Skip to content

v0.2.0 - Provider-Agnostic LLM, Investigative Agent, Native Timestamps

Choose a tag to compare

View all tags
@ramannanda9 ramannanda9 released this 30 Mar 23:37
· 6 commits to main since this release
v0.2.0
bc85ecf
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Highlights

  • Provider-agnostic LLM interface - LLMProvider ABC with AnthropicProvider (claude-sonnet-4-6) and OpenAIProvider (gpt-4o). Auto-detects from environment.
  • Investigative Agent - InvestigativeAgentPlugin with ReAct loop for Tier 2 root cause analysis. Auto-reflects Python functions into tool schemas.
  • Agent Tools - DataHub lineage/search (real SDK calls) and Slack incident search (async httpx).
  • Native Timestamps - run_ts/executed_at migrated to pa.timestamp("us", tz="UTC") for Iceberg partition pruning and type-safe comparisons.
  • SlackAlertPlugin Hardened - Webhook failures no longer crash the pipeline. Header text truncated to 150-char limit.
  • Performance Benchmarks - 29 benchmarks covering sketch compute, merge, profiling, and storage I/O.

Test Results

  • 83 unit tests passed, 1 skipped
  • 29 benchmarks passed

Breaking Changes

  • Provider modules renamed: anthropic.py -> anthropic_provider.py, openai.py -> openai_provider.py
  • Storage schema changed: run_ts/executed_at columns now pa.timestamp (requires re-ingestion of existing Parquet files)
Assets 2
Loading