Files
app/backend/tests/test_cli.py
T
stroblmeandClaude Opus 5 4479eeb726
Docs / docs (push) Successful in 38s
Playwright Tests / test-playwright (1, 2) (push) Successful in 2m56s
Playwright Tests / test-playwright (2, 2) (push) Successful in 2m7s
pre-commit / pre-commit (push) Failing after 2m17s
Test Backend / test-backend (push) Successful in 2m54s
Compose Smoke Test / test-compose (push) Successful in 44s
Playwright Tests / merge-reports (push) Successful in 1m17s
Follow a record into its fields, name the metrics, name the version
Three things the first export pass got wrong for a real study.

**Dotted paths.** A node returns a record, not a scalar — the numbers arrive
inside `final_metrics` — so `--metrics final_metrics.train_loss` yielded an
empty column and `--metrics final_metrics` yielded the whole record in one
cell. Both sides of the wide table now take dotted paths, and the defaults
reach the same depth: every number a result carries is a column named by its
path, and inputs are compared leaf by leaf, so two configurations differing in
one field give that field as the axis rather than two blobs that are merely
not equal. Lists stay whole — a curve belongs in the long table.

**`--list`.** Metric names are flow-qualified, so `--name train_loss` matched
nothing and said only that. `fluksio export metrics --list` prints the names
the selection carries, and an empty export made with `--name` points at it.

**A version to compare.** The CLI ships ahead of the engine and a stale one
answered a flat 404 with nothing anywhere in the API to tell how old it was.
The engine reports `version` on `/observability/summary`, `fluksio status`
prints it, and a 404 from export now names both versions — or says "older"
when the field itself predates the engine. Bumped to 0.1.5, which is what
makes the number worth reading.

Also formats `flow/metrics.py`, which had been committed unformatted and was
the last `ruff format --check` failure.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01A9Hdrmf2cwNABCnE5x9UJa
2026-08-27 20:42:19 +02:00

391 lines
13 KiB
Python

"""The command line, and the one ordering it depends on."""
import subprocess
import sys
from pathlib import Path
from fluksio.cli import load_or_create_secret_key
def test_importing_the_cli_does_not_build_the_settings() -> None:
"""`_configure_environment` has to run before anything reads a setting.
The settings are built on the first import of `fluksio.core.config`, and
every engine module reaches it within an import or two. If importing the
CLI pulled it in, `DATA_DIR` would be fixed at whatever directory the
command was run from — which is how the database ends up in the cwd.
"""
leaked = subprocess.run(
[
sys.executable,
"-c",
"import fluksio.cli, sys; print('fluksio.core.config' in sys.modules)",
],
capture_output=True,
text=True,
check=True,
)
assert leaked.stdout.strip() == "False", leaked.stdout
def test_the_secret_key_is_kept_rather_than_regenerated(tmp_path: Path) -> None:
"""A new key each start would sign out every session and orphan secrets.enc."""
path = tmp_path / "secret_key"
first = load_or_create_secret_key(path)
assert load_or_create_secret_key(path) == first
assert path.stat().st_mode & 0o777 == 0o600
def test_the_store_works_without_git(tmp_path: Path, monkeypatch) -> None:
"""A pip install on a locked-down host may have no git.
Flows are files, and that is what the store is for; the history is the part
that needs git. Losing it must not be a refusal to start.
"""
import subprocess as sp
from fluksio.flow.store import FlowStore
real_run = sp.run
def no_git(cmd, *args, **kwargs):
if cmd and cmd[0] == "git":
raise FileNotFoundError(2, "No such file or directory", "git")
return real_run(cmd, *args, **kwargs)
monkeypatch.setattr(sp, "run", no_git)
monkeypatch.setattr(FlowStore, "_warned_no_git", False)
store = FlowStore(tmp_path / "flows")
assert store.head() == ""
store.write_requirements("numpy\n")
assert store.read_requirements() == "numpy\n"
def test_run_arguments_are_typed_by_the_flow_they_are_for() -> None:
"""`--lr 0.05` is a float because the flow says `lr` is one."""
from fluksio.sdk import SyncError
from fluksio.sdk.cli import _params
definition = {
"inputs": [
{"spec": {"name": "lr", "dtype": "float"}},
{"spec": {"name": "epochs", "dtype": "int"}},
{"spec": {"name": "resume", "dtype": "bool"}},
]
}
assert _params(definition, ["--lr", "0.05", "--epochs", "3", "--resume"]) == {
"lr": 0.05,
"epochs": 3,
"resume": True,
}
assert _params(definition, ["--lr=1e-4"]) == {"lr": 0.0001}
import pytest
with pytest.raises(SyncError, match="not an input of this flow"):
_params(definition, ["--nonesuch", "1"])
def test_serve_uses_the_installation_the_directory_belongs_to(
tmp_path: Path, monkeypatch
) -> None:
"""A repository with its own venv gets its own engine, not the machine's."""
from fluksio import cli
monkeypatch.chdir(tmp_path)
monkeypatch.setattr(cli, "DEFAULT_HOME", tmp_path / "home" / ".fluksio")
# Nothing above it yet: one is made here rather than in the home directory.
made = cli._data_dir(None)
assert made == (tmp_path / ".fluksio").resolve()
assert (made / ".gitignore").exists()
# And is then found again from anywhere below it.
deep = tmp_path / "src" / "deep"
deep.mkdir(parents=True)
monkeypatch.chdir(deep)
assert cli._data_dir(None) == made
# `--global` asks for the shared one even so.
assert (
cli._data_dir(None, shared=True) == (tmp_path / "home" / ".fluksio").resolve()
)
# And `--data-dir` still names any directory outright.
named = cli._data_dir(str(tmp_path / "named"))
assert named == (tmp_path / "named").resolve()
def test_enrolling_needs_only_a_claim_code() -> None:
"""The default portal is what makes first-run one flag rather than two."""
from fluksio.cli import DEFAULT_PORTAL, _parser
args = _parser().parse_args(["enroll", "ABC-123"])
assert args.portal == DEFAULT_PORTAL
assert DEFAULT_PORTAL.startswith("https://")
# And a portal of your own still wins.
mine = _parser().parse_args(["enroll", "ABC-123", "--portal", "https://hub.me"])
assert mine.portal == "https://hub.me"
def test_run_syncs_by_default_and_can_be_told_not_to() -> None:
from fluksio.cli import _parser
assert _parser().parse_args(["run", "train"]).no_sync is False
assert _parser().parse_args(["run", "train", "--no-sync"]).no_sync is True
def test_a_sweep_is_the_product_of_the_parameters_given() -> None:
"""`--param lr=0.1,0.01 --param epochs=1,2` is four runs, typed by the flow."""
import pytest
from fluksio.sdk import SyncError
from fluksio.sdk.cli import _grid
definition = {
"inputs": [
{"spec": {"name": "lr", "dtype": "float"}},
{"spec": {"name": "epochs", "dtype": "int"}},
]
}
grid = _grid(definition, ["lr=0.1,0.01", "epochs=1,2"], seed=7)
assert [entry["params"] for entry in grid] == [
{"lr": 0.1, "epochs": 1},
{"lr": 0.1, "epochs": 2},
{"lr": 0.01, "epochs": 1},
{"lr": 0.01, "epochs": 2},
]
assert all(entry["seed"] == 7 for entry in grid)
with pytest.raises(SyncError, match="not an input of this flow"):
_grid(definition, ["nonesuch=1"], seed=None)
with pytest.raises(SyncError, match="name=value"):
_grid(definition, ["lr"], seed=None)
def test_the_local_engine_is_asked_for_rather_than_guessed() -> None:
from fluksio.cli import _parser
parser = _parser()
assert parser.parse_args(["run", "train"]).local is False
assert parser.parse_args(["run", "train", "--local"]).local is True
assert parser.parse_args(["runs", "--local"]).local is True
assert parser.parse_args(["sweep", "train", "--param", "lr=1"]).local is False
def test_a_local_run_always_waits(monkeypatch) -> None:
"""The engine is this process, so a run nobody waits for is thrown away."""
from contextlib import contextmanager
from fluksio.cli import _parser
from fluksio.sdk import cli
submitted: dict[str, object] = {}
class FakeHandle:
id = "run-1"
status = "ok"
result: dict[str, object] = {}
def wait(self, timeout: float = 0.0) -> "FakeHandle":
submitted["waited"] = True
return self
class FakeClient:
def get_flow(self, name: str) -> dict[str, object]:
return {"definition": {"inputs": []}}
def submit(self, flow, params, seed=None, no_cache=False, cause="sdk"):
submitted["flow"] = flow
submitted["no_cache"] = no_cache
submitted["cause"] = cause
return FakeHandle()
def run(self, run_id: str) -> dict[str, object]:
return {"nodes": [{"status": "cached"}, {"status": "ok"}]}
@contextmanager
def fake_engine():
yield FakeClient()
monkeypatch.setattr(cli, "_engine_client", fake_engine)
args = _parser().parse_args(["run", "train", "--local", "--no-sync", "--no-cache"])
assert cli.cmd_run(args, []) == 0
# "cli" rather than the SDK's default: the history says which asked.
assert submitted == {
"flow": "train",
"no_cache": True,
"waited": True,
"cause": "cli",
}
def test_ctrl_c_while_waiting_cancels_the_run(monkeypatch) -> None:
"""Interrupting means stop the run, not walk away leaving it going."""
from contextlib import contextmanager
from fluksio.cli import _parser
from fluksio.sdk import cli
cancelled: list[str] = []
class FakeHandle:
id = "run-1"
status = "running"
result: dict[str, object] = {}
def wait(self, timeout: float = 0.0):
raise KeyboardInterrupt
class FakeClient:
def get_flow(self, name: str) -> dict[str, object]:
return {"definition": {"inputs": []}}
def submit(self, flow, params, seed=None, no_cache=False, cause="sdk"):
return FakeHandle()
def cancel(self, run_id: str) -> None:
cancelled.append(run_id)
@contextmanager
def fake_engine():
yield FakeClient()
monkeypatch.setattr(cli, "_engine_client", fake_engine)
args = _parser().parse_args(["run", "train", "--local", "--no-sync"])
assert cli.cmd_run(args, []) == 130
assert cancelled == ["run-1"]
def test_an_artifact_input_may_be_named_rather_than_pasted() -> None:
"""The engine resolves either spelling; the CLI just stops mangling them."""
import json
from fluksio.sdk.cli import _coerce
assert _coerce("@run:123-abc.dataset", "artifact") == "@run:123-abc.dataset"
digest = "sha256:" + "a1" * 32
assert _coerce(digest, "artifact") == digest
# A reference a script already holds still arrives as the object it is.
reference = {"digest": digest, "size": 3}
assert _coerce(json.dumps(reference), "artifact") == reference
def test_the_run_prompt_keeps_declared_values_for_anything_left_blank() -> None:
"""Enter through the lot is what running with the defaults looks like."""
from unittest.mock import patch
from fluksio.sdk.cli import _ask_params
definition = {
"inputs": [
{"spec": {"name": "lr", "dtype": "float"}, "initial": 0.05},
{"spec": {"name": "epochs", "dtype": "int"}, "initial": 10},
]
}
with patch("builtins.input", side_effect=["", ""]):
# Nothing sent: an input the caller leaves out keeps what it declares,
# which is the same contract the browser's dialog has.
assert _ask_params(definition) == {}
with patch("builtins.input", side_effect=["0.01", "50"]):
assert _ask_params(definition) == {"lr": 0.01, "epochs": 50}
def test_the_run_prompt_names_an_answer_of_the_wrong_type() -> None:
from unittest.mock import patch
import pytest
from fluksio.sdk import SyncError
from fluksio.sdk.cli import _ask_params
definition = {"inputs": [{"spec": {"name": "lr", "dtype": "float"}}]}
with patch("builtins.input", side_effect=["fast"]):
with pytest.raises(SyncError, match="lr"):
_ask_params(definition)
def test_an_export_is_parsed_with_its_selection() -> None:
"""`export` is a group of two tables, and both take the same filters."""
from fluksio.cli import _parser
from fluksio.sdk.cli import cmd_export_metrics, cmd_export_runs
parser = _parser()
metrics = parser.parse_args(
[
"export",
"metrics",
"--flow",
"train",
"--run",
"a",
"--run",
"b",
"--name",
"loss,val",
"--stride",
"5",
]
)
assert metrics.func is cmd_export_metrics
assert metrics.run == ["a", "b"]
assert metrics.stride == 5
assert metrics.format == "csv"
runs = parser.parse_args(["export", "runs", "--params", "lr", "--format", "jsonl"])
assert runs.func is cmd_export_runs
assert runs.params == "lr"
assert runs.format == "jsonl"
def test_the_metric_names_are_asked_for_rather_than_guessed() -> None:
"""A name is flow-qualified, so `--list` is what says what would match."""
from fluksio.cli import _parser
from fluksio.sdk.cli import _list_names
parser = _parser()
assert parser.parse_args(["export", "metrics"]).list_names is False
assert parser.parse_args(["export", "metrics", "--list"]).list_names is True
class Engine:
def runs(self, flow="", limit=0, **filters):
assert filters == {"status": "ok"}
return [{"id": "r-empty"}, {"id": "r-1"}]
def metrics(self, run_id, name="", stride=1):
# The newest run failed before it measured anything; the next one
# carries the vocabulary.
return [] if run_id == "r-empty" else [{"name": "train.train_loss"}]
args = parser.parse_args(
["export", "metrics", "--flow", "train", "--status", "ok", "--list"]
)
assert _list_names(Engine(), args) == 0
def test_an_engine_without_the_route_is_named_rather_than_404() -> None:
"""A client ships ahead of the engine; a flat 404 does not say so."""
from fluksio.sdk.cli import _too_old
class Old:
def summary(self):
return {"status": "ok", "version": "0.1.4"}
class Ancient:
def summary(self):
# Older than the field itself.
return {"status": "ok"}
assert "engine is 0.1.4" in _too_old(Old())
assert "engine is older" in _too_old(Ancient())
assert "pip install -U fluksio" in _too_old(Ancient())