Skip to content

python-api.md

Latest commit

 

History

History
119 lines (81 loc) · 2.89 KB

python-api.md

File metadata and controls

119 lines (81 loc) · 2.89 KB

Python API Reference

This page documents the current, maintained Python interfaces for building on fleet-rlm.

Core Runtime Classes

fleet_rlm.core.interpreter.ModalInterpreter

Primary sandbox execution runtime.

Typical usage:

from fleet_rlm import ModalInterpreter

with ModalInterpreter(timeout=600, secret_name="LITELLM", volume_name="rlm-volume-dspy") as interp:
    result = interp.execute("print('hello')")

Key capabilities:

  • lifecycle control (start, shutdown, context managers)
  • sync/async execution (execute, aexecute)
  • execution profile support used by server and delegate workflows

fleet_rlm.core.agent.chat_agent.RLMReActChatAgent

Interactive ReAct orchestration module used by CLI and server chat surfaces.

Key behaviors:

  • document loading and active-alias management
  • command dispatch execution
  • sync/async chat turn helpers
  • streaming event generation for WebSocket clients

Runner Functions (fleet_rlm.runners)

Current maintained runner surface:

  • build_react_chat_agent(...)
  • run_react_chat_once(...)
  • arun_react_chat_once(...)
  • run_long_context(...)
  • check_secret_presence(...)
  • check_secret_key(...)

build_react_chat_agent(...)

Constructs an RLMReActChatAgent with runtime controls such as:

  • ReAct/RLM iteration budgets
  • recursion depth
  • Modal timeout/secret/volume
  • guardrail settings
  • delegate LM settings

run_react_chat_once(...) and arun_react_chat_once(...)

Single-turn wrappers around the interactive ReAct agent.

Common output shape includes:

  • assistant_response
  • optional trajectory metadata (when enabled)
  • turn/session metadata and warnings

run_long_context(...)

Long-document analysis/summarization helper backed by DSPy RLM signatures.

Modes:

  • analyze → AnalyzeLongDocument
  • summarize → SummarizeLongDocument

Secret Diagnostics

  • check_secret_presence(secret_name="LITELLM")
  • check_secret_key(secret_name="LITELLM", key="DSPY_LLM_API_KEY")

These execute Modal-side checks for required environment keys.

Signatures (fleet_rlm.core.agent.signatures)

Current maintained signatures include:

  • AnalyzeLongDocument
  • SummarizeLongDocument
  • ExtractFromLogs
  • GroundedAnswerWithCitations
  • IncidentTriageFromLogs
  • CodeChangePlan
  • CoreMemoryUpdateProposal
  • VolumeFileTreeSignature
  • MemoryActionIntentSignature
  • MemoryStructureAuditSignature
  • MemoryStructureMigrationPlanSignature
  • ClarificationQuestionSignature

Minimal Example

from fleet_rlm.runners import run_long_context

result = run_long_context(
    docs_path="README.md",
    query="Summarize the architecture",
    mode="analyze",
)
print(result["answer"])

Import Verification

uv run python -c "from fleet_rlm.runners import run_long_context, run_react_chat_once"
uv run python -c "from fleet_rlm.core.agent.signatures import AnalyzeLongDocument"