Skip to content

AI-powered virtual executive team — a single coherent executive persona backed by 8 specialist agents (FastAPI + Next.js).

test(onboarding): cover ARR parsing in the business-model step

ninjoana297caa968f6
diff --git a/packages/core/tests/unit/test_onboarding_wizard_arr_parse.py b/packages/core/tests/unit/test_onboarding_wizard_arr_parse.py
new file mode 100644
index 00000000..676a6d47
--- /dev/null
+++ b/packages/core/tests/unit/test_onboarding_wizard_arr_parse.py
@@ -0,0 +1,71 @@
+"""ARR parsing in the onboarding wizard's business-model step.
+
+Regression test for a crash that aborted onboarding entirely. The ARR
+heuristic searched ``\\$?([\\d,]+)\\s*[Mm]``, whose character class matches a
+bare comma with no digits. Any business-model answer containing a comma
+followed by a word starting with M captured ``","``, which
+``.replace(",", "")`` reduced to the empty string, and ``float("")`` raised
+``ValueError`` out of ``build_profile_from_answers``. The wizard returned 500,
+the profile was never written, and every subsequent step failed.
+
+The trigger is an ordinary sentence, not a malformed one — see the parametrised
+cases below.
+"""
+from __future__ import annotations
+
+import pytest
+
+from openexecutive.onboarding.wizard import build_profile_from_answers
+
+
+ "text",
+ [
+ "Consultoría estratégica, marketing y operaciones",
+ "Servicios profesionales, mantenimiento de sistemas",
+ "Vendemos software B2B, modelo de suscripción",
+ "We sell software, mostly to mid-market teams",
+ "Design and build, maintenance included",
+ ],
+)
+def test_comma_before_m_word_does_not_crash(text: str) -> None:
+ """A comma followed by an m-word must not be read as a magnitude."""
+ profile = build_profile_from_answers({"business_model": text})
+
+ assert profile["target_customer"]["profile"] == text
+ assert "annual_revenue_arr" not in profile
+
+
+ ("text", "expected"),
+ [
+ ("We do $2M in ARR", 2_000_000.0),
+ ("Roughly 12M annually", 12_000_000.0),
+ ("ARR is $1,500M across all lines", 1_500_000_000.0),
+ ],
+)
+def test_magnitude_still_parses(text: str, expected: float) -> None:
+ """A real magnitude is still picked up after the narrowing."""
+ profile = build_profile_from_answers({"business_model": text})
+
+ assert profile["annual_revenue_arr"] == expected
+
+
+def test_digits_before_an_m_word_are_not_a_magnitude() -> None:
+ """The word boundary keeps "300 clients, marketing" from meaning $300M.
+
+ Without it the capture starts at a real digit, so the crash guard alone
+ would not help — it would silently record a fabricated ARR instead.
+ """
+ profile = build_profile_from_answers(
+ {"business_model": "We have 300 clients, marketing is word of mouth"}
+ )
+
+ assert "annual_revenue_arr" not in profile
+
+
+def test_no_magnitude_leaves_the_field_unset() -> None:
+ profile = build_profile_from_answers({"business_model": "A boutique consultancy"})
+
+ assert "annual_revenue_arr" not in profile
+ assert profile["target_customer"]["pain_points"] == []