ed/manual_slop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

REGRESSSIOSSSOONNNNSSSS

Browse Source
This commit is contained in:
Edward R. Gonzalez
2026-03-06 21:39:50 -05:00
parent b88fdfde03
commit 12dba31c1d
3 changed files with 95 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files

100
src/ai_client.py
View File

@@ -254,28 +254,35 @@ def _classify_gemini_error(exc: Exception) -> ProviderError:
def _classify_deepseek_error(exc: Exception) -> ProviderError:
body = ""
if isinstance(exc, requests.exceptions.HTTPError) and exc.response is not None:
body = exc.response.text.lower()
try:
# Try to get the detailed error from DeepSeek's JSON response
err_data = exc.response.json()
if "error" in err_data:
body = str(err_data["error"].get("message", exc.response.text))
else:
body = exc.response.text
except:
body = exc.response.text
else:
body = str(exc).lower()
body = str(exc)
if "429" in body or "rate" in body:
return ProviderError("rate_limit", "deepseek", exc)
if "401" in body or "403" in body or "auth" in body or "api key" in body:
return ProviderError("auth", "deepseek", exc)
if "402" in body or "balance" in body or "billing" in body:
return ProviderError("balance", "deepseek", exc)
if "quota" in body or "limit exceeded" in body:
return ProviderError("quota", "deepseek", exc)
if "connection" in body or "timeout" in body or "network" in body:
return ProviderError("network", "deepseek", exc)
body_l = body.lower()
if "429" in body_l or "rate" in body_l:
return ProviderError("rate_limit", "deepseek", Exception(body))
if "401" in body_l or "403" in body_l or "auth" in body_l or "api key" in body_l:
return ProviderError("auth", "deepseek", Exception(body))
if "402" in body_l or "balance" in body_l or "billing" in body_l:
return ProviderError("balance", "deepseek", Exception(body))
if "quota" in body_l or "limit exceeded" in body_l:
return ProviderError("quota", "deepseek", Exception(body))
if "connection" in body_l or "timeout" in body_l or "network" in body_l:
return ProviderError("network", "deepseek", Exception(body))
if "400" in body or "bad request" in body:
# Try to wrap the original error with the response body for better debugging
if body:
new_exc = Exception(f"Bad Request (400): {body}")
return ProviderError("unknown", "deepseek", new_exc)
# If we have a body for a 400 error, wrap it
if "400" in body_l or "bad request" in body_l:
return ProviderError("unknown", "deepseek", Exception(f"DeepSeek Bad Request: {body}"))
return ProviderError("unknown", "deepseek", exc)
return ProviderError("unknown", "deepseek", Exception(body))
def set_provider(provider: str, model: str) -> None:
global _provider, _model
@@ -393,7 +400,7 @@ def _list_anthropic_models() -> list[str]:
raise _classify_anthropic_error(exc) from exc
def _list_deepseek_models(api_key: str) -> list[str]:
return ["deepseek-chat", "deepseek-reasoner", "deepseek-v3", "deepseek-r1"]
return ["deepseek-chat", "deepseek-reasoner"]
TOOL_NAME: str = "run_powershell"
@@ -1417,7 +1424,7 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
"Content-Type": "application/json",
}
is_reasoner = _model == "deepseek-reasoner"
is_reasoner = _model in ("deepseek-reasoner", "deepseek-r1")
# Update history following Anthropic pattern
with _deepseek_history_lock:
@@ -1432,12 +1439,41 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
for round_idx in range(MAX_TOOL_ROUNDS + 2):
current_api_messages: list[dict[str, Any]] = []
with _deepseek_history_lock:
for msg in _deepseek_history:
current_api_messages.append(msg)
sys_msg = {"role": "system", "content": f"{_get_combined_system_prompt()}\n\n<context>\n{md_content}\n</context>"}
current_api_messages.insert(0, sys_msg)
# DeepSeek R1 (Reasoner) can be extremely strict about the 'system' role.
# For maximum compatibility, we'll only use 'system' for non-reasoner models.
if not is_reasoner:
sys_msg = {"role": "system", "content": f"{_get_combined_system_prompt()}\n\n<context>\n{md_content}\n</context>"}
current_api_messages.append(sys_msg)
with _deepseek_history_lock:
for i, msg in enumerate(_deepseek_history):
# Create a clean copy of the message for the API
role = msg.get("role")
api_msg = {"role": role}
content = msg.get("content")
if i == 0 and is_reasoner:
# Prepend system instructions to the first user message for R1
content = f"System Instructions:\n{_get_combined_system_prompt()}\n\nContext:\n{md_content}\n\n---\n\n{content}"
if role == "assistant":
# OpenAI/DeepSeek: content MUST be a string if tool_calls is absent
# If tool_calls is present, content can be null
if msg.get("tool_calls"):
api_msg["content"] = content or None
api_msg["tool_calls"] = msg["tool_calls"]
else:
api_msg["content"] = content or ""
if msg.get("reasoning_content"):
api_msg["reasoning_content"] = msg["reasoning_content"]
elif role == "tool":
api_msg["content"] = content or ""
api_msg["tool_call_id"] = msg.get("tool_call_id")
else:
api_msg["content"] = content or ""
current_api_messages.append(api_msg)
request_payload: dict[str, Any] = {
"model": _model,
@@ -1445,9 +1481,13 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
"stream": stream,
}
if stream:
request_payload["stream_options"] = {"include_usage": True}
if not is_reasoner:
request_payload["temperature"] = _temperature
request_payload["max_tokens"] = _max_tokens
# DeepSeek max_tokens is for the output, clamp to 8192 which is their hard limit for V3/Chat
request_payload["max_tokens"] = min(_max_tokens, 8192)
tools = _get_deepseek_tools()
if tools:
request_payload["tools"] = tools
@@ -1482,6 +1522,10 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
continue
try:
chunk = json.loads(chunk_str)
if not chunk.get("choices"):
if chunk.get("usage"):
current_usage = cast(dict[str, Any], chunk["usage"])
continue
delta = cast(dict[str, Any], chunk.get("choices", [{}])[0].get("delta", {}))
if delta.get("content"):
content_chunk = cast(str, delta["content"])
@@ -1533,7 +1577,8 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
full_assistant_text = thinking_tags + assistant_text
with _deepseek_history_lock:
msg_to_store: dict[str, Any] = {"role": "assistant", "content": assistant_text}
# DeepSeek/OpenAI: If tool_calls are present, content can be null but should usually be present
msg_to_store: dict[str, Any] = {"role": "assistant", "content": assistant_text or None}
if reasoning_content:
msg_to_store["reasoning_content"] = reasoning_content
if tool_calls_raw:
@@ -1583,7 +1628,6 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
tool_results_for_history.append({
"role": "tool",
"tool_call_id": call_id,
"name": name,
"content": truncated,
})
_append_comms("IN", "tool_result", {"name": name, "id": call_id, "output": out})

12
src/api_hooks.py
View File

@@ -40,6 +40,15 @@ class HookServerInstance(ThreadingHTTPServer):
super().__init__(server_address, RequestHandlerClass)
self.app = app
def _serialize_for_api(obj: Any) -> Any:
if hasattr(obj, "to_dict"):
return obj.to_dict()
if isinstance(obj, list):
return [_serialize_for_api(x) for x in obj]
if isinstance(obj, dict):
return {k: _serialize_for_api(v) for k, v in obj.items()}
return obj
class HookHandler(BaseHTTPRequestHandler):
"""Handles incoming HTTP requests for the API hooks."""
def do_GET(self) -> None:
@@ -184,7 +193,8 @@ class HookHandler(BaseHTTPRequestHandler):
try:
gettable = _get_app_attr(app, "_gettable_fields", {})
for key, attr in gettable.items():
result[key] = _get_app_attr(app, attr, None)
val = _get_app_attr(app, attr, None)
result[key] = _serialize_for_api(val)
finally: event.set()
lock = _get_app_attr(app, "_pending_gui_tasks_lock")
tasks = _get_app_attr(app, "_pending_gui_tasks")