Skip to content

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

12.7 KB
Raw Blame
"""
Tests for document deletion functionality.

Tests the DocumentDeletionService methods:
- delete_document
- delete_blob_only
- remove_from_collection
"""

import pytest
import uuid

from sqlalchemy import create_engine, event
from sqlalchemy.orm import sessionmaker

from local_deep_research.database.models.library import (
    Document,
    DocumentBlob,
    DocumentChunk,
    DocumentCollection,
    Collection,
    SourceType,
)
from local_deep_research.constants import (
    FILE_PATH_BLOB_DELETED,
)
from local_deep_research.database.models.base import Base


@pytest.fixture
def test_engine():
    """Create an in-memory SQLite database for testing."""
    engine = create_engine("sqlite:///:memory:")

    # Enable foreign key enforcement for SQLite (required for CASCADE to work)
    @event.listens_for(engine, "connect")
    def set_sqlite_pragma(dbapi_connection, connection_record):
        cursor = dbapi_connection.cursor()
        cursor.execute("PRAGMA foreign_keys=ON")
        cursor.close()

    Base.metadata.create_all(engine)
    yield engine
    engine.dispose()


@pytest.fixture
def test_session(test_engine):
    """Create a session for testing."""
    Session = sessionmaker(bind=test_engine)
    session = Session()
    yield session
    session.close()


@pytest.fixture
def source_type(test_session):
    """Create a source type for testing."""
    st = SourceType(
        id=str(uuid.uuid4()),
        name="user_upload",
        display_name="User Upload",
    )
    test_session.add(st)
    test_session.commit()
    return st


@pytest.fixture
def test_collection(test_session):
    """Create a test collection."""
    collection = Collection(
        id=str(uuid.uuid4()),
        name="Test Collection",
        description="A test collection",
    )
    test_session.add(collection)
    test_session.commit()
    return collection


@pytest.fixture
def test_document(test_session, source_type, test_collection):
    """Create a test document with blob and collection link."""
    doc_id = str(uuid.uuid4())
    doc = Document(
        id=doc_id,
        source_type_id=source_type.id,
        document_hash=f"hash_{doc_id[:8]}",
        file_size=1024,
        file_type="pdf",
        title="Test Document",
        filename="test.pdf",
        storage_mode="database",
        text_content="This is test content.",
    )
    test_session.add(doc)
    test_session.flush()

    # Add blob
    blob = DocumentBlob(
        document_id=doc_id,
        pdf_binary=b"fake pdf content",
        blob_hash="fake_hash",
    )
    test_session.add(blob)

    # Add to collection
    doc_collection = DocumentCollection(
        document_id=doc_id,
        collection_id=test_collection.id,
        indexed=False,
    )
    test_session.add(doc_collection)

    test_session.commit()
    return doc


class TestDocumentDeletionService:
    """Unit tests for DocumentDeletionService."""

    def test_delete_document_removes_blob(self, test_session, test_document):
        """Verify DocumentBlob is deleted with document."""
        doc_id = test_document.id

        # Verify blob exists before deletion
        blob_before = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        assert blob_before is not None, "Blob should exist before deletion"

        # Delete document (manually simulating service behavior)
        doc = test_session.get(Document, doc_id)
        test_session.delete(doc)
        test_session.commit()

        # Verify blob is deleted (CASCADE should handle this)
        blob_after = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        assert blob_after is None, "Blob should be deleted with document"

    def test_delete_document_removes_collection_links(
        self, test_session, test_document, test_collection
    ):
        """Verify DocumentCollection links are deleted."""
        doc_id = test_document.id

        # Verify link exists before deletion
        link_before = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id)
            .first()
        )
        assert link_before is not None, "Collection link should exist"

        # Delete document
        doc = test_session.get(Document, doc_id)
        test_session.delete(doc)
        test_session.commit()

        # Verify link is deleted
        link_after = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id)
            .first()
        )
        assert link_after is None, "Collection link should be deleted"

    def test_delete_blob_keeps_document(self, test_session, test_document):
        """Verify document metadata and text_content preserved after blob deletion."""
        doc_id = test_document.id

        # Delete only the blob
        blob = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        test_session.delete(blob)
        test_session.commit()

        # Verify document still exists with text
        doc = test_session.get(Document, doc_id)
        assert doc is not None, "Document should still exist"
        assert doc.text_content is not None, "Text content should be preserved"
        assert doc.title == "Test Document", "Title should be preserved"

    def test_remove_from_collection_deletes_orphan(
        self, test_session, test_document, test_collection
    ):
        """Document in single collection: deleted completely."""
        doc_id = test_document.id

        # Remove the collection link (simulate remove_from_collection)
        link = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id, collection_id=test_collection.id)
            .first()
        )
        test_session.delete(link)
        test_session.flush()

        # Check if document is in any collection
        remaining = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id)
            .count()
        )
        assert remaining == 0, "Document should not be in any collection"

        # If orphaned, delete document
        if remaining == 0:
            doc = test_session.get(Document, doc_id)
            test_session.delete(doc)
        test_session.commit()

        # Verify document is deleted
        doc_after = test_session.get(Document, doc_id)
        assert doc_after is None, "Orphaned document should be deleted"

    def test_remove_from_collection_keeps_if_in_other(
        self, test_session, test_document, test_collection
    ):
        """Document in multiple collections: only unlinked."""
        doc_id = test_document.id

        # Create another collection and link document to it
        other_collection = Collection(
            id=str(uuid.uuid4()),
            name="Other Collection",
        )
        test_session.add(other_collection)
        test_session.flush()

        other_link = DocumentCollection(
            document_id=doc_id,
            collection_id=other_collection.id,
            indexed=False,
        )
        test_session.add(other_link)
        test_session.commit()

        # Remove from first collection
        link = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id, collection_id=test_collection.id)
            .first()
        )
        test_session.delete(link)
        test_session.commit()

        # Document should still exist (in other collection)
        doc = test_session.get(Document, doc_id)
        assert doc is not None, "Document should exist (in other collection)"

        # Document should only be in other collection
        links = (
            test_session.query(DocumentCollection)
            .filter_by(document_id=doc_id)
            .all()
        )
        assert len(links) == 1, "Document should be in exactly one collection"
        assert links[0].collection_id == other_collection.id


class TestBlobDeletion:
    """Unit tests for blob-only deletion."""

    def test_delete_blob_updates_storage_mode(
        self, test_session, test_document
    ):
        """Verify storage_mode set to 'none' after blob deletion."""
        doc_id = test_document.id

        # Delete blob and update storage mode
        blob = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        test_session.delete(blob)

        doc = test_session.get(Document, doc_id)
        doc.storage_mode = "none"
        doc.file_path = FILE_PATH_BLOB_DELETED
        test_session.commit()

        # Verify update
        doc_after = test_session.get(Document, doc_id)
        assert doc_after.storage_mode == "none"
        assert doc_after.file_path == FILE_PATH_BLOB_DELETED

    def test_delete_blob_frees_bytes(self, test_session, test_document):
        """Verify bytes freed calculation is correct."""
        doc_id = test_document.id

        # Get blob size
        blob = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        blob_size = len(blob.pdf_binary) if blob.pdf_binary else 0
        assert blob_size > 0, "Blob should have content"

        # Delete blob
        test_session.delete(blob)
        test_session.commit()

        # Verify blob is deleted
        blob_after = (
            test_session.query(DocumentBlob)
            .filter_by(document_id=doc_id)
            .first()
        )
        assert blob_after is None


class TestDocumentChunkCleanup:
    """Tests for DocumentChunk cleanup (no FK constraint)."""

    def test_chunks_require_manual_cleanup(
        self, test_session, test_document, test_collection
    ):
        """Verify DocumentChunks must be manually deleted (no FK)."""
        doc_id = test_document.id
        collection_name = f"collection_{test_collection.id}"

        # Create some chunks for the document
        for i in range(3):
            chunk = DocumentChunk(
                chunk_hash=f"chunk_hash_{i}_{uuid.uuid4().hex[:8]}",
                source_type="document",
                source_id=doc_id,
                collection_name=collection_name,
                chunk_text=f"Test chunk {i} content",
                chunk_index=i,
                start_char=i * 100,
                end_char=(i + 1) * 100,
                word_count=10,
                embedding_id=str(uuid.uuid4()),
                embedding_model="test-model",
                embedding_model_type="sentence_transformers",
            )
            test_session.add(chunk)
        test_session.commit()

        # Verify chunks exist
        chunks_before = (
            test_session.query(DocumentChunk)
            .filter_by(source_id=doc_id, source_type="document")
            .count()
        )
        assert chunks_before == 3, "Should have 3 chunks"

        # Delete document WITHOUT manually deleting chunks first
        doc = test_session.get(Document, doc_id)
        test_session.delete(doc)
        test_session.commit()

        # Chunks should still exist (no FK constraint!)
        chunks_after = (
            test_session.query(DocumentChunk)
            .filter_by(source_id=doc_id, source_type="document")
            .count()
        )
        # This shows the orphan problem - chunks remain after document deletion
        assert chunks_after == 3, "Chunks remain orphaned (no FK)"

    def test_manual_chunk_deletion(
        self, test_session, test_document, test_collection
    ):
        """Verify manual chunk deletion works."""
        doc_id = test_document.id
        collection_name = f"collection_{test_collection.id}"

        # Create a chunk
        chunk = DocumentChunk(
            chunk_hash=f"chunk_hash_{uuid.uuid4().hex[:8]}",
            source_type="document",
            source_id=doc_id,
            collection_name=collection_name,
            chunk_text="Test chunk content",
            chunk_index=0,
            start_char=0,
            end_char=100,
            word_count=10,
            embedding_id=str(uuid.uuid4()),
            embedding_model="test-model",
            embedding_model_type="sentence_transformers",
        )
        test_session.add(chunk)
        test_session.commit()

        # Manually delete chunks before document
        deleted = (
            test_session.query(DocumentChunk)
            .filter(
                DocumentChunk.source_id == doc_id,
                DocumentChunk.source_type == "document",
            )
            .delete(synchronize_session=False)
        )
        assert deleted == 1, "Should delete 1 chunk"

        # Now delete document
        doc = test_session.get(Document, doc_id)
        test_session.delete(doc)
        test_session.commit()

        # Verify all cleaned up
        doc_after = test_session.get(Document, doc_id)
        assert doc_after is None