chore: release v0.3.1

- correctness/stability patch: look-ahead filter, router crash-safety, checkpoint
  identity, crypto sentiment sources, configurable retries, Bedrock API-key auth
- adds Claude Sonnet 5 / Fable 5 support
This commit is contained in:
Yijia-Xiao
2026-07-05 14:29:07 +00:00
parent 0f70af2f31
commit 01477f9afb
4 changed files with 46 additions and 4 deletions

View File

@@ -25,8 +25,10 @@ NVIDIA_API_KEY=
#OPENAI_COMPATIBLE_API_KEY=
# AWS Bedrock (provider "bedrock", install with: pip install ".[bedrock]").
# Auth uses the standard AWS credential chain; set the region (and optionally a
# named profile). No single API key.
# Auth: either a Bedrock API key (bearer token, no AWS access keys) OR the AWS
# credential chain (env keys / ~/.aws/credentials / IAM role / AWS_PROFILE). Set
# the region either way; a bearer token takes precedence when both are present.
#AWS_BEARER_TOKEN_BEDROCK=
#AWS_DEFAULT_REGION=us-west-2
#AWS_PROFILE=
@@ -55,6 +57,10 @@ NVIDIA_API_KEY=
# honor it). Unset leaves each provider at its default. See the README
# "Reproducibility" note — no setting makes LLM output fully deterministic.
#TRADINGAGENTS_TEMPERATURE=0.0
# LLM SDK retry budget forwarded to every provider. Unset leaves each SDK at its
# own default (usually 2). Raise it to ride out bursty 429 rate-limit throttling
# on rate-limited deployments (e.g. Azure OpenAI) instead of aborting the run.
#TRADINGAGENTS_LLM_MAX_RETRIES=6
# Provider-specific reasoning/thinking depth (optional; unset = provider
# default). Setting one also skips the matching interactive prompt.
#TRADINGAGENTS_OPENAI_REASONING_EFFORT=medium