Files
aiturk-hermes-ide/apps/desktop/assets/codex-bridge/test_model_gecis.py
T
yilsemandClaude Opus 5.5 c8993d58cc feat(codex-bridge): web search for added models, forgiving spoken model names
Added models had no web search: Codex's built-in one runs on OpenAI's side and
the bridge drops it from their tool list. Measured in a voice session on
11 Oct 2026: a model looked for a search tool ten times and ended the turn
with no answer. The bridge now tells added models two ready commands (ara.py:
search and read a page, standard library only, no key). Codex's own models
never see the note: their requests do not reach the bridge.

Spoken model names: a half-spoken word matches the start of a name part
("think" -> thinking, "obli" -> obliterated), and when one candidate is the
others minus extra words, the one without the unsaid suffix is picked
("Qwen 3.8 Obliterated" -> not the ":thinking" one).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-11 07:25:30 +03:00

171 lines
11 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
"""Bridge 8: "modeli X yap" is handled by the bridge; a per-chat route survives until the user picks another model."""
import tempfile
import unittest
import model_gecis as g
MODELS = [("gpt-6-sol", "GPT-6-Sol"), ("gpt-6-astra", "GPT-6-Astra"), ("gpt-6.1-sol", "GPT-6.1-Sol"),
("ide/9router-ag-gemini-3.8-flash-low", "agy/gemini-3.8-flash-low"),
("ide/9router-ag-gemini-3.8-flash-high", "agy/gemini-3.8-flash-high"),
("ide/9router-cc-claude-opus-5-5", "cc/opus-5-5"), ("ide/9router-cc-claude-opus-5-5-nt", "cc/opus-5-5 (no-think)"),
("ide/9router-nano-z-ai-glm-5.3-flash-uncensored", "nano/glm-5.3-flash-uncensored"),
("ide/agnes-agnes-2.0-flash", "agnes/agnes-2.0-flash"), ("ide/agnes-agnes-2.5-flash", "agnes/agnes-2.5-flash")]
def user(text):
return {"type": "message", "role": "user", "content": [{"type": "input_text", "text": text}]}
def voice(text):
return user(f"<realtime_delegation>\n <input>{text}</input>\n <transcript_delta>user: x</transcript_delta>\n</realtime_delegation>")
def pick(text):
command = g.komut_coz(text)
return g.eslestir(command["ad"], MODELS) if command and command["tur"] == "gec" else command
class Komut(unittest.TestCase):
def test_spoken_names_resolve_to_one_model(self):
for text, slug in [("modeli gemini flash low yap", "ide/9router-ag-gemini-3.8-flash-low"),
("Modeli Gemini 3.8 Flash High'a geçir", "ide/9router-ag-gemini-3.8-flash-high"),
("modeli klod opus 5 5 yap", "ide/9router-cc-claude-opus-5-5"),
("modeli opus 5 5'e geç", "ide/9router-cc-claude-opus-5-5"),
("modeli gpt 6 astra yap", "gpt-6-astra"), ("model geç gpt 6.1 sol", "gpt-6.1-sol"),
("modeli glm flaş sansürsüz yap", "ide/9router-nano-z-ai-glm-5.3-flash-uncensored")]:
self.assertEqual((pick(text) or {}).get("slug"), slug, text)
def test_half_spoken_words_and_the_plain_variant(self):
"""Measured 11 Oct 2026 in a voice session: "obli think" was not found three times, and "Qwen 3.8
Obliterated" asked which one although the user had not said "thinking"."""
qwen = MODELS + [("ide/9router-nano-qwen-qwen3.8-27b-obliterated", "nano/qwen3.8-27b-obliterated"),
("ide/9router-nano-qwen-qwen3.8-27b-obliterated-thinking", "nano/qwen3.8-27b-obliterated:thinking"),
("ide/9router-nano-qwen-qwen3.8-27b-uncensored", "nano/qwen3.8-27b-uncensored")]
def sec(text):
return (g.eslestir(g.komut_coz(text)["ad"], qwen) or {}).get("slug")
thinking = "ide/9router-nano-qwen-qwen3.8-27b-obliterated-thinking"
self.assertEqual(sec("Modeli 3.8 Obliterated Think yap"), thinking)
self.assertEqual(sec("Modeli Qwen... modeli Qwen 3.8 .27 B... Obli... Think yap"), thinking)
self.assertEqual(sec("Modeli Qwen ... 3.8 Obliterated yap"), "ide/9router-nano-qwen-qwen3.8-27b-obliterated")
self.assertEqual(sec("modeli klod opus yap"), "ide/9router-cc-claude-opus-5-5") # not the "(no-think)" one
self.assertIn("adaylar", g.eslestir(g.komut_coz("modeli qwen 3.8 yap")["ad"], qwen)) # obliterated or uncensored: ask
def test_ambiguous_name_lists_candidates_and_changes_nothing(self):
self.assertEqual(pick("modeli agnes yap"), {"adaylar": ["agnes/agnes-2.0-flash", "agnes/agnes-2.5-flash"]})
self.assertIn("adaylar", pick("modeli flash yap")) # several unrelated Flash models: ask
self.assertIn("adaylar", pick("modeli sol yap")) # GPT-6-Sol or GPT-6.1-Sol: ask
self.assertIn("adaylar", pick("modeli gpt 999 astra yap")) # words fit, version does not: ask
def test_negations_and_general_questions_are_not_commands(self):
for text in ("modeli gemini flash low yapma", "modeli değiştirme opus kalsın", "GPT 6 Astra modeli için test yap",
"hangi model daha ucuz", "en iyi model hangisi", "modeli seçili dosyaya göre gemini flash low yap"):
result = pick(text)
self.assertFalse(result and ("slug" in result or result.get("tur") == "sor"), text)
self.assertEqual(g.komut_coz("hangi modelsin"), {"tur": "sor"})
def test_ordinary_sentences_are_not_commands(self):
for text in ("Masaüstünde kaç klasör var", "bu modeli kullanarak bir site yap", "yapay zeka modeli nedir",
"hangi modeli önerirsin bu iş için bana söyle bakalım lütfen",
"model " + "çok uzun bir cümle " * 5 + "yap"):
self.assertFalse(pick(text) and "slug" in pick(text), text)
self.assertEqual(g.komut_coz("seçili model hangisi"), {"tur": "sor"})
self.assertEqual(g.komut_coz("bu chatte ana model kim"), {"tur": "sor"})
def test_voice_delegation_yields_only_the_spoken_input(self):
text, index, spoken = g.son_kullanici_metni({"input": [user("a"), voice("modeli opus yap")]})
self.assertEqual((text, index, spoken), ("modeli opus yap", 1, True))
self.assertEqual(g.son_kullanici_metni({"input": [user("<skill>x</skill>")]})[0], None)
self.assertEqual(g.son_kullanici_metni({"input": [{"type": "function_call_output", "output": "x"}]})[0], None)
class TurOrtasi(unittest.TestCase):
def test_message_after_tool_items_is_mid_turn_and_after_final_answer_is_not(self):
final = {"type": "message", "role": "assistant", "phase": "final_answer", "content": "ok"}
call = {"type": "function_call", "call_id": "c", "name": "shell", "arguments": "{}"}
out = {"type": "function_call_output", "call_id": "c", "output": "x"}
self.assertFalse(g.tur_ortasi({"input": [user("a"), final, user(".")]}, 2))
self.assertTrue(g.tur_ortasi({"input": [user("a"), call, out, user(".")]}, 3))
self.assertTrue(g.tur_ortasi({"input": [user("a"), call, out, voice("bekliyorum"), user(".")]}, 4))
self.assertTrue(g.tur_ortasi({"input": [user("a"), {"type": "message", "role": "assistant", "phase": "commentary",
"content": "bakıyorum"}, user(".")]}, 2))
self.assertFalse(g.tur_ortasi({"input": [user(".")]}, 0))
def test_handover_note_replaces_only_the_spoken_input(self):
req = {"input": [voice("modeli opus yap")]}
g.metni_degistir(req, 0, g.devir_notu("cc/opus-5-5", "GPT-6-Sol"))
text = req["input"][0]["content"][0]["text"]
self.assertIn("<transcript_delta>", text)
self.assertIn("cc/opus-5-5 olarak devam ediyorum", text)
self.assertNotIn("modeli opus yap</input>", text)
class Gecmis(unittest.TestCase):
def test_handled_commands_in_history_become_notes_but_the_last_message_is_left_alone(self):
req = {"input": [user("modeli gemini flash low yap"), user("Masaüstünde kaç dosya var"),
voice("modeli opus 5 5 yap"), user("modeli gpt 6 astra yap")]}
self.assertEqual(g.gecmisi_temizle(req, MODELS), 2)
self.assertIn("agy/gemini-3.8-flash-low olarak değiştirdi", req["input"][0]["content"][0]["text"])
self.assertEqual(req["input"][1]["content"][0]["text"], "Masaüstünde kaç dosya var")
self.assertIn("<transcript_delta>", req["input"][2]["content"][0]["text"])
self.assertEqual(req["input"][3]["content"][0]["text"], "modeli gpt 6 astra yap")
self.assertEqual(g.gecmisi_temizle(req, MODELS), 0)
class Bilgi(unittest.TestCase):
def test_model_is_told_once_that_the_chat_model_changes_by_saying_so(self):
req = {"input": [user("selam")]}
self.assertTrue(g.bilgi_ekle(req))
self.assertEqual(req["input"][0]["role"], "developer")
self.assertIn("modeli <ad> yap", req["input"][0]["content"][0]["text"])
self.assertFalse(g.bilgi_ekle(req))
self.assertEqual(len(req["input"]), 2)
self.assertFalse(g.bilgi_ekle({"input": []}))
def test_a_command_without_a_name_asks_which_model(self):
for text in ("modeli değiştir", "başka modele geç", "modeli değiştirir misin"):
self.assertEqual(g.komut_coz(text), {"tur": "nasil"}, text)
self.assertIsNone(g.komut_coz("modeli değiştirme"))
class Dogal(unittest.TestCase):
GECIT = [("ide/9router/ModelARK/dola-seed-2-1-turbo-260628", "9router/ModelARK/dola-seed-2-1-turbo-260628"),
("ide/9router/ModelARK/seed-2-0-pro-260328", "9router/ModelARK/seed-2-0-pro-260328")]
def test_command_without_the_word_model_and_a_spelled_out_name(self):
for text in ("Seed iki nokta bir turbo yap", "Evet. Seed, S E E D iki nokta bir turbo yap"):
command = g.komut_coz(text)
self.assertEqual((command["ad"], command["kesin"]), (["seed", "2", "1", "turbo"], False), text)
self.assertEqual(g.eslestir(command["ad"], self.GECIT)["slug"], "ide/9router/ModelARK/dola-seed-2-1-turbo-260628")
self.assertTrue(g.komut_coz("Modeli Seed iki nokta bir turbo yap")["kesin"])
self.assertEqual(g.eslestir(g.komut_coz("gemini flash low'a geç")["ad"], MODELS)["slug"],
"ide/9router-ag-gemini-3.8-flash-low")
def test_ordinary_requests_ending_in_a_verb_are_not_commands(self):
for text in ("test yap", "kahve yap", "yapma"):
self.assertIsNone(g.komut_coz(text), text)
for text in ("şu dosyayı düzelt ve commit yap", "rapor hazırla gemini flash low ile karşılaştırma yap"):
command = g.komut_coz(text)
self.assertFalse(command["kesin"])
self.assertFalse((g.eslestir(command["ad"], MODELS) or {}).get("slug"), text)
class Yon(unittest.TestCase):
def test_route_lasts_until_the_user_picks_another_model_in_the_list(self):
with tempfile.TemporaryDirectory() as home:
y = g.Yonlendirme(home)
y.koy("chat", "gpt-6-sol", "ide/x", simdi=1000)
self.assertEqual(g.Yonlendirme(home).al("chat", "gpt-6-sol", simdi=1001), "ide/x") # survives a restart
self.assertIsNone(y.al("other", "gpt-6-sol", simdi=1001))
self.assertIsNone(y.al("chat", "gpt-6-astra", simdi=1002)) # user picked another
self.assertIsNone(y.al("chat", "gpt-6-sol", simdi=1003)) # and it stays cleared
y.koy("chat", "gpt-6-sol", "ide/x", simdi=1000)
y.koy("chat", "gpt-6-sol", "gpt-6-sol", simdi=1001) # back to the picked one
self.assertIsNone(y.al("chat", "gpt-6-sol", simdi=1002))
y.koy("old", "a", "b", simdi=1000)
self.assertIsNone(y.al("old", "a", simdi=1000 + g.OMUR_SN + 1))
if __name__ == "__main__":
unittest.main()