DocChatAgent: huge speedup, cost-reduction using RelevanceExtractorAgent Added global setting `quiet` to suppress all output, using ctx mgr `temporary_settings`
0.1.100
Better imports: `import langroid as lr` and expose key classes at top level
0.1.99
Minor fixes to batch convenience functions.
0.1.98
Convenience methods for batch/async/concurrent running of tasks, agent methods. See tests/main/test_batch.py
0.1.97
* convenience fn to run batch of tasks on list of inputs, async/concurrently (see tests/main/test_chat_agent_async.py) * bool option AgentConfig.show_stats to control showing token usage/cost stats.
0.1.96
Fix some edge cases in LLM config. Make OpenAI GPT4 default in Agent.base