fix(quick-router): route "how much is" arithmetic to the calculator

[Yurii214]

Summary

arithmetic_expression strips the calculate / what is lead-ins before the all-math-chars gate, but not how much is — so arithmetic behind that lead-in keeps its question words, fails the gate, and falls through to the LLM. This is inconsistent: how much is 20% of 50 already routes today, because percentage_expression runs first and strips no lead-in.

utterance	on main	expected
how much is 45 times 6	ABSTAIN	calculate{45 * 6}
how much is 100 divided by 4	ABSTAIN	calculate{100 / 4}
how much is 2 plus 2	ABSTAIN	calculate{2 + 2}

The what is / calculate forms already resolve today.

Changes

• Add how much is to the lead-in strip chain in arithmetic_expression, mirroring the existing what is / calculate entries.
• Negatives are unchanged: the two post-strip gates still require a digit AND an operator char AND all-math-chars, so how much is a ticket (no digit) and how much is 5 dollars in euros (no operator) still abstain; how much is 20% of 50 stays on the earlier percentage path.
• Add routes_how_much_is_arithmetic_to_calculate (three phrasings: times / divided by / plus).

Real Behavior Proof

• I have built and run the affected code locally (or noted why I could not).
• I have verified the change end-to-end on Jetson hardware.
• I have NOT verified on Jetson hardware, and I explain the equivalent verification path or validation gap below.

Tested profile / hardware (check all that apply):

• jetson
• raspberry_pi
• portable_sbc
• laptop
• mac
• CI-only / docs-only
• Not run locally

Ran cargo test -p genie-core --no-default-features --lib tools::quick (108 passed) and --test calc_input_test (13 passed); the new routes_how_much_is_arithmetic_to_calculate test fails on main (all three utterances route to None → panic) and passes with the change. cargo clippy and cargo fmt --check are clean. The change is pure text-classification logic with no hardware dependency, so the laptop run exercises the full affected path.

Summary by CodeRabbit

• Bug Fixes

  • Improved routing for natural-language arithmetic requests that start with “how much is …”.
  • Such utterances now correctly route to the calculator tool and handle multiplication, division, and addition as expected.

• Tests

  • Added a regression test to verify “how much is …” arithmetic produces the correct calculator tool calls.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 610806ff-f9d0-4e02-b73a-42fd86c805c8

📥 Commits

Reviewing files that changed from the base of the PR and between 4f3a38c and 96535d1.

📒 Files selected for processing (1)

• crates/genie-core/src/tools/quick.rs

🚧 Files skipped from review as they are similar to previous changes (1)

• crates/genie-core/src/tools/quick.rs

---

📝 Walkthrough

Walkthrough

The arithmetic router recognizes "how much is " prompts and sends them to the deterministic calculator path. Regression coverage verifies multiplication, division, and addition expressions.

Changes

Arithmetic routing

Layer / File(s)	Summary
Prefix routing and regression coverage crates/genie-core/src/tools/quick.rs	The arithmetic parser accepts the "how much is " prefix, and tests verify calculate tool calls with digitized expressions.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Suggested reviewers: matedev01

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: routing “how much is” arithmetic to the calculator.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[matedev01]

LGTM — arithmetic_expression stripped 'calculate '/'what is ' lead-ins but not 'how much is', so that phrasing kept its question words and failed the all-math-chars gate. Verified: clippy -D warnings clean, calc/quick-router tests (130) pass, fmt clean, live BFCL strict_accuracy 96.15% (no regression).