From 6193ce27240b1fa04d7e0653d3546bdcd35eafc8 Mon Sep 17 00:00:00 2001
From: Russell Ballestrini <russell@unturf.com>
Date: Tue, 6 Jan 2026 15:26:31 -0500
Subject: [PATCH] Fix domain vault init check in finish_crawl/get_stats

- Add _initialized check to finish_crawl() and get_stats() in all 3 vault classes
- Prevents FileNotFoundError when git commands run on non-existent directories
- Update test_crawl.py to use pytest tmp_path fixture
- Add network marker to skip functional tests in CI with '-m "not network"'
---
 neopig/domain_vault.py | 18 ++++++++++++++++++
 pytest.ini             |  2 ++
 tests/test_crawl.py    | 41 +++++++++++++++++++++++++++++------------
 3 files changed, 49 insertions(+), 12 deletions(-)

diff --git a/neopig/domain_vault.py b/neopig/domain_vault.py
index ea06dcc..6f9be28 100644
--- a/neopig/domain_vault.py
+++ b/neopig/domain_vault.py
@@ -372,6 +372,9 @@ class DomainHtmlVault:
 
     async def finish_crawl(self, stats: Dict[str, Any]) -> Optional[str]:
         """Finish crawl, log it, and commit if changes."""
+        if not self._initialized:
+            await self.init()
+
         log = await self._load_crawl_log()
         log['crawls'].append({
             'timestamp': datetime.now(timezone.utc).isoformat(),
@@ -386,6 +389,9 @@ class DomainHtmlVault:
 
     async def get_stats(self) -> Dict[str, Any]:
         """Get vault statistics."""
+        if not self._initialized:
+            await self.init()
+
         log = await self._load_crawl_log()
         commits = await self._git.log(5)
 
@@ -510,6 +516,9 @@ class DomainMediaVault:
 
     async def finish_crawl(self, stats: Dict[str, Any]) -> Optional[str]:
         """Finish crawl and commit if changes."""
+        if not self._initialized:
+            await self.init()
+
         media_new = stats.get('media_new', 0)
         if media_new > 0:
             return await self._git.commit(f"Crawl: {media_new} media files")
@@ -517,6 +526,9 @@ class DomainMediaVault:
 
     async def get_stats(self) -> Dict[str, Any]:
         """Get vault statistics."""
+        if not self._initialized:
+            await self.init()
+
         index = await self._load_index()
         commits = await self._git.log(5)
 
@@ -649,6 +661,9 @@ class DomainLinkpeekVault:
 
     async def finish_crawl(self, stats: Dict[str, Any]) -> Optional[str]:
         """Finish crawl and commit if changes."""
+        if not self._initialized:
+            await self.init()
+
         screenshots_new = stats.get('screenshots_new', 0)
         if screenshots_new > 0:
             return await self._git.commit(f"Crawl: {screenshots_new} screenshots")
@@ -656,6 +671,9 @@ class DomainLinkpeekVault:
 
     async def get_stats(self) -> Dict[str, Any]:
         """Get vault statistics."""
+        if not self._initialized:
+            await self.init()
+
         index = await self._load_index()
         commits = await self._git.log(5)
 
diff --git a/pytest.ini b/pytest.ini
index f9815bd..b29b317 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -5,3 +5,5 @@ testpaths = tests
 python_files = test_*.py
 python_classes = Test*
 python_functions = test_*
+markers =
+    network: tests that require network access (deselect with '-m "not network"')
diff --git a/tests/test_crawl.py b/tests/test_crawl.py
index 003009e..1d5bb14 100644
--- a/tests/test_crawl.py
+++ b/tests/test_crawl.py
@@ -3,12 +3,17 @@
 Functional test for neopig crawler.
 
 Tests crawling upload.unturf.com for images.
+
+These tests hit live network and are marked with @pytest.mark.network.
+To skip in CI: pytest -m "not network"
 """
 
 import asyncio
 import sys
 from pathlib import Path
 
+import pytest
+
 # Add current dir to path
 sys.path.insert(0, str(Path(__file__).parent))
 
@@ -16,17 +21,21 @@ from neopig import NeoPig
 from neopig.async_web_fetcher import CrawlMode
 
 
-async def test_crawl_upload_unturf():
+@pytest.mark.network
+async def test_crawl_upload_unturf(tmp_path):
     """Test crawling upload.unturf.com for images."""
     print("=" * 60)
     print("neopig functional test")
     print("Target: https://upload.unturf.com")
     print("=" * 60)
 
-    # Use test database and vault
+    # Use temp directory for test database and vault
+    db_path = tmp_path / "test_neopig.db"
+    vault_path = tmp_path / "test_vault"
+
     pig = NeoPig(
-        db_path="test_neopig.db",
-        vault_path="test_vault"
+        db_path=str(db_path),
+        vault_path=str(vault_path)
     )
     await pig.init()
 
@@ -59,7 +68,7 @@ async def test_crawl_upload_unturf():
 
     # Check database
     from neopig.database import Database
-    db = Database("test_neopig.db")
+    db = Database(str(db_path))
     await db.init()
     db_stats = await db.get_stats()
     print(f"\n  Database stats:")
@@ -79,15 +88,20 @@ async def test_crawl_upload_unturf():
     return stats
 
 
-async def test_crawl_modes():
+@pytest.mark.network
+async def test_crawl_modes(tmp_path):
     """Test different crawl modes."""
     print("\n" + "=" * 60)
     print("Testing crawl modes...")
     print("=" * 60)
 
+    # Use temp directory for test database and vault
+    db_path = tmp_path / "test_neopig.db"
+    vault_path = tmp_path / "test_vault"
+
     pig = NeoPig(
-        db_path="test_neopig.db",
-        vault_path="test_vault"
+        db_path=str(db_path),
+        vault_path=str(vault_path)
     )
     await pig.init()
 
@@ -107,10 +121,13 @@ async def test_crawl_modes():
 
 
 if __name__ == "__main__":
+    import tempfile
     print("neopig functional test suite\n")
 
-    # Run tests
-    asyncio.run(test_crawl_upload_unturf())
-    asyncio.run(test_crawl_modes())
+    # Run tests with temp directory
+    with tempfile.TemporaryDirectory() as tmp:
+        tmp_path = Path(tmp)
+        asyncio.run(test_crawl_upload_unturf(tmp_path))
+        asyncio.run(test_crawl_modes(tmp_path))
 
-    print("\nDone! Check test_vault/ and test_neopig.db for results.")
+    print("\nDone!")