core.runner¶
- class celi_framework.core.runner.CELIConfig(job_description, tool_implementations, llm_cache, primary_model_name, max_tokens, model_url, simulate_live=False, token_budget=0, sequential=False, force_tool_every_n=5)
Bases:
object-
force_tool_every_n:
int= 5
-
job_description:
JobDescription
-
llm_cache:
bool
-
max_tokens:
int
-
model_url:
Optional[str]
-
primary_model_name:
str
-
sequential:
bool= False
-
simulate_live:
bool= False
-
token_budget:
int= 0
-
tool_implementations:
Optional[ToolImplementations]
-
force_tool_every_n:
- celi_framework.core.runner.enable_llm_caching(simulate_live=False)¶
Enables LLM caching for all calls using the celi_framework.utils.llms functions.
If simulate_live is set to true, imposes a delay on each call to simulate live calls. Useful when testing timing or for demos.
- celi_framework.core.runner.get_celi_llm_cache()¶
- celi_framework.core.runner.run_celi(celi_config)¶
- async celi_framework.core.runner.run_process_runner(celi_config)¶