test(mcp): add e2e tests for tool registration, validation, and integration

Phase 1 (no credentials): tool registration schemas, parameter validation,
uninitialized scraper error handling (29 tests).
Phase 2 (integration): full e2e through MCP protocol for all 11 tools (20 tests).
Adds pytest-asyncio auto mode and shared fixtures.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: kgritesh

pyproject.toml

@@ -79,6 +79,7 @@ linkedin_spider = ["py.typed"]
 "*" = ["*.md", "*.txt", "*.yml", "*.yaml"]
 
 [tool.pytest.ini_options]
+asyncio_mode = "auto"
 testpaths = ["tests"]
 python_files = ["test_*.py", "*_test.py"]
 addopts = [

tests/conftest.py

@@ -0,0 +1,43 @@
+from __future__ import annotations
+
+import os
+from collections.abc import AsyncGenerator, Generator
+
+import pytest
+from dotenv import load_dotenv
+from fastmcp import Client
+
+import linkedin_spider.mcp.server as mcp_server
+from linkedin_spider.core.config import ScraperConfig
+from linkedin_spider.core.scraper import LinkedinSpider
+from linkedin_spider.mcp.server import mcp_app
+
+load_dotenv()
+
+
+@pytest.fixture(scope="session")
+def spider() -> Generator[LinkedinSpider, None, None]:
+    """Session-scoped LinkedinSpider authenticated via LINKEDIN_COOKIE from .env."""
+    cookie = os.environ.get("LINKEDIN_COOKIE")
+    if not cookie:
+        pytest.skip("LINKEDIN_COOKIE not set in environment")
+
+    config = ScraperConfig(headless=True)
+    scraper = LinkedinSpider(li_at_cookie=cookie, config=config)
+    yield scraper
+    scraper.close()
+
+
+@pytest.fixture(scope="session")
+def mcp_scraper(spider: LinkedinSpider) -> Generator[LinkedinSpider, None, None]:
+    """Inject the session-scoped spider into the MCP server global, reset on teardown."""
+    mcp_server._scraper_instance = spider
+    yield spider
+    mcp_server._scraper_instance = None
+
+
+@pytest.fixture
+async def mcp_client() -> AsyncGenerator[Client, None]:
+    """Provide a connected FastMCP Client for in-memory MCP testing."""
+    async with Client(mcp_app) as client:
+        yield client

tests/test_e2e.py

@@ -0,0 +1,238 @@
+"""End-to-end tests for linkedin-spider scraper actions.
+
+One scrape call per test.  Non-deterministic outputs are validated by
+structure, types, and minimum result counts — not exact values.
+Parametrized where multiple inputs exercise different code paths.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+
+import pytest
+
+from linkedin_spider.core.scraper import LinkedinSpider
+
+PROFILE_URLS = [
+    "https://www.linkedin.com/in/williamhgates/",
+    "https://www.linkedin.com/in/satyanadella/",
+]
+COMPANY_URLS = [
+    "https://www.linkedin.com/company/microsoft/",
+    "https://www.linkedin.com/company/google/",
+]
+PROFILE_EXPECTED_KEYS = {
+    "name",
+    "headline",
+    "location",
+    "about",
+    "experience",
+    "education",
+    "profile_url",
+}
+COMPANY_EXPECTED_KEYS = {
+    "name",
+    "company_url",
+    "tagline",
+    "industry",
+    "location",
+    "followers",
+    "employee_count",
+}
+
+
+def _assert_non_empty_str(value: Any, field: str) -> None:
+    assert isinstance(value, str), f"{field} should be str, got {type(value)}"
+    assert value and value != "N/A", f"{field} should not be empty/N/A"
+
+
+# ── profile ────────────────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+@pytest.mark.parametrize("profile_url", PROFILE_URLS)
+def test_scrape_profile(spider: LinkedinSpider, profile_url: str) -> None:
+    result = spider.scrape_profile(profile_url)
+
+    assert result is not None, "scrape_profile returned None"
+    assert PROFILE_EXPECTED_KEYS.issubset(result.keys()), (
+        f"Missing keys: {PROFILE_EXPECTED_KEYS - result.keys()}"
+    )
+    _assert_non_empty_str(result["name"], "name")
+    assert isinstance(result["experience"], list)
+    assert isinstance(result["education"], list)
+    assert result["profile_url"] == profile_url
+
+
+@pytest.mark.integration
+def test_scrape_profile_invalid_url(spider: LinkedinSpider) -> None:
+    assert spider.scrape_profile("https://example.com/not-a-profile") is None
+
+
+# ── company ────────────────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+@pytest.mark.parametrize("company_url", COMPANY_URLS)
+def test_scrape_company(spider: LinkedinSpider, company_url: str) -> None:
+    result = spider.scrape_company(company_url)
+
+    assert result is not None, "scrape_company returned None"
+    assert COMPANY_EXPECTED_KEYS.issubset(result.keys()), (
+        f"Missing keys: {COMPANY_EXPECTED_KEYS - result.keys()}"
+    )
+    _assert_non_empty_str(result["name"], "name")
+    assert result["company_url"] == company_url
+
+
+@pytest.mark.integration
+def test_scrape_company_invalid_url(spider: LinkedinSpider) -> None:
+    assert spider.scrape_company("https://example.com/not-a-company") is None
+
+
+# ── search profiles ───────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+@pytest.mark.parametrize(
+    ("query", "max_results", "filters"),
+    [
+        ("software engineer", 3, None),
+        ("data scientist", 2, {"location": "San Francisco"}),
+        ("product manager", 2, {"industry": "Technology, Information and Internet"}),
+    ],
+    ids=["no-filter", "location-filter", "industry-filter"],
+)
+def test_search_profiles(
+    spider: LinkedinSpider,
+    query: str,
+    max_results: int,
+    filters: dict[str, str] | None,
+) -> None:
+    results = spider.search_profiles(query, max_results=max_results, filters=filters)
+
+    assert isinstance(results, list)
+    assert len(results) >= 1, f"Expected at least 1 result for '{query}'"
+    assert len(results) <= max_results
+
+    first = results[0]
+    expected_keys = {"name", "headline", "location", "profile_url"}
+    assert expected_keys.issubset(first.keys()), f"Missing keys: {expected_keys - first.keys()}"
+
+
+# ── search posts ──────────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+@pytest.mark.parametrize(
+    ("keywords", "date_posted"),
+    [
+        ("artificial intelligence", None),
+        ("startup funding", "past-week"),
+    ],
+    ids=["no-date-filter", "past-week"],
+)
+def test_search_posts(
+    spider: LinkedinSpider,
+    keywords: str,
+    date_posted: str | None,
+) -> None:
+    results = spider.search_posts(
+        keywords,
+        max_results=2,
+        max_comments=0,
+        date_posted=date_posted,
+    )
+
+    assert isinstance(results, list)
+    assert len(results) >= 1, f"Expected at least 1 post for '{keywords}'"
+
+    post = results[0]
+    expected_keys = {
+        "author_name",
+        "author_headline",
+        "author_profile_url",
+        "post_text",
+        "hashtags",
+        "links",
+        "post_url",
+        "likes_count",
+        "comments_count",
+        "reposts_count",
+    }
+    assert expected_keys.issubset(post.keys()), f"Missing keys: {expected_keys - post.keys()}"
+    assert isinstance(post["likes_count"], int)
+    assert isinstance(post["comments_count"], int)
+    assert isinstance(post["reposts_count"], int)
+    assert isinstance(post["hashtags"], list)
+    assert isinstance(post["links"], list)
+
+
+# ── conversations list ────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+def test_scrape_conversations_list(spider: LinkedinSpider) -> None:
+    results = spider.scrape_conversations_list(max_results=3)
+
+    assert isinstance(results, list)
+    if not results:
+        pytest.skip("No conversations available")
+
+    convo = results[0]
+    expected_keys = {"participant_name", "timestamp", "message_snippet"}
+    assert expected_keys.issubset(convo.keys()), f"Missing keys: {expected_keys - convo.keys()}"
+
+
+# ── conversation messages ─────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+def test_scrape_conversation_messages(spider: LinkedinSpider) -> None:
+    result = spider.scrape_conversation_messages()
+
+    assert result is not None
+    assert "messages" in result
+    assert "total_messages" in result
+    assert isinstance(result["messages"], list)
+    assert isinstance(result["total_messages"], int)
+
+
+# ── incoming connections ──────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+def test_scrape_incoming_connections(spider: LinkedinSpider) -> None:
+    results = spider.scrape_incoming_connections(max_results=3)
+
+    assert isinstance(results, list)
+    if not results:
+        pytest.skip("No incoming connections available")
+
+    conn = results[0]
+    expected_keys = {"name", "profile_url", "headline"}
+    assert expected_keys.issubset(conn.keys()), f"Missing keys: {expected_keys - conn.keys()}"
+
+
+# ── outgoing connections ──────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+def test_scrape_outgoing_connections(spider: LinkedinSpider) -> None:
+    results = spider.scrape_outgoing_connections(max_results=3)
+
+    assert isinstance(results, list)
+    if not results:
+        pytest.skip("No outgoing connections available")
+
+    conn = results[0]
+    expected_keys = {"name", "profile_url", "headline"}
+    assert expected_keys.issubset(conn.keys()), f"Missing keys: {expected_keys - conn.keys()}"
+
+
+# ── keep alive ────────────────────────────────────────────────────────────
+
+
+@pytest.mark.integration
+def test_keep_alive(spider: LinkedinSpider) -> None:
+    assert spider.keep_alive() is True