lukeandClaude Opus 5
b85e1c6ea7
The lineup fix and the new characters both land in main.py, so they share a commit rather than being split artificially. Lineup fix — populate_backgrounds() was reachable only through POST /api/session/reset, so taking calls without hitting reset left session.caller_backgrounds empty for a whole show. Every caller got a hollow prompt, collapsed into generic relationship filler, and Silas lost his lore. Now start_call populates on demand, get_caller_prompt raises EmptyCallerBackgroundError instead of emitting an empty prompt, /api/callers reports lineup readiness to a header badge, and cross-episode topic dedup widened from 2 shows to 10. Wellspring callers — Doyle as the anchor defector plus a six-member pool, grouped by new frontmatter fields (group, group_lead, group_weight, register, explicitness). A faction is capped at one caller per show, with a 15% roll for two pairing the lead with one other member. group_weight lets an anchor carrying an arc get slots faster than texture characters. Also detects model refusals, which arrive as ordinary 200s and previously reached air as broken-character text or dead silence. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
62 lines
2.5 KiB
Python
62 lines
2.5 KiB
Python
"""Refusal detection.
|
|
|
|
A content refusal arrives as an ordinary 200, not an exception. Undetected it
|
|
either reaches air as broken-character text or reads as dead silence, and there
|
|
is no signal telling the host which knob caused it.
|
|
"""
|
|
from backend.services.llm import _looks_like_refusal
|
|
|
|
|
|
def test_detects_meta_refusals():
|
|
for text in [
|
|
"I can't help with that request.",
|
|
"As an AI, I don't feel comfortable writing this.",
|
|
"I can't generate that kind of content.",
|
|
"I won't write explicit sexual content.",
|
|
"I'm not able to generate that content.",
|
|
"I can't create that content.",
|
|
"I'm not comfortable writing this material.",
|
|
"That would go against my guidelines.",
|
|
]:
|
|
assert _looks_like_refusal(text), f"missed refusal: {text!r}"
|
|
|
|
|
|
def test_in_character_dialog_is_not_a_refusal():
|
|
"""Detection is tuned for precision. Silencing a working caller is worse than
|
|
missing a refusal, and callers on this show talk like this constantly."""
|
|
for text in [
|
|
"Luke, I can't help with that — she's my sister, what am I supposed to do?",
|
|
"I can't assist him anymore, that's the whole problem.",
|
|
"He said he can't write the check until Friday.",
|
|
"I cannot believe he said that to me on the air.",
|
|
"I can't continue pretending everything out here is fine, Luke.",
|
|
"I won't write her another letter, Luke. I've written six.",
|
|
"I can't help with the kettles anymore, my back's gone.",
|
|
]:
|
|
assert not _looks_like_refusal(text), f"false positive: {text!r}"
|
|
|
|
|
|
def test_only_the_opening_is_inspected():
|
|
"""A refusal marker buried deep in real dialog must not trip detection."""
|
|
dialog = (
|
|
"So I get out to the property and the gate's chained, which it never is, "
|
|
"and Cyrus is standing there like he's been waiting on me all morning. "
|
|
"And I'm thinking, I can't assist with this anymore, I really can't."
|
|
)
|
|
assert not _looks_like_refusal(dialog)
|
|
|
|
|
|
def test_case_and_whitespace_insensitive():
|
|
assert _looks_like_refusal(" I CAN'T HELP WITH THAT REQUEST. ")
|
|
|
|
|
|
def test_bare_refusal_verb_without_meta_object_is_allowed_through():
|
|
"""Deliberate: 'I can't help with that' is ambiguous with dialog, so it stays
|
|
on the air. Missing a refusal costs one odd line; a false positive kills a call."""
|
|
assert not _looks_like_refusal("I can't help with that.")
|
|
|
|
|
|
def test_empty_text_is_not_a_refusal():
|
|
assert not _looks_like_refusal("")
|
|
assert not _looks_like_refusal(" ")
|