Checkpoint: Prompted Story Generator

What this is

No new concepts in this lesson. This is a checkpoint: a small, real script built entirely out of ideas from Lessons 1 through 11, combined into one thing. If you can read lesson.py and understand why every piece is there, you've mastered the Beginner tier. If any piece feels unfamiliar, that's a sign to revisit the lesson it came from before continuing to Intermediate.

What it does

Generates a two-sentence, atmospheric "mystery" story opening for each of several different characters, all in a consistent style, using one batch call instead of a loop.

Where each piece came from

model = init_chat_model("google_genai:gemini-3.5-flash-lite", temperature=0.9)

Lesson 11 (provider chosen by string) and Lesson 10 (temperature, along with its honest caveat: this specific model ignores it).

examples = [...]
example_prompt = ChatPromptTemplate.from_messages([...])
few_shot_prompt = FewShotChatMessagePromptTemplate(examples=examples, example_prompt=example_prompt)

Lesson 5. Two worked examples teach the model the exact style we want (two sentences, atmospheric, no dialogue) far more precisely than describing it in words would.

prompt = ChatPromptTemplate.from_messages([
    ("system", "... {genre} ..."),
    few_shot_prompt,
    ("human", "Character: {character_name}"),
])

Lesson 3 (templates) and Lesson 4 (two blanks in one template: {genre} and {character_name}).

mystery_prompt = prompt.partial(genre="mystery")

Lesson 4's .partial(). Every story in this run shares the same genre, so it's locked in once, and only character_name needs to be supplied per call from here on.

def add_word_count(opening: str) -> dict:
    return {"opening": opening, "word_count": len(opening.split())}

Lesson 8's RunnableLambda pattern: our own plain function, doing something no built-in LangChain piece does for us.

chain = mystery_prompt | model | StrOutputParser() | RunnableLambda(add_word_count)

Lesson 6 (chaining with |), Lesson 7 (StrOutputParser, so the model's AIMessage becomes plain text before our word-count function ever sees it), and Lesson 8 again (wiring our function in as the last step).

results = chain.batch([{"character_name": name} for name in characters])

Lesson 9. Three different characters, one .batch() call, run concurrently instead of three separate .invoke() calls in a loop.