procurement-ingest/tests/conftest.py

"""Shared pytest configuration for the procurement-ingest test suite.

The Lambda handlers create boto3 clients at module import time, so a
region and dummy credentials must be present in the environment before
any handler module is imported. Setting them here at conftest import
time guarantees they exist before test collection touches a handler.
"""

import importlib.util
import os
import sys
from pathlib import Path

import pytest

os.environ.setdefault("AWS_DEFAULT_REGION", "us-east-1")
os.environ.setdefault("AWS_ACCESS_KEY_ID", "testing")
os.environ.setdefault("AWS_SECRET_ACCESS_KEY", "testing")
os.environ.setdefault("AWS_SESSION_TOKEN", "testing")

REPO_ROOT = Path(__file__).resolve().parents[1]
_SHARED_DIR = REPO_ROOT / "lambdas" / "shared"


# Sibling modules imported by bare name from the handlers (the Lambda runtime
# puts each function's own directory on sys.path; the CDK bundling then cp's the
# shared modules in flat beside handler.py so those bare imports resolve too).
# template_parser/derived_fields are duplicated PER PIPELINE, so their bare
# names MUST be bound to the right pipeline's file around each handler exec --
# relying on sys.path ordering (or on whatever a previously collected suite left
# in sys.modules) silently binds a handler to the OTHER pipeline's sibling.
# ses_auth/email_parsing/emf are now single-sourced under lambdas/shared/ (Phase
# 3); the loop below resolves them from there via a shared-dir fallback.
#
# Phase 5 decomposed each God-handler into flat siblings (prompts/telemetry/
# extraction/enrichment/persistence). The order below is DEPENDENCY-TOPOLOGICAL,
# not alphabetical: the loader binds each bare name in sys.modules right after
# exec'ing it, so a sibling whose module body does `from <x> import ...` must
# appear AFTER <x> here or its exec ImportErrors. The load-bearing edges are
# emf < telemetry, prompts < extraction, and derived_fields + telemetry <
# enrichment. WO has no enrichment/derived_fields sibling -- the loader's
# `if not sibling_path.exists(): continue` silently skips them there, so one
# unified tuple serves both pipelines.
_SIBLING_MODULES = (
    "ses_auth",
    "email_parsing",
    "emf",
    "prompts",
    "template_parser",
    "derived_fields",
    "telemetry",
    "extraction",
    "enrichment",
    "persistence",
)


def _load_module(path, module_name):
    if module_name in sys.modules:
        return sys.modules[module_name]
    spec = importlib.util.spec_from_file_location(module_name, path)
    module = importlib.util.module_from_spec(spec)
    sys.modules[module_name] = module
    spec.loader.exec_module(module)
    return module


def load_handler(relative_path, module_name):
    """Load a Lambda handler module by file path under a unique name.

    The handler files all share the basename ``handler.py`` and are not
    importable as packages, so a plain ``import handler`` would collide
    across Lambdas. The same loader serves the ``ses_auth.py`` modules,
    which are likewise duplicated per pipeline and not importable as
    packages.

    Handler modules import their siblings by bare name (e.g. ``from
    template_parser import try_deterministic_parse``). Each sibling is loaded
    from the handler's own directory under a unique module name and registered
    under its bare name only for the duration of the handler exec, then the
    previous binding is restored -- so this loader is deterministic regardless
    of collection order and of what the per-Lambda test suites (which put
    their own module dir on sys.path) have already cached in sys.modules.
    """
    path = REPO_ROOT / relative_path
    if module_name in sys.modules:
        return sys.modules[module_name]
    # Keep the handler dir on sys.path for parity with the Lambda runtime.
    handler_dir = str(path.parent)
    if handler_dir not in sys.path:
        sys.path.insert(0, handler_dir)
    if path.name != "handler.py":
        # Leaf modules (e.g. ses_auth.py itself) have no sibling imports.
        return _load_module(path, module_name)
    saved = {}
    for sibling in _SIBLING_MODULES:
        # Per-pipeline siblings (template_parser/derived_fields) resolve next to
        # the handler; the shared, single-sourced siblings (ses_auth/
        # email_parsing/emf) fall back to lambdas/shared/. No ambiguity: post
        # Phase 3 the shared names exist ONLY under shared/, the per-pipeline
        # names ONLY next to the handler.
        sibling_path = path.parent / f"{sibling}.py"
        if not sibling_path.exists():
            sibling_path = _SHARED_DIR / f"{sibling}.py"
        if not sibling_path.exists():
            continue
        saved[sibling] = sys.modules.get(sibling)
        sys.modules[sibling] = _load_module(sibling_path, f"{module_name}__{sibling}")
    try:
        module = _load_module(path, module_name)
    finally:
        for sibling, previous in saved.items():
            if previous is not None:
                sys.modules[sibling] = previous
            else:
                sys.modules.pop(sibling, None)
    return module


@pytest.fixture(scope="session")
def po_handler():
    """The PO email processor handler module."""
    return load_handler(
        "lambdas/po/email_processor/handler.py",
        "po_email_processor_handler",
    )


@pytest.fixture(scope="session")
def po_persistence(po_handler):
    """The PO persistence sibling (owns PO_TABLE + save_* after Phase 5).

    load_handler registers each sibling under ``<handler>__<sibling>`` in
    sys.modules (only the bare-name binding is restored afterwards), so the
    persistence module stays reachable here by its unique name. The root
    tests/test_po_merge.py keys its moto table on ``po_persistence.PO_TABLE``
    rather than a hardcoded literal.
    """
    return sys.modules["po_email_processor_handler__persistence"]


@pytest.fixture(scope="session")
def po_enrichment(po_handler):
    """The PO enrichment sibling (owns pad_zip after Phase 5).

    pad_zip is PURE and NOT re-exported by handler (handler's body never calls
    it -- it runs inside enrich_parsed), so tests/test_pad_zip.py dereferences it
    on the owning module instead of po_handler.
    """
    return sys.modules["po_email_processor_handler__enrichment"]


@pytest.fixture(scope="session")
def wo_handler():
    """The WO email processor handler module."""
    return load_handler(
        "lambdas/wo/email_processor/handler.py",
        "wo_email_processor_handler",
    )


@pytest.fixture(params=["po_handler", "wo_handler"])
def email_handler(request):
    """Parametrized fixture yielding each email processor handler module."""
    return request.getfixturevalue(request.param)


@pytest.fixture
def ses_auth():
    """The single-sourced ses_auth module (lambdas/shared/, Phase 3).

    Previously parameterized over the two per-pipeline copies to prove they
    stayed byte-identical; now there is exactly one copy, so this loads it
    once -- halving the test_ses_auth run.
    """
    return load_handler("lambdas/shared/ses_auth.py", "shared_ses_auth")