quantecon_wasm has no Timer and no timeit, so qe.Timer() cannot run in a wasm lecture

[mmcky]

quantecon_wasm exposes only the Matlab-like timers. Timer and timeit — the API the lectures have standardised on and the one the style guide mandates — are absent, so qe.Timer() raises AttributeError in a wasm build.

Current state

quantecon_wasm/util/__init__.py ends at:

from .timing import tic, tac, toc, loop_timer

and quantecon_wasm/util/timing.py defines exactly __Timer__, tic, tac, toc, loop_timer. There is no Timer class and no timeit. The package also has no quantecon_wasm/timings/ directory, which upstream Timer depends on for get_default_precision — so this is a two-part port, not a single-file copy.

Why the gap exists

	version	last commit
QuantEcon.py-wasm	v0.7.2-1	2025-01-24
QuantEcon.py	v0.11.4	current

Timer landed upstream in QuantEcon/QuantEcon.py#783 (2025-08-16) and timeit in QuantEcon/QuantEcon.py#794 (2025-08-27) — both after this package's last commit, so it never had the chance to pick them up. This is a concrete instance of the drift that #3 is about.

Why it matters now, not later

The collision is scheduled rather than hypothetical. lecture-wasm imports quantecon_wasm as qe in 4 lectures, and the series mirrors lecture-python-intro, where the style rule qe-code-004 ("use the quantecon.Timer context manager") targets timing magics that are still present in the mirrored copies:

lecture-wasm file	timing magics awaiting conversion
lectures/monte_carlo.md	8 (%%time)
lectures/inequality.md	2 (%%time)
lectures/markov_chains_I.md	2 (%time)

The moment any of those is converted upstream and synced, the wasm build breaks on qe.Timer(). QuantEcon/compliance-lecture-style already carries the recommendation to convert them.

Separately, QuantEcon/QuantEcon.py#833 deprecates tic/tac/toc/loop_timer upstream (step 1 of QuantEcon/QuantEcon.py#786; removal targeted at v1.0). That PR does not affect this package — the fork has its own copy of timing.py with no warning — but when step 2 removes those functions upstream, this package becomes the only place in the org still carrying them, and the two APIs diverge in both directions at once: quantecon will have Timer/timeit and no tic/toc, quantecon_wasm the exact opposite.

Suggested scope

1. Port quantecon/timings/ (__init__.py, timings.py) to quantecon_wasm/timings/.
2. Port Timer and timeit from quantecon/util/timing.py into quantecon_wasm/util/timing.py, and export both from quantecon_wasm/util/__init__.py.
3. Port the corresponding TestTimer / timeit cases into quantecon_wasm/util/tests/test_timing.py.
4. Keep tic/tac/toc/loop_timer in place for now — nothing in wasm calls them, and removing them is a separate decision that should follow upstream's step 2 rather than lead it.

Worth checking during the port whether anything in Timer/timeit relies on behaviour Pyodide does not provide; time.perf_counter is available under Pyodide, but that is the thing to confirm rather than assume.

How this surfaced

An audit of whether QuantEcon/QuantEcon.py#833 would affect the lecture series. It does not — all nine repos in the lectures workspace manifest are clean of the deprecated timers — but the sweep turned up this package as the one place in the org whose timing API cannot follow the lectures. lecture-wasm uses neither API today, so nothing is broken right now.

🤖 Generated with Claude Code