Saving Your Index

The Librarian's Notebook

At the end of the day the library closes and the librarian goes home. Everything she knew -- which words are in which books -- lives in her head (the computer's memory), and memory is wiped when the program stops.

So before she leaves, she writes every book into a notebook: a JSON file on disk. Tomorrow she opens the notebook, files every book again, and answers every question exactly as she did yesterday.

from pysearch.engine import SearchEngine
from pysearch.persistence import load_engine, save_engine

engine = SearchEngine()
engine.add("cats", "Cats purr and nap in the sun")
save_engine(engine, "library.json")

tomorrow = load_engine("library.json")
tomorrow.search("purr")   # [("cats", ...)]

Why Save the Books, Not the Index?

The notebook keeps each book's original text, not the index built from it. Rebuilding the index on load is quick, and it means a saved file still works even if we improve how the index is built later.

The file also records a format number. If a future version ever saves things differently, an old program refuses the new file instead of misreading it.

Saving Safely

Saving writes a fresh copy beside the old file and then swaps it in, in one instant step. A crash halfway through leaves yesterday's notebook untouched, never half of each.

When the Notebook Is Damaged

If the file is missing, isn't text at all, isn't valid JSON, or isn't a saved search engine, load_engine raises a PersistenceError that names the file. It never quietly hands back an empty library.