.3 Challenge It

Thinking Session Prompt

What does this code print, and why is it probably a bug?

temperature = 35
if temperature > 30:
    print("Hot")
if temperature > 20:
    print("Warm")
if temperature > 10:
    print("Mild")

What to Look For

It prints “Hot”, “Warm”, and “Mild”, all three. The programmer probably wanted only “Hot”. The fix is elif instead of separate if statements. AI generates this bug when it builds conditions incrementally without considering mutual exclusivity.

10.2.4 What You Should Have Learned

if/elif/else chains are mutually exclusive, only one branch runs
Separate if statements are independent, multiple can run
Guard clauses simplify deeply nested logic
Truthy/falsy: empty collections, 0, None, “” are all falsy
Ternary: x = "yes" if condition else "no"

10.3 The Gap

When you read AI-generated conditional logic, you can now trace the flow. You know whether conditions are mutually exclusive or independent. You can spot when AI should have used elif instead of if, or when a deeply nested block could be flattened with guard clauses.

In the Building Session, you will add categorised responses to your chatbot, different behaviour for greetings, questions, farewells, and default cases.

10.4 Building Session

10.4.1 The Spec

Add response categories to your chatbot:

Greetings: respond when input starts with “hello”, “hi”, “hey”
Questions: respond when input contains “?”
Farewells: respond when input is “bye”, “goodbye”, “quit”
Default: respond to everything else
Use an if/elif/else chain, not separate if statements

10.4.2 Prompt It

Building Session Prompt

Update the get_response() function in my chatbot to v1.0. Add categorised responses using an if/elif/else chain:

Greetings (input starts with “hello”, “hi”, or “hey”): respond warmly
Questions (input contains “?”): respond helpfully
Farewells (“bye” or “goodbye”): respond and signal to exit
Numbers: the existing even/odd check
Default: a catch-all response

The function should return a tuple: (response_text, should_exit). The main loop checks should_exit to know when to break.

Update the main() function to handle the tuple return.

10.4.3 Read the Code

Your AI will produce something like this:

def get_response(user_input):
    """Categorise input and return (response, should_exit)."""
    text = user_input.lower().strip()

    # Check categories in priority order
    if text in ("bye", "goodbye", "quit"):
        return "Goodbye! It was nice chatting.", True
    elif text.startswith(("hello", "hi", "hey")):
        return "Hello! Great to hear from you.", False
    elif "?" in text:
        return "Good question! Let me think...", False
    else:
        try:
            number = int(user_input)
            parity = "even" if number % 2 == 0 else "odd"
            return f"{number} is {parity}.", False
        except ValueError:
            return f"Interesting. '{user_input}'.", False

What to Notice

The elif chain ensures only one category matches. Farewells are checked first, if someone says “bye”, that takes priority over everything else. .startswith() can take a tuple of options. The function returns a tuple (response, should_exit). The boolean flag lets the caller decide what to do.

10.4.4 Stretch It

Building Session Prompt

Add a mood system. Track a mood variable that starts at “neutral”. If the user sends 3 greetings in a row, mood becomes “happy”. If the user sends 3 questions in a row, mood becomes “curious”. Include the mood in the response.

10.5 Your Chatbot So Far

Ch 1-8: Loop, types, memory, formatting, validation, operators, functions
Ch 9: Refactored into functions
Ch 10: Categorised responses with if/elif/else, farewell detection

10.6 Quick Reference

# if/elif/else (mutually exclusive)
if x > 90:
    grade = "A"
elif x > 80:
    grade = "B"
else:
    grade = "C"

# Ternary expression
status = "even" if x % 2 == 0 else "odd"

# Guard clause
def process(data):
    if not data:
        return None
    # main logic here...

# Truthy/falsy checks
if items:        # True if list is not empty
if not name:     # True if string is empty
if count:        # True if not zero

# Making Decisions ## The Wall AI generated a chain of `if`, `elif`, and `else` statements. The logic was wrong. It matched the first condition when it should have matched the third. You could not trace through the conditions to find the bug because you did not understand how Python evaluates them. You rearranged the order and broke something else. Conditional logic is where AI makes its most subtle mistakes. The code runs, the syntax is valid, but the logic is backwards. If you cannot trace through conditions yourself, you will never catch these bugs. This chapter fixes that. ## Thinking Session ### Getting Oriented ::: {.callout-note title="Thinking Session Prompt"} Explain Python's if, elif, and else statements. I want to understand: how does Python decide which branch to take? What happens if multiple conditions are true? And what is the difference between separate if statements and an if/elif chain? ::: Your AI should explain that `elif` chains are mutually exclusive, only the first true condition executes. Separate `if` statements each evaluate independently. This distinction is critical: AI sometimes uses separate `if` statements when it should use `elif`, causing multiple branches to fire. ### Go Deeper ::: {.callout-note title="Thinking Session Prompt"} What are common patterns for conditional logic in Python? I see AI using nested if statements, ternary expressions (x if condition else y), and guard clauses. When is each pattern appropriate? ::: ::: {.callout-tip title="What to Look For"} Guard clauses (early returns for special cases) flatten deeply nested code. Ternary expressions are good for simple either/or assignments but become unreadable when nested. If your AI shows deeply nested if/else, ask it to refactor with guard clauses. ::: ::: {.callout-note title="Thinking Session Prompt"} What values are "truthy" and "falsy" in Python? AI writes things like `if items:` instead of `if len(items) > 0:`. Are these the same? When does this shorthand cause problems? ::: ### Challenge It ::: {.callout-note title="Thinking Session Prompt"} What does this code print, and why is it probably a bug? ```python temperature = 35 if temperature > 30: print("Hot") if temperature > 20: print("Warm") if temperature > 10: print("Mild") ``` ::: ::: {.callout-tip title="What to Look For"} It prints "Hot", "Warm", and "Mild", all three. The programmer probably wanted only "Hot". The fix is `elif` instead of separate `if` statements. AI generates this bug when it builds conditions incrementally without considering mutual exclusivity. ::: ### What You Should Have Learned - `if/elif/else` chains are mutually exclusive, only one branch runs - Separate `if` statements are independent, multiple can run - Guard clauses simplify deeply nested logic - Truthy/falsy: empty collections, 0, None, "" are all falsy - Ternary: `x = "yes" if condition else "no"` ## The Gap When you read AI-generated conditional logic, you can now trace the flow. You know whether conditions are mutually exclusive or independent. You can spot when AI should have used `elif` instead of `if`, or when a deeply nested block could be flattened with guard clauses. In the Building Session, you will add categorised responses to your chatbot, different behaviour for greetings, questions, farewells, and default cases. ## Building Session ### The Spec Add response categories to your chatbot: - Greetings: respond when input starts with "hello", "hi", "hey" - Questions: respond when input contains "?" - Farewells: respond when input is "bye", "goodbye", "quit" - Default: respond to everything else - Use an if/elif/else chain, not separate if statements ### Prompt It ::: {.callout-note title="Building Session Prompt"} Update the get_response() function in my chatbot to v1.0. Add categorised responses using an if/elif/else chain: - Greetings (input starts with "hello", "hi", or "hey"): respond warmly - Questions (input contains "?"): respond helpfully - Farewells ("bye" or "goodbye"): respond and signal to exit - Numbers: the existing even/odd check - Default: a catch-all response The function should return a tuple: (response_text, should_exit). The main loop checks should_exit to know when to break. Update the main() function to handle the tuple return. ::: ### Read the Code Your AI will produce something like this: ```python def get_response(user_input): """Categorise input and return (response, should_exit).""" text = user_input.lower().strip() # Check categories in priority order if text in ("bye", "goodbye", "quit"): return "Goodbye! It was nice chatting.", True elif text.startswith(("hello", "hi", "hey")): return "Hello! Great to hear from you.", False elif "?" in text: return "Good question! Let me think...", False else: try: number = int(user_input) parity = "even" if number % 2 == 0 else "odd" return f"{number} is {parity}.", False except ValueError: return f"Interesting. '{user_input}'.", False ``` ::: {.callout-tip title="What to Notice"} The `elif` chain ensures only one category matches. Farewells are checked first, if someone says "bye", that takes priority over everything else. `.startswith()` can take a tuple of options. The function returns a tuple `(response, should_exit)`. The boolean flag lets the caller decide what to do. ::: ### Stretch It ::: {.callout-note title="Building Session Prompt"} Add a mood system. Track a `mood` variable that starts at "neutral". If the user sends 3 greetings in a row, mood becomes "happy". If the user sends 3 questions in a row, mood becomes "curious". Include the mood in the response. ::: ## Your Chatbot So Far - Ch 1-8: Loop, types, memory, formatting, validation, operators, functions - Ch 9: Refactored into functions - **Ch 10: Categorised responses with if/elif/else, farewell detection** ## Quick Reference ```python # if/elif/else (mutually exclusive) if x > 90: grade = "A" elif x > 80: grade = "B" else: grade = "C" # Ternary expression status = "even" if x % 2 == 0 else "odd" # Guard clause def process(data): if not data: return None # main logic here... # Truthy/falsy checks if items: # True if list is not empty if not name: # True if string is empty if count: # True if not zero ```