teknium1
d10bb2ab6f
`scripts/run_tests.sh tests/<dir>/` is how a change gets its regression
coverage run, so a test filed under the wrong directory is a test nobody
runs when that code changes. Two kinds of drift had accumulated.
Parallel directories for one source package, folded into the mirror:
tests/acp -> tests/acp_adapter (its __init__/conftest move with it)
tests/cli -> tests/hermes_cli (prompt_toolkit fixture merged into
hermes_cli/conftest.py)
tests/run_agent -> tests/agent (backoff fixture becomes
agent/conftest.py)
tests/relay -> tests/gateway/relay
tests/state -> tests/hermes_state
246 loose files at tests/ root, routed by the package they import/patch:
hermes_cli, hermes_state, agent, gateway, tools, plugins, tui_gateway, cron.
Installer and desktop-update script tests go to tests/scripts/{install,
desktop_update}/. 43 tests of root-level modules (batch_runner, utils,
hermes_constants, packaging) stay at the root.
Filenames drop their issue numbers (95 files: test_89315_x.py -> test_x.py);
the number stays in the module docstring where it has context.
Collisions: test_cli_skin_integration.py existed in both tests/ and tests/cli
with different subsets — merged into one (10 tests, all kept);
run_agent/test_pre_compress_memory_context.py -> agent/..._handoff.py;
tests/test_account_usage.py -> agent/test_account_usage_fetch.py;
tests/test_web_server.py -> hermes_cli/test_web_server_ws_ping.py.
Deleted: test_minisweagent_path.py (empty since PR #2804),
test_model_picker_scroll.py (tested a private copy of the logic, imported
nothing), test_process_loop_event_loop_warning.py (asserted asyncio behaviour,
imported nothing from Hermes).
Repo-root path arithmetic (Path(__file__).parents[N], dirname chains) is
bumped for the 202 files that changed depth and verified by evaluating every
such expression against the new location. classify_changes' desktop-updater
lane prefix, tests-os.yml's ignore glob and every in-tree path comment follow
the moves. tests/test_tests_tree_layout.py keeps the tree from drifting back.
128 lines
3.7 KiB
Python
128 lines
3.7 KiB
Python
"""Tests for AIAgent._repair_tool_call — tool-name normalization.
|
|
|
|
Regression guard for #14784: Claude-style models sometimes emit
|
|
class-like tool-call names (``TodoTool_tool``, ``Patch_tool``,
|
|
``BrowserClick_tool``, ``PatchTool``). Before the fix they returned
|
|
"Unknown tool" even though the target tool was registered under a
|
|
snake_case name. The repair routine now normalizes CamelCase,
|
|
strips trailing ``_tool`` / ``-tool`` / ``tool`` suffixes (up to
|
|
twice to handle double-tacked suffixes like ``TodoTool_tool``), and
|
|
falls back to fuzzy match.
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
from types import SimpleNamespace
|
|
|
|
import pytest
|
|
|
|
|
|
VALID = {
|
|
"todo",
|
|
"patch",
|
|
"browser_click",
|
|
"browser_navigate",
|
|
"web_search",
|
|
"read_file",
|
|
"write_file",
|
|
"terminal",
|
|
"execute_code",
|
|
"session_search",
|
|
}
|
|
|
|
|
|
@pytest.fixture
|
|
def repair():
|
|
"""Return a bound _repair_tool_call built on a minimal shell agent.
|
|
|
|
We avoid constructing a real AIAgent (which pulls in credential
|
|
resolution, session DB, etc.) because the repair routine only
|
|
reads self.valid_tool_names. A SimpleNamespace stub is enough to
|
|
bind the unbound function.
|
|
"""
|
|
from run_agent import AIAgent
|
|
stub = SimpleNamespace(valid_tool_names=VALID)
|
|
return AIAgent._repair_tool_call.__get__(stub, AIAgent)
|
|
|
|
|
|
class TestExistingBehaviorStillWorks:
|
|
"""Pre-existing repairs must keep working (no regressions)."""
|
|
|
|
def test_lowercase_already_matches(self, repair):
|
|
assert repair("browser_click") == "browser_click"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestClassLikeEmissions:
|
|
"""Regression coverage for #14784 — CamelCase + _tool suffix variants."""
|
|
|
|
def test_camel_case_no_suffix(self, repair):
|
|
assert repair("BrowserClick") == "browser_click"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class TestEdgeCases:
|
|
"""Edge inputs that must not crash or produce surprising results."""
|
|
|
|
def test_empty_string(self, repair):
|
|
assert repair("") is None
|
|
|
|
|
|
|
|
|
|
|
|
class TestVolcEngineXmlPollution:
|
|
"""Regression coverage for #33007 — VolcEngine ``api/plan`` endpoint
|
|
leaks raw XML attribute fragments into ``tool_use.name``.
|
|
|
|
Observed in production with the ``anthropic_messages`` API mode:
|
|
|
|
terminal" parameter="command" string="true
|
|
execute_code" parameter="code" string="true
|
|
session_search" parameter="session_id" string="true
|
|
|
|
The fix trims at the first ``"``/``'``/``<``/``>`` so the rest of
|
|
the repair pipeline can resolve the cleaned name to a real tool.
|
|
"""
|
|
|
|
def test_terminal_with_xml_attribute_pollution(self, repair):
|
|
# Exact pattern from the bug report (terminal call).
|
|
polluted = 'terminal" parameter="command" string="true'
|
|
assert repair(polluted) == "terminal"
|
|
|
|
|
|
|
|
|
|
def test_tool_name_with_trailing_quote_only(self, repair):
|
|
# Minimal leak — just a stray trailing quote, no full attribute.
|
|
assert repair('terminal"') == "terminal"
|
|
|
|
|
|
|
|
def test_clean_tool_name_unaffected_by_sanitizer(self, repair):
|
|
# Pure passthrough — no XML/quote chars, no change.
|
|
assert repair("execute_code") == "execute_code"
|
|
assert repair("session_search") == "session_search"
|
|
|
|
def test_space_separated_name_still_normalizes(self, repair):
|
|
# Critical: the XML strip must NOT consume whitespace, or the
|
|
# legitimate ``"write file" -> write_file`` repair path breaks.
|
|
assert repair("write file") == "write_file"
|
|
|
|
|
|
def test_leading_quote_falls_through_to_fuzzy_match(self, repair):
|
|
# Sanitizer only trims when the XML char is at idx > 0 — a
|
|
# name that *starts* with a quote is left untouched so the
|
|
# rest of the pipeline (fuzzy match at 0.7 cutoff) can still
|
|
# recover the obvious target.
|
|
assert repair('"terminal"') == "terminal"
|