Add pytest suite and enable tests in CI (#95)

The repo had no automated tests, so regressions in the email
parsing pipeline could only be caught manually. The reusable
ci-python-sam workflow already supports a run-tests input at the
pinned SHA; enable it and add a first suite covering the pure
functions pad_zip and parse_raw_email.

tests/conftest.py sets AWS_DEFAULT_REGION and dummy credentials
before any handler import because the handlers create boto3
clients at module import time, and loads each handler.py via
importlib under a unique module name since the files share a
basename. pytest.ini scopes discovery to tests/ so the manual
test_local.py script at the repo root is not collected.

.github/workflows/ci.yaml

@@ -7,6 +7,7 @@ jobs:
   ci:
     uses: Sea-Haven-Industries/.github/.github/workflows/ci-python-sam.yaml@fd60e4c9041784f666ac0fdefb9bec3c7fbf5143 # main
     with:
-      source-dirs: "lambdas cdk"
+      source-dirs: "lambdas cdk tests"
+      run-tests: true
       run-cdk-synth: true
       run-sam-validate: false

pytest.ini

@@ -0,0 +1,5 @@
+[pytest]
+; Restrict discovery to tests/. test_local.py at the repo root matches the
+; default test_*.py glob but is a manual script that imports a handler (and
+; therefore creates boto3 clients) at collection time.
+testpaths = tests

tests/conftest.py

@@ -0,0 +1,60 @@
+"""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]
+
+
+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.
+    """
+    path = REPO_ROOT / relative_path
+    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
+
+
+@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 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)

tests/test_pad_zip.py

@@ -0,0 +1,29 @@
+"""Tests for pad_zip in the PO email processor."""
+
+import pytest
+
+
+@pytest.mark.parametrize(
+    ("raw", "expected"),
+    [
+        # Missing values pass through untouched
+        (None, None),
+        ("", ""),
+        # Leading zeros restored (Coupa drops them: "MA 2149" -> "02149")
+        ("2149", "02149"),
+        ("7001", "07001"),
+        ("501", "00501"),
+        # Already-valid 5-digit zips are unchanged
+        ("85704", "85704"),
+        ("02149", "02149"),
+        # Zip+4: leading zeros restored on the base, suffix preserved
+        ("2149-1234", "02149-1234"),
+        ("85704-1234", "85704-1234"),
+        # Surrounding whitespace is handled
+        (" 2149 ", "02149"),
+        # Non-numeric values are left alone
+        ("ABC12", "ABC12"),
+    ],
+)
+def test_pad_zip(po_handler, raw, expected):
+    assert po_handler.pad_zip(raw) == expected

tests/test_parse_raw_email.py

@@ -0,0 +1,91 @@
+"""Tests for parse_raw_email in the PO and WO email processors.
+
+Synthetic MIME messages are built with email.message.EmailMessage so the
+tests exercise real multipart parsing without shipping .eml fixtures.
+"""
+
+from email.message import EmailMessage
+
+SUBJECT = "Amazon.com Services LLC Purchase Order 2D-18206023"
+SENDER = "do_not_reply@coupahost.com"
+TO = "orders@seahavenind.com"
+CC = "ap@seahavenind.com"
+DATE = "Tue, 14 Jul 2026 09:30:00 -0400"
+
+PLAIN_BODY = "PO 2D-18206023 issued to Sea Haven Industries."
+HTML_BODY = "<html><body><p>PO 2D-18206023 issued.</p></body></html>"
+
+
+def _base_message():
+    msg = EmailMessage()
+    msg["Subject"] = SUBJECT
+    msg["From"] = SENDER
+    msg["To"] = TO
+    msg["Cc"] = CC
+    msg["Date"] = DATE
+    return msg
+
+
+def make_plain_email():
+    msg = _base_message()
+    msg.set_content(PLAIN_BODY)
+    return msg.as_bytes()
+
+
+def make_multipart_alternative_email():
+    msg = _base_message()
+    msg.set_content(PLAIN_BODY)
+    msg.add_alternative(HTML_BODY, subtype="html")
+    return msg.as_bytes()
+
+
+def make_multipart_html_only_email():
+    msg = _base_message()
+    msg.set_content(HTML_BODY, subtype="html")
+    # An attachment forces the message to multipart/mixed
+    msg.add_attachment(
+        b"%PDF-1.4 fake",
+        maintype="application",
+        subtype="pdf",
+        filename="po.pdf",
+    )
+    return msg.as_bytes()
+
+
+def test_headers_extracted(email_handler):
+    result = email_handler.parse_raw_email(make_plain_email())
+    assert result["subject"] == SUBJECT
+    assert result["sender"] == SENDER
+    assert result["to"] == TO
+    assert result["date"] == DATE
+
+
+def test_simple_plain_body(email_handler):
+    result = email_handler.parse_raw_email(make_plain_email())
+    assert result["body"].strip() == PLAIN_BODY
+
+
+def test_multipart_prefers_text_plain(email_handler):
+    result = email_handler.parse_raw_email(make_multipart_alternative_email())
+    assert result["body"].strip() == PLAIN_BODY
+    assert "<html>" not in result["body"]
+
+
+def test_multipart_falls_back_to_html(email_handler):
+    result = email_handler.parse_raw_email(make_multipart_html_only_email())
+    assert "PO 2D-18206023 issued." in result["body"]
+    assert "<html>" in result["body"]
+
+
+def test_wo_handler_extracts_cc(wo_handler):
+    result = wo_handler.parse_raw_email(make_plain_email())
+    assert result["cc"] == CC
+
+
+def test_missing_headers_default_to_empty(email_handler):
+    msg = EmailMessage()
+    msg.set_content(PLAIN_BODY)
+    result = email_handler.parse_raw_email(msg.as_bytes())
+    assert result["subject"] == ""
+    assert result["sender"] == ""
+    assert result["to"] == ""