Daily Product Lesson

A small automation that emails me one product-management lesson every morning — a written takeaway plus a few verbatim key quotes — drawn at random from my shelf of product books. Built to turn books I'd already read (but rarely revisited) into a daily, low-effort learning habit.

Built by Larkin Domench as a hands-on product exercise: scoping a real problem, making explicit build-vs-quality tradeoffs, and shipping a working system.

---

The problem

I own several great product/strategy books. I read them once and the ideas faded. I didn't want another app to check or a course to finish — I wanted the good parts to come to me, in small daily doses, with enough of the original text to actually land. Spaced, passive, zero-effort review.

Who it's for

Me, first — but the pattern fits any knowledge worker who wants to compound what they read instead of letting it decay.

Key product decisions (and the tradeoffs)

Decision	Options I weighed	What I chose & why
How lessons are generated	Live LLM each morning vs. a pre-curated bank	Pre-curated bank. Higher quality, $0 and reliable at send time, no runtime dependency. The "intelligence" is front-loaded once.
How much book text to include	Long verbatim excerpts vs. a few short key quotes	Started with long excerpts; switched to 2–4 short key quotes after they felt too heavy to read daily. (Real feedback → real iteration.)
Delivery	In-app/draft vs. real email	Real email via Gmail, so it lands in my inbox with no extra step. Auto-detects SSL/465 when the network blocks 587.
Quote integrity	Trust the generator vs. verify	Built a validator that fails the build if any quote isn't verbatim in its source book — guards against paraphrase/hallucination.
Scheduling	Cloud cron vs. local	Local launchd — the job needs the book files on my machine and no server to pay for.

How it works

Books (PDF/EPUB/MOBI)
   │  extract_books.py            → clean text per book
   ▼
book_text/*.txt
   │  curated once into …         → summary + takeaway + verbatim key quotes
   ▼
lessons.json  ──validate_lessons.py──►  every quote must appear verbatim in its book
   │
   │  launchd @ 7:00 AM daily
   ▼
daily_lesson.py  → random lesson (no repeats until the bank cycles) → Gmail SMTP → inbox

What I learned

• Front-loading quality beats clever runtime. A boring, deterministic daily job + a well-curated dataset is more reliable than a smart job that can fail.
• Verbatim integrity needs a test, not trust. The validator caught real paraphrase errors I'd have shipped.
• Small feedback loops matter. The excerpt-length change came from one round of "this feels heavy" — exactly how a product should evolve.

Run it yourself

Note: this repo intentionally does not include the books or their extracted text (copyrighted). It ships the code, architecture, and sample data so you can point it at your own library.

python3 -m venv .venv && source .venv/bin/activate
pip install pymupdf ebooklib beautifulsoup4 lxml mobi

cp config.example.json config.json     # set your to/from + SMTP
cp lessons.sample.json lessons.json     # or generate your own bank

# add books to a folder, then:
python extract_books.py                 # books → book_text/*.txt
python validate_lessons.py              # verify quotes are verbatim
python daily_lesson.py --dry-run        # preview, no send
python daily_lesson.py                  # send today's lesson

Email auth uses a Gmail app password read from $GMAIL_APP_PASSWORD or ~/.product_lesson_secret. Schedule it with launchd (macOS) or cron.

Files

File	Purpose
daily_lesson.py	Picks a lesson (no repeats) and emails it
extract_books.py	Book files → clean text
validate_lessons.py	Schema + verbatim-quote check
find_passage.py	Helper to locate exact quotes in a book
books.json	Canonical book names → author + text file
config.example.json / lessons.sample.json	Templates (real versions git-ignored)

Tech

Python · PyMuPDF / EbookLib / mobi (multi-format extraction) · Gmail SMTP (SSL) · macOS launchd scheduling.

Testing

validate_lessons.py and find_passage.py have a pytest suite covering the verbatim-quote gate (valid lesson, paraphrase, missing fields, duplicate id, unknown book, too many quotes) and passage lookup. CI runs it on every push.

pip install -r requirements-dev.txt
pytest -v