Commit Graph
19 Commits
Author SHA1 Message Date
dmahan93 59b53f0a23 fix: skip tests when atroposlib/minisweagent unavailable in CI
- test_agent_loop_tool_calling.py: import atroposlib at module level
  to trigger skip (environments.agent_loop is now importable without
  atroposlib due to __init__.py graceful fallback)
- test_modal_sandbox_fixes.py: skip TestToolResolution tests when
  minisweagent not installed
2026-03-11 06:52:55 -07:00
dmahan93 d198a647e2 fix: guard all atroposlib imports for CI without atropos installed
- environments/__init__.py: try/except on atroposlib imports so
  submodules like tool_call_parsers remain importable standalone
- test_agent_loop.py, test_tool_call_parsers.py,
  test_managed_server_tool_support.py: skip at module level when
  atroposlib is missing
2026-03-11 06:52:55 -07:00
dmahan93 0f53275169 test: skip atropos-dependent tests when atroposlib not installed
Guard all test files that import from environments/ or atroposlib
with try/except + pytest.skip(allow_module_level=True) so they
gracefully skip instead of crashing when deps aren't available.
2026-03-11 06:52:55 -07:00
dmahan93 366de72a38 add a local vllm instance 2026-03-11 06:52:55 -07:00
dmahan93 13f5459670 fix: use ManagedServer for vLLM in TBLite eval + local_vllm config
TBLite eval was bypassing ManagedServer and calling ServerManager
directly, which uses /v1/chat/completions — not available on the
atropos vllm_api_server (/generate only).

Now uses _use_managed_server() to detect vLLM/SGLang backends and
route through ManagedServer (Phase 2) with proper tool_parser and
/generate endpoint. Falls back to Phase 1 for OpenAI endpoints.

Also adds local_vllm.yaml config for running against a local vLLM
server with Docker sandboxes.
2026-03-11 06:52:55 -07:00
dmahan93 93333387d6 fix: handle dict and object tool_calls in agent loop
vLLM's ToolCallTranslator returns tool_calls as dicts, while
OpenAI API returns them as objects with .id, .function.name etc.
Normalize both formats in the agent loop.
2026-03-11 06:52:26 -07:00
dmahan93 1f9e7cd659 test: 5 vLLM integration tests + fallback tool call parser
Tests hit a real vLLM server (Qwen/Qwen3-4B-Thinking-2507) via
ManagedServer Phase 2. Auto-skip if server isn't running.

Tests verify:
- Single tool call through full agent loop
- Multi-tool calls across turns
- ManagedServer produces SequenceNodes with tokens/logprobs
- Direct response without tools
- Thinking model produces <think> blocks

Also adds fallback parser in agent_loop.py: when ManagedServer's
ToolCallTranslator can't parse (vLLM not installed), hermes-agent's
standalone parsers extract <tool_call> tags from raw content.
2026-03-11 06:52:26 -07:00
dmahan93 09fc64c6b6 add eval output to gitignore 2026-03-11 06:52:26 -07:00
dmahan93 84147f4d81 refactor: update to new atropos tool-calling API
Migrate from old tool_call_parser (instance) to new ToolCallTranslator
pattern from atropos add-openai-endpoint-for-managed-server branch:

- Set tool_parser on ServerManager (string name, e.g. 'hermes')
- Use managed_server(tokenizer=..., preserve_think_blocks=...)
  instead of managed_server(tokenizer=..., tool_call_parser=instance)
- ManagedServer now handles tool call translation internally via
  ToolCallTranslator (bidirectional raw text <-> OpenAI tool_calls)
- Remove old parser loading code (get_parser/KeyError fallback)

The hermes-agent tool_call_parsers/ directory is preserved as a
standalone fallback for environments that don't use vLLM's parsers.
2026-03-11 06:52:26 -07:00
dmahan93 ee4b20b55b test: 9 agent loop tool-calling integration tests
Real LLM calls via OpenRouter using stepfun/step-3.5-flash:free (zero cost).
Falls back to paid models if free model is unavailable.

Tests: single tool call, multi-tool single turn, multi-turn chains,
unknown tool rejection, max_turns limit, direct response (no tools),
tool error handling, AgentResult structure, conversation history.
2026-03-11 06:52:26 -07:00
dmahan93 ed27b826c5 feat: add eval_concurrency limit + Docker local config for TBLite
- Add eval_concurrency config field with asyncio.Semaphore
- Add local.yaml config using Docker backend (sandboxed, no cloud costs)
- Register docker_image alongside modal_image for backend flexibility
- Default: 8 parallel tasks for local runs
2026-03-11 06:52:26 -07:00
dmahan93 b03aefaf20 test: 13 tests for Modal sandbox infra fixes 2026-03-11 06:51:42 -07:00
dmahan93 d7f4db53f5 fix: Modal sandbox eval infra (9 fixes for TBLite baseline)
Fixes discovered while running TBLite baseline evaluation:

1. ephemeral_disk param not supported in modal 1.3.5 - check before passing
2. Modal legacy image builder requires working pip - add ensurepip fix via
   setup_dockerfile_commands to handle task images with broken pip
3. Host cwd leaked into Modal sandbox - add /home/ to host prefix check
4. Tilde ~ not expanded by subprocess.run(cwd=) in sandboxes - use /root
5. install_pipx must stay True for swerex-remote to be available

Dependencies also needed (not in this commit):
- git submodule update --init mini-swe-agent
- uv pip install swe-rex boto3
2026-03-11 06:51:42 -07:00
dmahan93 2c97bf3936 Add tests for atropos tool calling integration
- test_tool_call_parsers.py: 16 tests for parser registry, hermes parser
  (single/multiple/truncated/malformed), and ParseResult contract validation
- test_agent_loop.py: 21 tests for HermesAgentLoop with mock servers
  (text responses, tool calls, max turns, unknown tools, API errors,
  extra_body forwarding, managed state, blocked tools, reasoning extraction)
- test_managed_server_tool_support.py: 9 tests validating API compatibility
  between hermes-agent and atroposlib's ManagedServer tool_call_parser support
  (gracefully skips on baseline atroposlib, passes on tool_call_support branch)
2026-03-11 06:51:26 -07:00
dmahan93 7791174ced feat: add --fuck-it-ship-it flag to bypass dangerous command approvals
Adds a fun alias for skipping all dangerous command approval prompts.
When passed, sets HERMES_YOLO_MODE=1 which causes check_dangerous_command()
to auto-approve everything.

Available on both top-level and chat subcommand:
  hermes --fuck-it-ship-it
  hermes chat --fuck-it-ship-it

Includes 5 tests covering normal blocking, yolo bypass, all patterns,
and edge cases (empty string env var).
2026-03-08 18:36:37 -05:00
dmahan93 9079a27814 fix: prompt box and response box span full terminal width on wide screens
- Replace hardcoded '─' * 200 horizontal rules with Window(char='─')
  so prompt_toolkit fills the entire terminal width automatically
- Use shutil.get_terminal_size().columns instead of Rich Console.width
  for response box, separator line, and input height calculation
  (more reliable inside patch_stdout context)
2026-03-05 15:57:05 -08:00
Dakota e5e77381f0 Made to be more descriptive from comments 2025-09-12 18:10:36 -05:00
Dakota 066514e2a9 add more architecture docs 2025-09-12 17:47:32 -05:00
Dakota 045a1737f8 - message graphs 2025-09-12 10:25:14 -05:00