From 39f6a851ddd3dd3fedf733f65d70151501c8b7bd Mon Sep 17 00:00:00 2001 From: Liang Jiaqing Date: Thu, 12 Feb 2026 22:52:47 +0800 Subject: [PATCH] refactor: optimize ClaudeSession context trimming with tag compression, fix tool_use parsing to use last match --- sidercall.py | 87 +++++++++++++++++++++++++++++----------------------- 1 file changed, 49 insertions(+), 38 deletions(-) diff --git a/sidercall.py b/sidercall.py index cb23370..804786d 100644 --- a/sidercall.py +++ b/sidercall.py @@ -26,47 +26,30 @@ class SiderLLMSession: full_text = self._core.chat(prompt, model, stream=False) if stream: return iter([full_text]) # gen有奇怪的空回复或死循环行为，sider足够快 return full_text - -class GeminiSession: - def __init__(self, api_key=None, default_model="gemini-2.0-flash-001", proxy=proxy): - self.api_key = api_key or google_api_key - if not self.api_key: raise ValueError("google_api_key 未配置或为空，请在 mykey.py 中设置") - self.default_model = default_model - self.proxies = {"http":proxy, "https":proxy} if proxy else None - def ask(self, prompt, model=None, stream=False): - if model is None: model = self.default_model - url = f"https://generativelanguage.googleapis.com/v1/models/{model}:generateContent?key={self.api_key}" - headers = {"Content-Type":"application/json"} - data = {"contents":[{"role":"user","parts":[{"text":prompt}]}]} - try: - kw = {"headers":headers, "json":data, "timeout":60, 'proxies': self.proxies} - r = requests.post(url, **kw) - except Exception as e: - return f"[GeminiError] request failed: {e}" - if r.status_code != 200: - body = r.text[:500].replace("\n"," ") - return f"[GeminiError] HTTP {r.status_code}: {body}" - try: - obj = r.json(); cands = obj.get("candidates") or [] - if not cands: return "[GeminiError] empty candidates" - parts = (cands[0].get("content") or {}).get("parts") or [] - full_text = "".join(p.get("text","") for p in parts) - except Exception as e: - return f"[GeminiError] invalid response format: {e}" - return iter([full_text]) if stream else full_text - + class ClaudeSession: - def __init__(self, api_key, api_base, model="claude-opus", context_win=24000): + def __init__(self, api_key, api_base, model="claude-opus", context_win=12000): self.api_key, self.api_base, self.default_model, self.context_win = api_key, api_base.rstrip('/'), model, context_win self.raw_msgs, self.lock = [], threading.Lock() def _trim_messages(self, messages): - total = sum(len(m['prompt'])//4 for m in messages) - if total <= self.context_win: return messages - target, current, result = self.context_win * 0.9, 0, [] + # 压缩4轮前的assistant消息：truncate / 块 + for i, msg in enumerate(messages): + if i < len(messages) - 4 and 'orig' not in msg: + msg['orig'] = msg['prompt'] + for tag in ('thinking', 'tool_use', 'tool_result'): + msg['prompt'] = re.sub( + rf'(<{tag}>)([\s\S]*?)()', + lambda m: m.group(1) + (m.group(2)[:200] + '...') + m.group(3) if len(m.group(2)) > 200 else m.group(0), + msg['prompt'] + ) + total = sum(len(m['prompt']) for m in messages) + if total <= self.context_win * 4: return messages + target, current, result = self.context_win * 4 * 0.9, 0, [] for msg in reversed(messages): - if (msg_len := len(msg['prompt'])//4) + current <= target: + if (msg_len := len(msg['prompt'])) + current <= target: result.append(msg); current += msg_len else: break + if current > 10000 * 4: print(f'[DEBUG] Whole context length {current//4}.') return result[::-1] or messages[-2:] def raw_ask(self, messages, model=None, temperature=0.5, max_tokens=4096): model = model or self.default_model @@ -195,6 +178,34 @@ class LLMSession: if stream: return _ask_gen() return ''.join(list(_ask_gen())) + +class GeminiSession: + def __init__(self, api_key=None, default_model="gemini-2.0-flash-001", proxy=proxy): + self.api_key = api_key or google_api_key + if not self.api_key: raise ValueError("google_api_key 未配置或为空，请在 mykey.py 中设置") + self.default_model = default_model + self.proxies = {"http":proxy, "https":proxy} if proxy else None + def ask(self, prompt, model=None, stream=False): + if model is None: model = self.default_model + url = f"https://generativelanguage.googleapis.com/v1/models/{model}:generateContent?key={self.api_key}" + headers = {"Content-Type":"application/json"} + data = {"contents":[{"role":"user","parts":[{"text":prompt}]}]} + try: + kw = {"headers":headers, "json":data, "timeout":60, 'proxies': self.proxies} + r = requests.post(url, **kw) + except Exception as e: + return f"[GeminiError] request failed: {e}" + if r.status_code != 200: + body = r.text[:500].replace("\n"," ") + return f"[GeminiError] HTTP {r.status_code}: {body}" + try: + obj = r.json(); cands = obj.get("candidates") or [] + if not cands: return "[GeminiError] empty candidates" + parts = (cands[0].get("content") or {}).get("parts") or [] + full_text = "".join(p.get("text","") for p in parts) + except Exception as e: + return f"[GeminiError] invalid response format: {e}" + return iter([full_text]) if stream else full_text class MockFunction: def __init__(self, name, arguments): @@ -291,11 +302,11 @@ class ToolClient: tool_calls = None tool_pattern = r"(.*?)" - tool_match = re.search(tool_pattern, remaining_text, re.DOTALL) + tool_all = re.findall(tool_pattern, remaining_text, re.DOTALL) json_str = "" - if tool_match: - json_str = tool_match.group(1).strip() + if tool_all: + json_str = tool_all[-1].strip() remaining_text = re.sub(tool_pattern, "", remaining_text, flags=re.DOTALL) elif '' in remaining_text: weaktoolstr = remaining_text.split('')[-1].strip() @@ -317,7 +328,7 @@ class ToolClient: if func_name: tool_calls = [MockToolCall(func_name, args)] except json.JSONDecodeError as e: print("[Warn] Failed to parse tool_use JSON:", json_str) - tool_calls = [MockToolCall('bad_json', {'msg': f'Failed to parse tool_use JSON: {str(e)}'})] + tool_calls = [MockToolCall('bad_json', {'msg': f'Failed to parse tool_use JSON: {json_str[:200]}'})] except Exception as e: print("[Error] Exception during tool_use parsing:", str(e), data)