fix(model-switch): normalize Unicode dashes from Telegram/iOS input

Telegram on iOS auto-converts double hyphens (--) to em dashes (—)
or en dashes (–) via autocorrect. This breaks /model flag parsing
since parse_model_flags() only recognizes literal '--provider' and
'--global'.

When the flag isn't parsed, the entire string (e.g. 'glm-5.1 —provider zai')
gets treated as the model name and fails with 'Model names cannot
contain spaces.'

Fix: normalize Unicode dashes (U+2012-U+2015) to '--' when they
appear before flag keywords (provider, global), before flag extraction.

The existing test suite in test_model_switch_provider_routing.py
already covers all four dash variants — this commit adds the code
that makes them pass.
This commit is contained in:
Roque 2026-04-10 06:38:27 -06:00 committed by Teknium
parent c6398fcaab
commit 92a23479c0
3 changed files with 64 additions and 0 deletions

View file

@ -6586,6 +6586,11 @@ class GatewayRunner:
import asyncio as _asyncio
args = event.get_command_args().strip()
# Normalize Unicode dashes (Telegram/iOS auto-converts -- to em/en dash)
import re as _re
args = _re.sub(r'[\u2012\u2013\u2014\u2015](days|source)', r'--\1', args)
days = 30
source = None

View file

@ -274,6 +274,11 @@ def parse_model_flags(raw_args: str) -> tuple[str, str, bool]:
is_global = False
explicit_provider = ""
# Normalize Unicode dashes (Telegram/iOS auto-converts -- to em/en dash)
# A single Unicode dash before a flag keyword becomes "--"
import re as _re
raw_args = _re.sub(r'[\u2012\u2013\u2014\u2015](provider|global)', r'--\1', raw_args)
# Extract --global
if "--global" in raw_args:
is_global = True

View file

@ -0,0 +1,54 @@
"""Tests for Unicode dash normalization in /insights command flag parsing.
Telegram on iOS auto-converts -- to em/en dashes. The /insights handler
normalizes these before parsing --days and --source flags.
"""
import re
import pytest
# The regex from gateway/run.py insights handler
_UNICODE_DASH_RE = re.compile(r'[\u2012\u2013\u2014\u2015](days|source)')
def _normalize_insights_args(raw: str) -> str:
"""Apply the same normalization as the /insights handler."""
return _UNICODE_DASH_RE.sub(r'--\1', raw)
class TestInsightsUnicodeDashFlags:
"""--days and --source must survive iOS Unicode dash conversion."""
@pytest.mark.parametrize("input_str,expected", [
# Standard double hyphen (baseline)
("--days 7", "--days 7"),
("--source telegram", "--source telegram"),
# Em dash (U+2014)
("\u2014days 7", "--days 7"),
("\u2014source telegram", "--source telegram"),
# En dash (U+2013)
("\u2013days 7", "--days 7"),
("\u2013source telegram", "--source telegram"),
# Figure dash (U+2012)
("\u2012days 7", "--days 7"),
# Horizontal bar (U+2015)
("\u2015days 7", "--days 7"),
# Combined flags with em dashes
("\u2014days 30 \u2014source cli", "--days 30 --source cli"),
])
def test_unicode_dash_normalized(self, input_str, expected):
result = _normalize_insights_args(input_str)
assert result == expected
def test_regular_hyphens_unaffected(self):
"""Normal --days/--source must pass through unchanged."""
assert _normalize_insights_args("--days 7 --source discord") == "--days 7 --source discord"
def test_bare_number_still_works(self):
"""Shorthand /insights 7 (no flag) must not be mangled."""
assert _normalize_insights_args("7") == "7"
def test_no_flags_unchanged(self):
"""Input with no flags passes through as-is."""
assert _normalize_insights_args("") == ""
assert _normalize_insights_args("30") == "30"