# QuestLog CLI — AI Learning Path Explorer (Click + pandas)

## PURPOSE
The video lists roughly a dozen competing "quest paths" for the Tutorial Island starter
pack and, for each of the six classes, a further pile of courses, cohorts, books, and
certificates. The presenter's own complaint is the paradox of choice. This CLI turns all
of that into one filterable catalog: give it your weekly hours, monthly budget, preferred
learning style, and the class you are aiming at, and it prints a ranked quest log.
It also ships a `classes` command that shows which class each track belongs to, so a
learner can see the Strategist / Rogue / Alchemist / Crafter / Sage / Machinist split.

## WHY THIS MATTERS
It teaches the transcript's two decisions in code. Decision one, Tutorial Island: the
starter pack (understand generative AI, prompt correctly, general tools). Decision two,
pick your class. The catalog encodes the real numbers the video gives — Google Essentials
"sub 10 hours", the 13-hour Google AI Professional Certificate across eight courses,
Andrew Ng's 12–15-hour pair that is free to audit (Pro lab/certificate at $25–30/mo),
the 30–50-hour Elements of AI, the free 10–15-hour Anthropic path, the 2–5-hour OpenAI
foundations, and the ~6-hour Microsoft/LinkedIn track. The ranking makes the trade-offs
visible instead of hidden in a webpage.

## HOW TO BUILD
Stack: **Click + pandas**. Load a catalog (inline constant or a CSV override), turn it
into a DataFrame, apply the filters as vectorized column masks, then compute a fit column
and sort. The essential snippet is the scoring pass:

```python
df["weeks"] = (df["hours"] / max(weeks_available, 1)).round(1)
df["fit"] = 100 - df["weeks"] * 5 - df["usd_month"] * 0.2
ranked = df.sort_values("fit", ascending=False).head(top)
```

Then `click.echo(ranked.to_string(index=False))` prints the log. Add `--csv` so the
catalog can live outside the code.

## WHEN TO USE IN PRODUCTION
Perfect for a bootcamp ops team, a newsletter author, or an L&D analyst who keeps the
catalog in a spreadsheet and needs a repeatable, scriptable recommendation without a
web server. It drops cleanly into CI to flag stale prices. Do NOT use it when the
learner needs interactive exploration (that is the dashboard project) or when the catalog
changes hourly (query the source API live instead of snapshotting a CSV).

## HOW IT HELPS STUDENTS
Students learn Click command design, DataFrame filtering and derived columns, and how to
separate data from logic. It is the fastest path from "spreadsheet" to "real tool".

## STARTER CODE
`starter/main.py` holds the full catalog, `quests` and `classes` commands, the fit
scoring, and `# TODO` markers for JSON/Markdown export and an LLM summary per quest.
