Verbosity levels & improved JSON parsing (#83)

* added different verbosity levels, better json parsing

* fix typing error

vision_agent/agent/vision_agent_v2.py

@@ -51,6 +51,21 @@ def extract_code(code: str) -> str:
  return code
 
 
+def extract_json(json_str: str) -> Dict[str, Any]:
+ try:
+ json_dict = json.loads(json_str)
+ except json.JSONDecodeError:
+ if "```json" in json_str:
+ json_str = json_str[json_str.find("```json") + len("```json") :]
+ json_str = json_str[: json_str.find("```")]
+ elif "```" in json_str:
+ json_str = json_str[json_str.find("```") + len("```") :]
+ # get the last ``` not one from an intermediate string
+ json_str = json_str[: json_str.find("}```")]
+ json_dict = json.loads(json_str)
+ return json_dict # type: ignore
+
+
 def write_plan(
  chat: List[Dict[str, str]],
  plan: Optional[List[Dict[str, Any]]],
@@ -65,8 +80,8 @@ def write_plan(
  context = USER_REQ_CONTEXT.format(user_requirement=user_requirements)
  prompt = PLAN.format(context=context, plan=str(plan), tool_desc=tool_desc)
  chat[-1]["content"] = prompt
- plan = json.loads(model.chat(chat).replace("```", "").strip())
- return plan["user_req"], plan["plan"] # type: ignore
+ new_plan = extract_json(model.chat(chat))
+ return new_plan["user_req"], new_plan["plan"]
 
 
 def write_code(
@@ -133,7 +148,7 @@ def debug_code(
  {"role": "system", "content": DEBUG_SYS_MSG},
  {"role": "user", "content": prompt},
  ]
- code_and_ref = json.loads(model.chat(messages).replace("```", "").strip())
+ code_and_ref = extract_json(model.chat(messages))
  if hasattr(model, "kwargs"):
  del model.kwargs["response_format"]
  return extract_code(code_and_ref["improved_impl"]), code_and_ref["reflection"]
@@ -149,7 +164,7 @@ def write_and_exec_code(
  exec: Execute,
  retrieved_ltm: str,
  max_retry: int = 3,
- verbose: bool = False,
+ verbosity: int = 0,
 ) -> Tuple[bool, str, str, Dict[str, List[str]]]:
  success = False
  counter = 0
@@ -159,6 +174,9 @@ def write_and_exec_code(
  user_req, subtask, retrieved_ltm, tool_info, orig_code, model
  )
  success, result = exec.run_isolation(code)
+ if verbosity == 2:
+ _CONSOLE.print(Syntax(code, "python", theme="gruvbox-dark", line_numbers=True))
+ _LOGGER.info(f"\tCode success: {success}, result: {str(result)}")
  working_memory: Dict[str, List[str]] = {}
  while not success and counter < max_retry:
  if subtask not in working_memory:
@@ -180,11 +198,11 @@ def write_and_exec_code(
  )
  success, result = exec.run_isolation(code)
  counter += 1
- if verbose:
+ if verbosity == 2:
  _CONSOLE.print(
  Syntax(code, "python", theme="gruvbox-dark", line_numbers=True)
  )
- _LOGGER.info(f"\tDebugging reflection, result: {reflection}, {result}")
+  _LOGGER.info(f"\tDebugging reflection: {reflection}, result: {result}")
 
  if success:
  working_memory[subtask].append(
@@ -204,7 +222,7 @@ def run_plan(
  code: str,
  tool_recommender: Sim,
  long_term_memory: Optional[Sim] = None,
- verbose: bool = False,
+ verbosity: int = 0,
 ) -> Tuple[str, str, List[Dict[str, Any]], Dict[str, List[str]]]:
  active_plan = [e for e in plan if "success" not in e or not e["success"]]
  current_code = code
@@ -235,7 +253,7 @@ def run_plan(
  tool_info,
  exec,
  retrieved_ltm,
- verbose=verbose,
+ verbosity=verbosity,
  )
  if task["type"] == "code":
  current_code = code
@@ -244,11 +262,11 @@ def run_plan(
 
  working_memory.update(working_memory_i)
 
- if verbose:
+ if verbosity == 1:
  _CONSOLE.print(
  Syntax(code, "python", theme="gruvbox-dark", line_numbers=True)
  )
- _LOGGER.info(f"\tCode success, result: {success}, {str(result)}")
+ _LOGGER.info(f"\tCode success: {success} result: {str(result)}")
 
  task["success"] = success
  task["result"] = result
@@ -283,23 +301,23 @@ def __init__(
  timeout: int = 600,
  tool_recommender: Optional[Sim] = None,
  long_term_memory: Optional[Sim] = None,
- verbose: bool = False,
+ verbosity: int = 0,
  ) -> None:
- self.planner = OpenAILLM(temperature=0.1, json_mode=True)
- self.coder = OpenAILLM(temperature=0.1)
+ self.planner = OpenAILLM(temperature=0.0, json_mode=True)
+ self.coder = OpenAILLM(temperature=0.0)
  self.exec = Execute(timeout=timeout)
  if tool_recommender is None:
  self.tool_recommender = Sim(TOOLS_DF, sim_key="desc")
  else:
  self.tool_recommender = tool_recommender
- self.verbose = verbose
+ self.verbosity = verbosity
  self._working_memory: Dict[str, List[str]] = {}
  if long_term_memory is not None:
  if "doc" not in long_term_memory.df.columns:
  raise ValueError("Long term memory must have a 'doc' column.")
  self.long_term_memory = long_term_memory
  self.max_retries = 3
- if self.verbose:
+ if self.verbosity:
  _LOGGER.setLevel(logging.INFO)
 
  def __call__(
@@ -355,7 +373,7 @@ def chat_with_workflow(
  working_code,
  self.tool_recommender,
  self.long_term_memory,
- self.verbose,
+ self.verbosity,
  )
  success = all(task["success"] for task in plan)
  working_memory.update(working_memory_i)