1. 动作优化

1. SummarizeCode动作：用于基于代码进行总结，思考bug、逻辑、todo
  2. CodeReview动作优化：目前强制要求回答问题，有更高的成功率了
2. 数据结构
  1. Document的标准化：Env->Repo->Document，其中Document/Asset/Code都只用Document
    1. 原用于检索的Document改为IndexableDocument
  2. Repo结构引入：用于Document装载与元数据装载
  3. RepoParser引入：写了一个简单的AST parser（后续可能要换tree-sitter），给出了整库symbol
3. 配置优化
  1. 默认更换为gpt-4-1106-preview，以获得最好的效果与成本
  2. 提供~/.metagpt作为配置最高优先级目录，从中读取config.yaml
  3. workspace可以灵活指定了，在config中配置
4. metagpt作为默认命令行，而非python startup.py
  1. 使用新的METAGPT_ROOT生成方式，而非寻找git，以便cli安装
  2. 命令行由fire换为了typer，它会带来相对更好的体验
  3. project_name可以灵活指定了，在metagpt命令行输入中配置
5. 其他
  1. BossRequirement -> UserRequirement
  2. 大量错误文本的修正，增加了可读性
  3. 中量提示词优化，稍微提升了一些准确率
  4. 暂时屏蔽了LongtermMemory相关逻辑，这个逻辑底层调用了langchain的FAISS，会带来~5秒加载耗时
  5. 修复了安装包中的部分描述错误

metagpt/utils/mermaid.py

@@ -10,7 +10,7 @@ import os
 from pathlib import Path
 
 from metagpt.config import CONFIG
-from metagpt.const import PROJECT_ROOT
+from metagpt.const import METAGPT_ROOT
 from metagpt.logs import logger
 from metagpt.utils.common import check_cmd_exists
 
@@ -69,7 +69,7 @@ async def mermaid_to_file(mermaid_code, output_file_without_suffix, width=2048,
             if stdout:
                 logger.info(stdout.decode())
             if stderr:
-                logger.error(stderr.decode())
+                logger.warning(stderr.decode())
     else:
         if engine == "playwright":
             from metagpt.utils.mmdc_playwright import mermaid_to_file
@@ -141,6 +141,6 @@ MMC2 = """sequenceDiagram
 
 if __name__ == "__main__":
     loop = asyncio.new_event_loop()
-    result = loop.run_until_complete(mermaid_to_file(MMC1, PROJECT_ROOT / f"{CONFIG.mermaid_engine}/1"))
-    result = loop.run_until_complete(mermaid_to_file(MMC2, PROJECT_ROOT / f"{CONFIG.mermaid_engine}/1"))
+    result = loop.run_until_complete(mermaid_to_file(MMC1, METAGPT_ROOT / f"{CONFIG.mermaid_engine}/1"))
+    result = loop.run_until_complete(mermaid_to_file(MMC2, METAGPT_ROOT / f"{CONFIG.mermaid_engine}/1"))
     loop.close()

metagpt/utils/token_counter.py

@@ -21,6 +21,7 @@ TOKEN_COSTS = {
     "gpt-4-32k": {"prompt": 0.06, "completion": 0.12},
     "gpt-4-32k-0314": {"prompt": 0.06, "completion": 0.12},
     "gpt-4-0613": {"prompt": 0.06, "completion": 0.12},
+    "gpt-4-1106-preview": {"prompt": 0.01, "completion": 0.03},
     "text-embedding-ada-002": {"prompt": 0.0004, "completion": 0.0},
     "chatglm_turbo": {"prompt": 0.0, "completion": 0.00069}  # 32k version, prompt + completion tokens=0.005￥/k-tokens
 }
@@ -37,6 +38,7 @@ TOKEN_MAX = {
     "gpt-4-32k": 32768,
     "gpt-4-32k-0314": 32768,
     "gpt-4-0613": 8192,
+    "gpt-4-1106-preview": 128000,
     "text-embedding-ada-002": 8192,
     "chatglm_turbo": 32768
 }
@@ -56,16 +58,17 @@ def count_message_tokens(messages, model="gpt-3.5-turbo-0613"):
         "gpt-4-32k-0314",
         "gpt-4-0613",
         "gpt-4-32k-0613",
+        "gpt-4-1106-preview",
     }:
         tokens_per_message = 3
         tokens_per_name = 1
     elif model == "gpt-3.5-turbo-0301":
         tokens_per_message = 4  # every message follows <|start|>{role/name}\n{content}<|end|>\n
         tokens_per_name = -1  # if there's a name, the role is omitted
-    elif "gpt-3.5-turbo" in model:
+    elif "gpt-3.5-turbo" == model:
         print("Warning: gpt-3.5-turbo may update over time. Returning num tokens assuming gpt-3.5-turbo-0613.")
         return count_message_tokens(messages, model="gpt-3.5-turbo-0613")
-    elif "gpt-4" in model:
+    elif "gpt-4" == model:
         print("Warning: gpt-4 may update over time. Returning num tokens assuming gpt-4-0613.")
         return count_message_tokens(messages, model="gpt-4-0613")
     else: