diff --git a/cwmscli/load/root.py b/cwmscli/load/root.py index 1366ad3..d42827a 100644 --- a/cwmscli/load/root.py +++ b/cwmscli/load/root.py @@ -4,14 +4,16 @@ import logging from dataclasses import dataclass from typing import Optional -from urllib.parse import urlparse +from urllib.parse import urljoin, urlparse import click +import requests from cwmscli import requirements as reqs from cwmscli.utils.deps import requires logger = logging.getLogger(__name__) +CDA_PROBE_TIMEOUT_SECONDS = 2.5 CONTEXT = dict( help_option_names=["-h", "--help"], @@ -40,11 +42,74 @@ def _norm_office(o: Optional[str]) -> str: return (o or "").strip().upper() +def _swagger_docs_url(api_root: str) -> str: + return urljoin(f"{api_root.rstrip('/')}/", "swagger-docs") + + +def _looks_like_cda_landing_page(response: requests.Response) -> bool: + # Some local CDA builds serve the UI but not the generated OpenAPI route. + server = response.headers.get("Server", "").lower() + if "cwms-data-api" in server: + return True + + content_type = response.headers.get("Content-Type", "").lower() + if "html" not in content_type: + return False + return "CDA - CWMS Data API" in response.text + + +def _validate_cda_api_root(api_root: str, *, role: str) -> None: + parsed = urlparse(api_root) + if parsed.scheme not in {"http", "https"} or not parsed.netloc: + raise click.ClickException( + f"{role} CDA URL must be an absolute http(s) URL, got {api_root!r}." + ) + + url = _swagger_docs_url(api_root) + try: + response = requests.get( + url, + headers={"Accept": "application/json"}, + timeout=CDA_PROBE_TIMEOUT_SECONDS, + ) + response.raise_for_status() + document = response.json() + if isinstance(document, dict) and ( + document.get("openapi") or document.get("swagger") + ): + return + except (requests.RequestException, ValueError) as openapi_error: + logger.debug( + "CDA OpenAPI probe for %s at %s did not succeed: %s", + role, + url, + openapi_error, + ) + + try: + response = requests.get(api_root, timeout=CDA_PROBE_TIMEOUT_SECONDS) + response.raise_for_status() + except requests.Timeout as e: + raise click.ClickException( + f"Could not validate {role} CDA at {api_root}: timed out fetching {api_root}." + ) from e + except requests.RequestException as e: + raise click.ClickException( + f"Could not validate {role} CDA at {api_root}: failed to fetch {api_root}: {e}" + ) from e + + if not _looks_like_cda_landing_page(response): + raise click.ClickException( + f"{role} URL {api_root} did not return a CDA OpenAPI document or CDA landing page." + ) + + def validate_cda_targets(func): @functools.wraps(func) def wrapper(*args, **kwargs): source_csv = kwargs.get("source_csv") target_csv = kwargs.get("target_csv") + skip_target_cda_check = kwargs.pop("skip_target_cda_check", False) if source_csv and target_csv: raise click.ClickException( @@ -92,6 +157,10 @@ def wrapper(*args, **kwargs): "This is allowed, but double-check intent.", ) + # Dry-runs still need a real target; otherwise users can validate a bad load command. + if target_cda and not skip_target_cda_check: + _validate_cda_api_root(target_cda, role="Target") + src_label = source_csv or source_cda or "-" tgt_label = target_csv or target_cda or "-" logger.info( @@ -134,6 +203,14 @@ def shared_source_target_options(f): envvar="CDA_API_KEY", help="Target API key used when no saved cwms-cli login token is available.", )(f) + f = click.option( + "--skip-target-cda-check", + envvar="CWMS_CLI_SKIP_TARGET_CDA_CHECK", + is_flag=True, + default=False, + show_default=True, + help="Skip the preflight check that --target-cda points to a CDA service.", + )(f) f = click.option( "--dry-run/--no-dry-run", is_flag=True, diff --git a/tests/commands/test_load_location_ids.py b/tests/commands/test_load_location_ids.py index 065da52..d5c1ac2 100644 --- a/tests/commands/test_load_location_ids.py +++ b/tests/commands/test_load_location_ids.py @@ -267,6 +267,9 @@ def test_cli_rejects_source_csv_and_target_csv_together(tmp_path, monkeypatch): monkeypatch.setattr( "cwmscli.utils.get_saved_login_token", lambda *args, **kwargs: None ) + monkeypatch.setattr( + "cwmscli.load.root._validate_cda_api_root", lambda *a, **k: None + ) src = tmp_path / "in.csv" src.write_text("name,office-id,active\nLOC_A,SWT,True\n") out = tmp_path / "out.csv" @@ -292,6 +295,9 @@ def test_cli_rejects_source_csv_with_explicit_source_cda(tmp_path, monkeypatch): monkeypatch.setattr( "cwmscli.utils.get_saved_login_token", lambda *args, **kwargs: None ) + monkeypatch.setattr( + "cwmscli.load.root._validate_cda_api_root", lambda *a, **k: None + ) src = tmp_path / "in.csv" src.write_text("name,office-id,active\nLOC_A,SWT,True\n") @@ -312,3 +318,61 @@ def test_cli_rejects_source_csv_with_explicit_source_cda(tmp_path, monkeypatch): ) assert result.exit_code != 0 assert "mutually exclusive" in result.output + + +def test_cli_can_skip_target_cda_check_with_flag(tmp_path, monkeypatch): + monkeypatch.setattr( + "cwmscli.load.root._validate_cda_api_root", + lambda *a, **k: (_ for _ in ()).throw( + AssertionError("target check should be skipped") + ), + ) + src = tmp_path / "in.csv" + src.write_text("name,office-id,active\nLOC_A,SWT,True\n") + + runner = CliRunner() + result = runner.invoke( + location_group, + [ + "ids-all", + "--source-csv", + str(src), + "--source-office", + "SWT", + "--target-cda", + "http://not-cda.example/cwms-data", + "--skip-target-cda-check", + "--dry-run", + ], + ) + + assert result.exit_code == 0, result.output + + +def test_cli_can_skip_target_cda_check_with_env_var(tmp_path, monkeypatch): + monkeypatch.setenv("CWMS_CLI_SKIP_TARGET_CDA_CHECK", "1") + monkeypatch.setattr( + "cwmscli.load.root._validate_cda_api_root", + lambda *a, **k: (_ for _ in ()).throw( + AssertionError("target check should be skipped") + ), + ) + src = tmp_path / "in.csv" + src.write_text("name,office-id,active\nLOC_A,SWT,True\n") + + runner = CliRunner() + result = runner.invoke( + location_group, + [ + "ids-all", + "--source-csv", + str(src), + "--source-office", + "SWT", + "--target-cda", + "http://not-cda.example/cwms-data", + "--dry-run", + ], + ) + + assert result.exit_code == 0, result.output diff --git a/tests/load/test_cda_target_validation.py b/tests/load/test_cda_target_validation.py new file mode 100644 index 0000000..022b56e --- /dev/null +++ b/tests/load/test_cda_target_validation.py @@ -0,0 +1,94 @@ +import click +import pytest + +from cwmscli.load.root import _validate_cda_api_root + + +class FakeResponse: + def __init__(self, payload=None, error=None, headers=None, text=""): + self.payload = payload + self.error = error + self.headers = headers or {} + self.text = text + + def raise_for_status(self): + if self.error: + raise self.error + + def json(self): + if isinstance(self.payload, Exception): + raise self.payload + return self.payload + + +def test_validate_cda_api_root_accepts_openapi_document(monkeypatch): + calls = [] + + def fake_get(url, headers=None, timeout=None): + calls.append((url, headers, timeout)) + return FakeResponse({"openapi": "3.0.1", "info": {"title": "CWMS Data API"}}) + + monkeypatch.setattr("cwmscli.load.root.requests.get", fake_get) + + _validate_cda_api_root("http://localhost:8082/cwms-data/", role="Target") + + assert calls == [ + ( + "http://localhost:8082/cwms-data/swagger-docs", + {"Accept": "application/json"}, + 2.5, + ) + ] + + +def test_validate_cda_api_root_rejects_non_openapi_document(monkeypatch): + monkeypatch.setattr( + "cwmscli.load.root.requests.get", + lambda *a, **k: FakeResponse({"message": "not cda"}), + ) + + with pytest.raises(click.ClickException, match="did not return a CDA OpenAPI"): + _validate_cda_api_root("https://example.test/not-cda", role="Target") + + +def test_validate_cda_api_root_accepts_cda_landing_page(monkeypatch): + calls = [] + + def fake_get(url, headers=None, timeout=None): + calls.append(url) + if url.endswith("/swagger-docs"): + return FakeResponse(ValueError("not json")) + return FakeResponse( + headers={"Content-Type": "text/html"}, + text="CDA - CWMS Data API", + ) + + monkeypatch.setattr("cwmscli.load.root.requests.get", fake_get) + + _validate_cda_api_root("http://localhost:7000/cwms-data", role="Target") + + assert calls == [ + "http://localhost:7000/cwms-data/swagger-docs", + "http://localhost:7000/cwms-data", + ] + + +def test_validate_cda_api_root_rejects_invalid_url(): + with pytest.raises(click.ClickException, match="absolute http"): + _validate_cda_api_root("sas-t7/sas-data/", role="Target") + + +def test_validate_cda_api_root_reports_request_failure(monkeypatch): + class FakeRequestException(Exception): + pass + + def fake_get(*args, **kwargs): + raise FakeRequestException("connection refused") + + monkeypatch.setattr("cwmscli.load.root.requests.get", fake_get) + monkeypatch.setattr( + "cwmscli.load.root.requests.RequestException", FakeRequestException + ) + + with pytest.raises(click.ClickException, match="failed to fetch"): + _validate_cda_api_root("http://localhost:9999/cwms-data", role="Target") diff --git a/tests/load/test_timeseries_data.py b/tests/load/test_timeseries_data.py index 21d194c..ba69870 100644 --- a/tests/load/test_timeseries_data.py +++ b/tests/load/test_timeseries_data.py @@ -12,6 +12,9 @@ def test_load_timeseries_data_command_allows_group_without_category_filters( monkeypatch, ): + monkeypatch.setattr( + "cwmscli.load.root._validate_cda_api_root", lambda *a, **k: None + ) calls = [] def fake_load(**kwargs):