From 842aac82fcda09a6879edfdcf40adfc12b053790 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E8=8E=98=E6=9D=83=20=E9=A9=AC?= Date: Sat, 2 Sep 2023 21:11:44 +0800 Subject: [PATCH] fixbug: summary too long --- metagpt/provider/openai_api.py | 45 ++++++++++++++++++++-------------- 1 file changed, 26 insertions(+), 19 deletions(-) diff --git a/metagpt/provider/openai_api.py b/metagpt/provider/openai_api.py index c08a34f7e..4764b6aad 100644 --- a/metagpt/provider/openai_api.py +++ b/metagpt/provider/openai_api.py @@ -226,38 +226,45 @@ class OpenAIGPTAPI(BaseGPTAPI, RateLimiter): return CONFIG.max_tokens_rsp return get_max_completion_tokens(messages, self.model, CONFIG.max_tokens_rsp) - async def get_summary(self, text: str, max_words=20): + async def get_summary(self, text: str, max_words=200): + max_token_count = DEFAULT_MAX_TOKENS + max_count = 100 + while max_count > 0: + if len(text) < max_token_count: + return await self._get_summary(text, max_words=max_words) + + text_windows = self.split_texts(text, window_size=max_token_count - max_words) + summaries = [] + for ws in text_windows: + response = await self._get_summary(ws, max_words=max_words) + summaries.append(response) + if len(summaries) == 1: + return summaries[0] + + # Merged and retry + text = "\n".join(summaries) + + max_count -= 1 # safeguard + raise openai.error.InvalidRequestError("text too long") + + async def _get_summary(self, text: str, max_words=20): """Generate text summary""" if len(text) < max_words: return text - language = CONFIG.language or DEFAULT_LANGUAGE - command = f"Translate the above content into a {language} summary of less than {max_words} words." + command = f"Translate the above content into a summary of less than {max_words} words." msg = text + "\n\n" + command logger.info(f"summary ask:{msg}") response = await self.aask(msg=msg, system_msgs=[]) logger.info(f"summary rsp: {response}") return response - async def get_context_title(self, text: str, max_token_count_per_ask=None, max_words=5) -> str: + async def get_context_title(self, text: str, max_words=5) -> str: """Generate text title""" - max_response_token_count = 50 - max_token_count = max_token_count_per_ask or CONFIG.MAX_TOKENS or DEFAULT_MAX_TOKENS - while True: - text_windows = self.split_texts(text, window_size=max_token_count - max_response_token_count) - - summaries = [] - for ws in text_windows: - response = await self.get_summary(ws, max_words=max_response_token_count) - summaries.append(response) - if len(summaries) == 1: - return summaries[0] - text = "\n".join(summaries) - if len(text) <= max_words * 2 and len(text) <= max_token_count: - break + summary = await self.get_summary(text, max_words) language = CONFIG.language or DEFAULT_LANGUAGE command = f"Translate the above summary into a {language} title of less than {max_words} words." - summaries.append(command) + summaries = [summary, command] msg = "\n".join(summaries) logger.info(f"title ask:{msg}") response = await self.aask(msg=msg, system_msgs=[])