kodit 0.2.5__py3-none-any.whl → 0.2.7__py3-none-any.whl

kodit/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '0.2.5'
-__version_tuple__ = version_tuple = (0, 2, 5)
+__version__ = version = '0.2.7'
+__version_tuple__ = version_tuple = (0, 2, 7)

kodit/application/services/indexing_application_service.py

@@ -1,6 +1,7 @@
 """Application service for indexing operations."""
 
 import structlog
+from sqlalchemy.ext.asyncio import AsyncSession
 
 from kodit.application.commands.snippet_commands import CreateIndexSnippetsCommand
 from kodit.application.services.snippet_application_service import (
@@ -8,6 +9,7 @@ from kodit.application.services.snippet_application_service import (
 )
 from kodit.domain.entities import Snippet
 from kodit.domain.enums import SnippetExtractionStrategy
+from kodit.domain.errors import EmptySourceError
 from kodit.domain.interfaces import ProgressCallback
 from kodit.domain.services.bm25_service import BM25DomainService
 from kodit.domain.services.embedding_service import EmbeddingDomainService
@@ -51,6 +53,7 @@ class IndexingApplicationService:
         text_search_service: EmbeddingDomainService,
         enrichment_service: EnrichmentDomainService,
         snippet_application_service: SnippetApplicationService,
+        session: AsyncSession,
     ) -> None:
         """Initialize the indexing application service.
 
@@ -62,11 +65,13 @@ class IndexingApplicationService:
             text_search_service: The text search domain service.
             enrichment_service: The enrichment domain service.
             snippet_application_service: The snippet application service.
+            session: The database session for transaction management.
 
         """
         self.indexing_domain_service = indexing_domain_service
         self.source_service = source_service
         self.snippet_application_service = snippet_application_service
+        self.session = session
         self.log = structlog.get_logger(__name__)
         self.bm25_service = bm25_service
         self.code_search_service = code_search_service
@@ -93,7 +98,12 @@ class IndexingApplicationService:
 
         # Create the index
         request = IndexCreateRequest(source_id=source.id)
-        return await self.indexing_domain_service.create_index(request)
+        index_view = await self.indexing_domain_service.create_index(request)
+
+        # Commit the index creation
+        await self.session.commit()
+
+        return index_view
 
     async def list_indexes(self) -> list[IndexView]:
         """List all available indexes with their details.
@@ -125,7 +135,7 @@ class IndexingApplicationService:
             progress_callback: Optional progress callback for reporting progress.
 
         Raises:
-            ValueError: If the index doesn't exist.
+            ValueError: If the index doesn't exist or no indexable snippets are found.
 
         """
         log_event("kodit.index.run")
@@ -138,8 +148,11 @@ class IndexingApplicationService:
 
         # Delete old snippets so we don't duplicate
         await self.indexing_domain_service.delete_all_snippets(index.id)
+        # Commit the deletion
+        await self.session.commit()
 
         # Create snippets for supported file types using the snippet application service
+        # (snippet_application_service handles its own commits)
         self.log.info("Creating snippets for files", index_id=index.id)
         command = CreateIndexSnippetsCommand(
             index_id=index.id, strategy=SnippetExtractionStrategy.METHOD_BASED
@@ -150,6 +163,11 @@ class IndexingApplicationService:
 
         snippets = await self.indexing_domain_service.get_snippets_for_index(index.id)
 
+        # Check if any snippets were extracted
+        if not snippets:
+            msg = f"No indexable snippets found for index {index.id}"
+            raise EmptySourceError(msg)
+
         # Create BM25 index
         self.log.info("Creating keyword index")
         reporter = Reporter(self.log, progress_callback)
@@ -184,6 +202,8 @@ class IndexingApplicationService:
 
         # Update index timestamp
         await self.indexing_domain_service.update_index_timestamp(index.id)
+        # Commit the timestamp update
+        await self.session.commit()
 
     async def _create_bm25_index(
         self, snippets: list[Snippet], progress_callback: ProgressCallback | None = None
@@ -264,6 +284,10 @@ class IndexingApplicationService:
                 "enrichment", processed, len(snippets), "Enriching snippets..."
             )
 
+        # Commit all snippet content updates as a single transaction
+        if enriched_contents:
+            await self.session.commit()
+
         await reporter.done("enrichment")
 
     async def _create_text_embeddings(

kodit/application/services/snippet_application_service.py

@@ -4,6 +4,7 @@ from pathlib import Path
 from typing import Any
 
 import structlog
+from sqlalchemy.ext.asyncio import AsyncSession
 
 from kodit.application.commands.snippet_commands import (
     CreateIndexSnippetsCommand,
@@ -28,6 +29,7 @@ class SnippetApplicationService:
         snippet_extraction_service: SnippetExtractionDomainService,
         snippet_repository: SnippetRepository,
         file_repository: FileRepository,
+        session: AsyncSession,
     ) -> None:
         """Initialize the snippet application service.
 
@@ -35,11 +37,13 @@ class SnippetApplicationService:
             snippet_extraction_service: Domain service for snippet extraction
             snippet_repository: Repository for snippet persistence
             file_repository: Repository for file operations
+            session: The database session for transaction management
 
         """
         self.snippet_extraction_service = snippet_extraction_service
         self.snippet_repository = snippet_repository
         self.file_repository = file_repository
+        self.session = session
         self.log = structlog.get_logger(__name__)
 
     async def extract_snippets_from_file(
@@ -140,4 +144,6 @@ class SnippetApplicationService:
                 message=f"Processing {file.cloned_path}...",
             )
 
+        # Commit all snippet creations in a single transaction
+        await self.session.commit()
         await reporter.done("create_snippets")

kodit/cli.py

@@ -19,6 +19,7 @@ from kodit.config import (
     with_app_context,
     with_session,
 )
+from kodit.domain.errors import EmptySourceError
 from kodit.domain.services.source_service import SourceService
 from kodit.domain.value_objects import MultiSearchRequest
 from kodit.infrastructure.indexing.indexing_factory import (
@@ -58,6 +59,7 @@ def create_snippet_application_service(
         snippet_extraction_service=snippet_extraction_service,
         snippet_repository=snippet_repository,
         file_repository=file_repository,
+        session=session,
     )
 
 
@@ -100,6 +102,7 @@ async def index(
     sources: list[str],
 ) -> None:
     """List indexes, or index data sources."""
+    log = structlog.get_logger(__name__)
     source_service = SourceService(
         clone_dir=app_context.get_clone_dir(),
         session_factory=lambda: session,
@@ -152,7 +155,17 @@ async def index(
 
         # Create a new progress callback for the indexing operations
         indexing_progress_callback = create_multi_stage_progress_callback()
-        await service.run_index(index.id, indexing_progress_callback)
+        try:
+            await service.run_index(index.id, indexing_progress_callback)
+        except EmptySourceError as e:
+            log.exception("Empty source error", error=e)
+            msg = f"""{e}. This could mean:
+• The repository contains no supported file types
+• All files are excluded by ignore patterns
+• The files contain no extractable code snippets
+Please check the repository contents and try again.
+"""
+            click.echo(msg)
 
 
 @cli.group()

kodit/domain/errors.py

@@ -0,0 +1,5 @@
+"""Domain errors."""
+
+
+class EmptySourceError(ValueError):
+    """Error raised when a source is empty."""

kodit/domain/services/source_service.py

@@ -49,7 +49,7 @@ class SourceService:
         """Create a source."""
         async with self._session_factory() as session:
             repo = SqlAlchemySourceRepository(session)
-            git_factory, folder_factory = self._build_factories(repo)
+            git_factory, folder_factory = self._build_factories(repo, session)
 
             if is_valid_clone_target(uri_or_path_like):
                 source = await git_factory.create(uri_or_path_like, progress_callback)
@@ -60,17 +60,18 @@ class SourceService:
             else:
                 raise ValueError(f"Unsupported source: {uri_or_path_like}")
 
-            await session.commit()  # one commit for all work
+            # Factories handle their own commits now
             return source
 
     def _build_factories(
-        self, repository: SourceRepository
+        self, repository: SourceRepository, session: AsyncSession
     ) -> tuple[GitSourceFactory, FolderSourceFactory]:
         # Git-specific collaborators
         git_wc = GitWorkingCopyProvider(self.clone_dir)
         git_factory = GitSourceFactory(
             repository=repository,
             working_copy=git_wc,
+            session=session,
         )
 
         # Folder-specific collaborators
@@ -78,6 +79,7 @@ class SourceService:
         folder_factory = FolderSourceFactory(
             repository=repository,
             working_copy=fold_wc,
+            session=session,
         )
 
         return git_factory, folder_factory

kodit/infrastructure/cloning/folder/factory.py

@@ -3,6 +3,7 @@
 from pathlib import Path
 
 import structlog
+from sqlalchemy.ext.asyncio import AsyncSession
 
 from kodit.domain.entities import AuthorFileMapping, Source, SourceType
 from kodit.domain.interfaces import NullProgressCallback, ProgressCallback
@@ -22,6 +23,7 @@ class FolderSourceFactory:
         self,
         repository: SourceRepository,
         working_copy: FolderWorkingCopyProvider,
+        session: AsyncSession,
     ) -> None:
         """Initialize the source factory."""
         self.log = structlog.get_logger(__name__)
@@ -29,6 +31,7 @@ class FolderSourceFactory:
         self.working_copy = working_copy
         self.metadata_extractor = FolderFileMetadataExtractor()
         self.author_extractor = NoOpAuthorExtractor()
+        self.session = session
 
     async def create(
         self, uri: str, progress_callback: ProgressCallback | None = None
@@ -55,7 +58,7 @@ class FolderSourceFactory:
         clone_path = await self.working_copy.prepare(directory.as_uri())
 
         # Create source record
-        source = await self.repository.create_source(
+        source = await self.repository.save(
             Source(
                 uri=directory.as_uri(),
                 cloned_path=str(clone_path),
@@ -63,12 +66,18 @@ class FolderSourceFactory:
             )
         )
 
+        # Commit source creation so we get an ID for foreign key relationships
+        await self.session.commit()
+
         # Get all files to process
         files = [f for f in clone_path.rglob("*") if f.is_file()]
 
         # Process files
         await self._process_files(source, files, progress_callback)
 
+        # Commit file processing
+        await self.session.commit()
+
         return source
 
     async def _process_files(

kodit/infrastructure/cloning/git/factory.py

@@ -5,6 +5,7 @@ from pathlib import Path
 
 import git
 import structlog
+from sqlalchemy.ext.asyncio import AsyncSession
 
 from kodit.domain.entities import AuthorFileMapping, Source, SourceType
 from kodit.domain.interfaces import NullProgressCallback, ProgressCallback
@@ -16,6 +17,7 @@ from kodit.infrastructure.cloning.metadata import (
     GitAuthorExtractor,
     GitFileMetadataExtractor,
 )
+from kodit.infrastructure.git.git_utils import sanitize_git_url
 from kodit.infrastructure.ignore.ignore_pattern_provider import GitIgnorePatternProvider
 
 
@@ -26,6 +28,7 @@ class GitSourceFactory:
         self,
         repository: SourceRepository,
         working_copy: GitWorkingCopyProvider,
+        session: AsyncSession,
     ) -> None:
         """Initialize the source factory."""
         self.log = structlog.get_logger(__name__)
@@ -33,6 +36,7 @@ class GitSourceFactory:
         self.working_copy = working_copy
         self.metadata_extractor = GitFileMetadataExtractor()
         self.author_extractor = GitAuthorExtractor(repository)
+        self.session = session
 
     async def create(
         self, uri: str, progress_callback: ProgressCallback | None = None
@@ -43,33 +47,41 @@ class GitSourceFactory:
             progress_callback = NullProgressCallback()
 
         # Normalize the URI
-        self.log.debug("Normalising git uri", uri=uri)
+        # Never log the raw URI in production
+        self.log.debug("Normalising git uri", uri="[REDACTED]" + uri[-4:])
         with tempfile.TemporaryDirectory() as temp_dir:
             git.Repo.clone_from(uri, temp_dir)
             remote = git.Repo(temp_dir).remote()
             uri = remote.url
 
+        # Sanitize the URI to remove any credentials
+        sanitized_uri = sanitize_git_url(uri)
+        self.log.debug("Sanitized git uri", sanitized_uri=sanitized_uri)
+
         # Check if source already exists
-        self.log.debug("Checking if source already exists", uri=uri)
-        source = await self.repository.get_by_uri(uri)
+        self.log.debug("Checking if source already exists", uri=sanitized_uri)
+        source = await self.repository.get_by_uri(sanitized_uri)
 
         if source:
             self.log.info("Source already exists, reusing...", source_id=source.id)
             return source
 
-        # Prepare working copy
+        # Prepare working copy (use original URI for cloning, sanitized for storage)
         clone_path = await self.working_copy.prepare(uri)
 
         # Create source record
-        self.log.debug("Creating source", uri=uri, clone_path=str(clone_path))
-        source = await self.repository.create_source(
+        self.log.debug("Creating source", uri=sanitized_uri, clone_path=str(clone_path))
+        source = await self.repository.save(
             Source(
-                uri=uri,
+                uri=sanitized_uri,
                 cloned_path=str(clone_path),
                 source_type=SourceType.GIT,
             )
         )
 
+        # Commit source creation so we get an ID for foreign key relationships
+        await self.session.commit()
+
         # Get files to process using ignore patterns
         ignore_provider = GitIgnorePatternProvider(clone_path)
         ignore_service = IgnoreService(ignore_provider)
@@ -83,6 +95,9 @@ class GitSourceFactory:
         self.log.info("Inspecting files", source_id=source.id, num_files=len(files))
         await self._process_files(source, files, progress_callback)
 
+        # Commit file processing
+        await self.session.commit()
+
         return source
 
     async def _process_files(
@@ -111,6 +126,11 @@ class GitSourceFactory:
 
             # Extract authors
             authors = await self.author_extractor.extract(path, source)
+
+            # Commit authors so they get IDs before creating mappings
+            if authors:
+                await self.session.commit()
+
             for author in authors:
                 await self.repository.upsert_author_file_mapping(
                     AuthorFileMapping(

kodit/infrastructure/cloning/git/working_copy.py

@@ -5,6 +5,8 @@ from pathlib import Path
 import git
 import structlog
 
+from kodit.infrastructure.git.git_utils import sanitize_git_url
+
 
 class GitWorkingCopyProvider:
     """Working copy provider for git-based sources."""
@@ -16,17 +18,23 @@ class GitWorkingCopyProvider:
 
     async def prepare(self, uri: str) -> Path:
         """Prepare a Git working copy."""
-        # Create a unique directory name for the clone
-        clone_path = self.clone_dir / uri.replace("/", "_").replace(":", "_")
+        # Sanitize the URI for directory name to prevent credential leaks
+        sanitized_uri = sanitize_git_url(uri)
+
+        # Create a unique directory name for the clone using the sanitized URI
+        clone_path = self.clone_dir / sanitized_uri.replace("/", "_").replace(":", "_")
         clone_path.mkdir(parents=True, exist_ok=True)
 
         try:
-            self.log.info("Cloning repository", uri=uri, clone_path=str(clone_path))
+            self.log.info(
+                "Cloning repository", uri=sanitized_uri, clone_path=str(clone_path)
+            )
+            # Use the original URI for cloning (with credentials if present)
             git.Repo.clone_from(uri, clone_path)
         except git.GitCommandError as e:
             if "already exists and is not an empty directory" not in str(e):
                 msg = f"Failed to clone repository: {e}"
                 raise ValueError(msg) from e
-            self.log.info("Repository already exists, reusing...", uri=uri)
+            self.log.info("Repository already exists, reusing...", uri=sanitized_uri)
 
         return clone_path

kodit/infrastructure/git/git_utils.py

@@ -1,6 +1,7 @@
 """Git utilities for infrastructure operations."""
 
 import tempfile
+from urllib.parse import urlparse, urlunparse
 
 import git
 
@@ -22,3 +23,65 @@ def is_valid_clone_target(target: str) -> bool:
             return False
         else:
             return True
+
+
+def sanitize_git_url(url: str) -> str:
+    """Remove credentials from a git URL while preserving the rest of the URL structure.
+
+    This function handles various git URL formats:
+    - HTTPS URLs with username:password@host
+    - HTTPS URLs with username@host (no password)
+    - SSH URLs (left unchanged)
+    - File URLs (left unchanged)
+
+    Args:
+        url: The git URL that may contain credentials.
+
+    Returns:
+        The sanitized URL with credentials removed.
+
+    Examples:
+        >>> sanitize_git_url("https://phil:token@dev.azure.com/org/project/_git/repo")
+        "https://dev.azure.com/org/project/_git/repo"
+        >>> sanitize_git_url("https://username@github.com/user/repo.git")
+        "https://github.com/user/repo.git"
+        >>> sanitize_git_url("git@github.com:user/repo.git")
+        "git@github.com:user/repo.git"
+
+    """
+    # Handle SSH URLs (they don't have credentials in the URL format)
+    if url.startswith(("git@", "ssh://")):
+        return url
+
+    # Handle file URLs
+    if url.startswith("file://"):
+        return url
+
+    try:
+        # Parse the URL
+        parsed = urlparse(url)
+
+        # If there are no credentials, return the URL as-is
+        if not parsed.username:
+            return url
+
+        # Reconstruct the URL without credentials
+        # Keep scheme, netloc (without username/password), path, params, query, fragment
+        sanitized_netloc = parsed.hostname
+        if parsed.port:
+            sanitized_netloc = f"{parsed.hostname}:{parsed.port}"
+
+        return urlunparse(
+            (
+                parsed.scheme,
+                sanitized_netloc,
+                parsed.path,
+                parsed.params,
+                parsed.query,
+                parsed.fragment,
+            )
+        )
+
+    except Exception:  # noqa: BLE001
+        # If URL parsing fails, return the original URL
+        return url

kodit/infrastructure/indexing/index_repository.py

@@ -42,7 +42,6 @@ class SQLAlchemyIndexRepository(IndexRepository):
 
         index = Index(source_id=source_id)
         self.session.add(index)
-        await self.session.commit()
 
         # Get source for the view
         source_query = select(Source).where(Source.id == source_id)
@@ -158,7 +157,6 @@ class SQLAlchemyIndexRepository(IndexRepository):
 
         if index:
             index.updated_at = datetime.now(UTC)
-            await self.session.commit()
 
     async def delete_all_snippets(self, index_id: int) -> None:
         """Delete all snippets for an index.
@@ -178,7 +176,6 @@ class SQLAlchemyIndexRepository(IndexRepository):
         # Now delete the snippets
         query = delete(Snippet).where(Snippet.index_id == index_id)
         await self.session.execute(query)
-        await self.session.commit()
 
     async def get_snippets_for_index(self, index_id: int) -> list[Snippet]:
         """Get all snippets for an index.
@@ -207,7 +204,6 @@ class SQLAlchemyIndexRepository(IndexRepository):
             content=snippet["content"],
         )
         self.session.add(db_snippet)
-        await self.session.commit()
 
     async def update_snippet_content(self, snippet_id: int, content: str) -> None:
         """Update the content of an existing snippet.
@@ -224,7 +220,6 @@ class SQLAlchemyIndexRepository(IndexRepository):
         if snippet:
             snippet.content = content
             # SQLAlchemy will automatically track this change
-            await self.session.commit()
 
     async def list_snippets_by_ids(self, ids: list[int]) -> list[tuple[dict, dict]]:
         """List snippets by IDs.

kodit/infrastructure/indexing/indexing_factory.py

@@ -53,6 +53,7 @@ def create_snippet_application_service(
         snippet_extraction_service=snippet_extraction_service,
         snippet_repository=snippet_repository,
         file_repository=file_repository,
+        session=session,
     )
 
 
@@ -108,4 +109,5 @@ def create_indexing_application_service(
         text_search_service=text_search_service,
         enrichment_service=enrichment_service,
         snippet_application_service=snippet_application_service,
+        session=session,
     )

kodit/infrastructure/sqlalchemy/embedding_repository.py

@@ -30,7 +30,6 @@ class SqlAlchemyEmbeddingRepository:
 
         """
         self.session.add(embedding)
-        await self.session.commit()
         return embedding
 
     async def get_embedding_by_snippet_id_and_type(
@@ -81,7 +80,6 @@ class SqlAlchemyEmbeddingRepository:
         embeddings = result.scalars().all()
         for embedding in embeddings:
             await self.session.delete(embedding)
-        await self.session.commit()
 
     async def list_semantic_results(
         self, embedding_type: EmbeddingType, embedding: list[float], top_k: int = 10

kodit/infrastructure/sqlalchemy/file_repository.py

@@ -21,6 +21,25 @@ class SqlAlchemyFileRepository(FileRepository):
         """
         self.session = session
 
+    async def get(self, id: int) -> File | None:  # noqa: A002
+        """Get a file by ID."""
+        return await self.session.get(File, id)
+
+    async def save(self, entity: File) -> File:
+        """Save entity."""
+        self.session.add(entity)
+        return entity
+
+    async def delete(self, id: int) -> None:  # noqa: A002
+        """Delete entity by ID."""
+        file = await self.get(id)
+        if file:
+            await self.session.delete(file)
+
+    async def list(self) -> Sequence[File]:
+        """List all entities."""
+        return (await self.session.scalars(select(File))).all()
+
     async def get_files_for_index(self, index_id: int) -> Sequence[File]:
         """Get all files for an index.
 
@@ -57,17 +76,3 @@ class SqlAlchemyFileRepository(FileRepository):
         query = select(File).where(File.id == file_id)
         result = await self.session.execute(query)
         return result.scalar_one_or_none()
-
-    async def save(self, file: File) -> File:
-        """Save file using SQLAlchemy.
-
-        Args:
-            file: The file to save
-
-        Returns:
-            The saved file
-
-        """
-        self.session.add(file)
-        await self.session.commit()
-        return file

kodit/infrastructure/sqlalchemy/repository.py

@@ -1,6 +1,7 @@
 """SQLAlchemy repository."""
 
 from collections.abc import Sequence
+from typing import cast
 
 from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -16,41 +17,43 @@ class SqlAlchemySourceRepository(SourceRepository):
         """Initialize the repository."""
         self._session = session
 
-    async def get(self, source_id: int) -> Source | None:
+    async def get(self, id: int) -> Source | None:  # noqa: A002
         """Get a source by ID."""
-        return await self._session.get(Source, source_id)
+        return await self._session.get(Source, id)
+
+    async def save(self, entity: Source) -> Source:
+        """Save entity."""
+        self._session.add(entity)
+        return entity
+
+    async def delete(self, id: int) -> None:  # noqa: A002
+        """Delete entity by ID."""
+        source = await self.get(id)
+        if source:
+            await self._session.delete(source)
+
+    async def list(self) -> Sequence[Source]:
+        """List all entities."""
+        stmt = select(Source)
+        return (await self._session.scalars(stmt)).all()
 
     async def get_by_uri(self, uri: str) -> Source | None:
         """Get a source by URI."""
         stmt = select(Source).where(Source.uri == uri)
-        return await self._session.scalar(stmt)  # None if no row
+        return cast("Source | None", await self._session.scalar(stmt))
 
-    async def list(self, *, source_type: SourceType | None = None) -> Sequence[Source]:
-        """List sources."""
+    async def list_by_type(
+        self, source_type: SourceType | None = None
+    ) -> Sequence[Source]:
+        """List sources by type."""
         stmt = select(Source)
         if source_type is not None:
             stmt = stmt.where(Source.type == source_type)
         return (await self._session.scalars(stmt)).all()
 
-    async def add(self, source: Source) -> None:
-        """Add a source."""
-        self._session.add(source)  # INSERT on flush
-        await self._session.flush()  # Flush to get the ID
-
-    async def create_source(self, source: Source) -> Source:
-        """Create a source and commit it."""
-        self._session.add(source)
-        await self._session.commit()
-        return source
-
-    async def remove(self, source: Source) -> None:
-        """Remove a source."""
-        await self._session.delete(source)  # DELETE on flush
-
     async def create_file(self, file: File) -> File:
         """Create a new file record."""
         self._session.add(file)
-        await self._session.commit()
         return file
 
     async def upsert_author(self, author: Author) -> Author:
@@ -59,14 +62,13 @@ class SqlAlchemySourceRepository(SourceRepository):
         stmt = select(Author).where(
             Author.name == author.name, Author.email == author.email
         )
-        existing_author = await self._session.scalar(stmt)
+        existing_author = cast("Author | None", await self._session.scalar(stmt))
 
         if existing_author:
             return existing_author
 
         # Author doesn't exist, create new one
         self._session.add(author)
-        await self._session.commit()
         return author
 
     async def upsert_author_file_mapping(
@@ -78,14 +80,15 @@ class SqlAlchemySourceRepository(SourceRepository):
             AuthorFileMapping.author_id == mapping.author_id,
             AuthorFileMapping.file_id == mapping.file_id,
         )
-        existing_mapping = await self._session.scalar(stmt)
+        existing_mapping = cast(
+            "AuthorFileMapping | None", await self._session.scalar(stmt)
+        )
 
         if existing_mapping:
             return existing_mapping
 
         # Mapping doesn't exist, create new one
         self._session.add(mapping)
-        await self._session.commit()
         return mapping
 
 
@@ -96,26 +99,35 @@ class SqlAlchemyAuthorRepository(AuthorRepository):
         """Initialize the repository."""
         self._session = session
 
-    async def get(self, author_id: int) -> Author | None:
+    async def get(self, id: int) -> Author | None:  # noqa: A002
         """Get an author by ID."""
-        return await self._session.get(Author, author_id)
+        return await self._session.get(Author, id)
 
-    async def get_by_name(self, name: str) -> Author | None:
-        """Get an author by name."""
-        return await self._session.scalar(select(Author).where(Author.name == name))
+    async def save(self, entity: Author) -> Author:
+        """Save entity."""
+        self._session.add(entity)
+        return entity
 
-    async def get_by_email(self, email: str) -> Author | None:
-        """Get an author by email."""
-        return await self._session.scalar(select(Author).where(Author.email == email))
+    async def delete(self, id: int) -> None:  # noqa: A002
+        """Delete entity by ID."""
+        author = await self.get(id)
+        if author:
+            await self._session.delete(author)
 
     async def list(self) -> Sequence[Author]:
         """List authors."""
         return (await self._session.scalars(select(Author))).all()
 
-    async def add(self, author: Author) -> None:
-        """Add an author."""
-        self._session.add(author)
+    async def get_by_name(self, name: str) -> Author | None:
+        """Get an author by name."""
+        return cast(
+            "Author | None",
+            await self._session.scalar(select(Author).where(Author.name == name)),
+        )
 
-    async def remove(self, author: Author) -> None:
-        """Remove an author."""
-        await self._session.delete(author)
+    async def get_by_email(self, email: str) -> Author | None:
+        """Get an author by email."""
+        return cast(
+            "Author | None",
+            await self._session.scalar(select(Author).where(Author.email == email)),
+        )

kodit/infrastructure/sqlalchemy/snippet_repository.py

@@ -21,19 +21,24 @@ class SqlAlchemySnippetRepository(SnippetRepository):
         """
         self.session = session
 
-    async def save(self, snippet: Snippet) -> Snippet:
-        """Save snippet using SQLAlchemy.
+    async def get(self, id: int) -> Snippet | None:  # noqa: A002
+        """Get a snippet by ID."""
+        return await self.session.get(Snippet, id)
 
-        Args:
-            snippet: The snippet to save
+    async def save(self, entity: Snippet) -> Snippet:
+        """Save entity."""
+        self.session.add(entity)
+        return entity
 
-        Returns:
-            The saved snippet
+    async def delete(self, id: int) -> None:  # noqa: A002
+        """Delete entity by ID."""
+        snippet = await self.get(id)
+        if snippet:
+            await self.session.delete(snippet)
 
-        """
-        self.session.add(snippet)
-        await self.session.commit()
-        return snippet
+    async def list(self) -> Sequence[Snippet]:
+        """List all entities."""
+        return (await self.session.scalars(select(Snippet))).all()
 
     async def get_by_id(self, snippet_id: int) -> Snippet | None:
         """Get a snippet by ID.
@@ -72,4 +77,3 @@ class SqlAlchemySnippetRepository(SnippetRepository):
         """
         query = delete(Snippet).where(Snippet.index_id == index_id)
         await self.session.execute(query)
-        await self.session.commit()

kodit/mcp.py

@@ -97,6 +97,7 @@ def create_snippet_application_service(
         snippet_extraction_service=snippet_extraction_service,
         snippet_repository=snippet_repository,
         file_repository=file_repository,
+        session=session,
     )
 
 

{kodit-0.2.5.dist-info → kodit-0.2.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: kodit
-Version: 0.2.5
+Version: 0.2.7
 Summary: Code indexing for better AI code generation
 Project-URL: Homepage, https://docs.helixml.tech/kodit/
 Project-URL: Documentation, https://docs.helixml.tech/kodit/

{kodit-0.2.5.dist-info → kodit-0.2.7.dist-info}/RECORD

@@ -1,23 +1,24 @@
 kodit/.gitignore,sha256=ztkjgRwL9Uud1OEi36hGQeDGk3OLK1NfDEO8YqGYy8o,11
 kodit/__init__.py,sha256=aEKHYninUq1yh6jaNfvJBYg-6fenpN132nJt1UU6Jxs,59
-kodit/_version.py,sha256=N3oBwJUFmS-AwCjqOcSlRW4GvSq-uJJMaBvoGfv1-hM,511
+kodit/_version.py,sha256=Xk20v7uvkFqkpy9aLJzVngs1eKQn0FYUP2oyA1MEQUU,511
 kodit/app.py,sha256=qKBWJ0VNSY_M6G3VFfAQ0133q5bnS99cUFD0p396taw,1032
-kodit/cli.py,sha256=SUoo9R-Jut0OlZLajD-nSPdhvHPOZ2SpN35SDyON8tc,11500
+kodit/cli.py,sha256=JnhTlG1s04O0m8AzsBdrwP8T_BqSZMPXnRLvI7T_Gxc,12004
 kodit/config.py,sha256=3yh7hfLSILjZK_qJMhcExwRcrWJ0b5Eb1JjjOvMPJZo,4146
 kodit/database.py,sha256=kI9yBm4uunsgV4-QeVoCBL0wLzU4kYmYv5qZilGnbPE,1740
 kodit/log.py,sha256=sHPHYetlMcKTor2VaFLMyao1_fZ_xhuzqXCAt5F5UMU,8575
-kodit/mcp.py,sha256=kiWyZ2Ptluh0jLDQNxKjxNmD18bEG1-zAFVOoZAPsWI,6192
+kodit/mcp.py,sha256=bUvG4by2CDN3QSkKNUer0yxpzAGbq-hg8HTU7S5fMv4,6217
 kodit/middleware.py,sha256=I6FOkqG9-8RH5kR1-0ZoQWfE4qLCB8lZYv8H_OCH29o,2714
 kodit/reporting.py,sha256=icce1ZyiADsA_Qz-mSjgn2H4SSqKuGfLKnw-yrl9nsg,2722
 kodit/application/__init__.py,sha256=mH50wTpgP9dhbKztFsL8Dda9Hi18TSnMVxXtpp4aGOA,35
 kodit/application/commands/__init__.py,sha256=AOVs25fwboBnMCWdgDB7fPbAYTljurAPVMkATIGRKuk,38
 kodit/application/commands/snippet_commands.py,sha256=WzRrnJOnLpIK8-wvN7c-ecGs_4LosQ_jR30dQkFqFBY,600
 kodit/application/services/__init__.py,sha256=p5UQNw-H5sxQvs5Etfte93B3cJ1kKW6DNxK34uFvU1E,38
-kodit/application/services/indexing_application_service.py,sha256=n7ViSfAPshUqe_gebgduRKwIPkB3Cl72CDw2U_qUJXY,13558
-kodit/application/services/snippet_application_service.py,sha256=NJxR3X78lTK0zXpYPdiCCw8NAJdLwyMyynfGnBfvONM,4759
+kodit/application/services/indexing_application_service.py,sha256=tfdEiFTVvqkZQ6I_ZkW5IECQJHZi35OJX4bj96lkZrc,14455
+kodit/application/services/snippet_application_service.py,sha256=2qW7ZIUQ63bdtAngf6xgQ09N-vygNpd12sXwyKwOcSs,5037
 kodit/domain/__init__.py,sha256=TCpg4Xx-oF4mKV91lo4iXqMEfBT1OoRSYnbG-zVWolA,66
 kodit/domain/entities.py,sha256=6XVuwDIQjkBw5Bm51io5ZUxB6_O4A774CDke2bfKWTY,5584
 kodit/domain/enums.py,sha256=Ik_h3D3eZ0FsSlPsU0ikm-Yv3Rmvzicffi9yBn19UIE,191
+kodit/domain/errors.py,sha256=yIsgCjM_yOFIg8l7l-t7jM8pgeAX4cfPq0owf7iz3DA,106
 kodit/domain/interfaces.py,sha256=Jkd0Ob4qSvhZHI9jRPFQ1n5Cv0SvU-y3Z-HCw2ikc4I,742
 kodit/domain/repositories.py,sha256=bdKxSKGI6XzrpzeKcv-NDV2JBirbEMRK-Y4UCZmDtoY,2706
 kodit/domain/value_objects.py,sha256=4Vs7Uk1wQgTjnCkZlOlw7E1Q8NiyAmBpFv38Lhs3WZ0,3869
@@ -28,7 +29,7 @@ kodit/domain/services/enrichment_service.py,sha256=XsXg3nV-KN4rqtC7Zro_ZiZ6RSq-1
 kodit/domain/services/ignore_service.py,sha256=boEN-IRLmUtwO9ZnuACaVFZbIKrtUG8YwnsXKEDIG28,1136
 kodit/domain/services/indexing_service.py,sha256=FEizu2GkvZA32xHOYXXch0LuHoWg6Z-BbJMPjZslzjc,5853
 kodit/domain/services/snippet_extraction_service.py,sha256=QW_99bXWpr8g6ZI-hp4Aj57VCSrUf71dLwQca5T6pyg,3065
-kodit/domain/services/source_service.py,sha256=CM9TKTvn-9xfsk3AoKvIoPQ5_MridXng2jY8skn6_5Q,2987
+kodit/domain/services/source_service.py,sha256=9XGS3imJn65v855cztsJSaaFod6LhkF2xfUVMaytx-A,3068
 kodit/infrastructure/__init__.py,sha256=HzEYIjoXnkz_i_MHO2e0sIVYweUcRnl2RpyBiTbMObU,28
 kodit/infrastructure/bm25/__init__.py,sha256=DmGbrEO34FOJy4e685BbyxLA7gPW1eqs2gAxsp6JOuM,34
 kodit/infrastructure/bm25/bm25_factory.py,sha256=I4eo7qRslnyXIRkBf-StZ5ga2Evrr5J5YFocTChFD3g,884
@@ -37,11 +38,11 @@ kodit/infrastructure/bm25/vectorchord_bm25_repository.py,sha256=0Db9XWFjiS4TFrsN
 kodit/infrastructure/cloning/__init__.py,sha256=IzIvX-yeRRFZ-lfvPVSEe_qXszO6DGQdjKwwDigexyQ,30
 kodit/infrastructure/cloning/metadata.py,sha256=C5LLmsUzi29RhSbzVDNqiShbekg7qdp1ihGUyFXy5yM,4277
 kodit/infrastructure/cloning/folder/__init__.py,sha256=w6ykrVtbYJlUDEXAjqgf6w2rMsUMCrrpIbl3QMjubgY,37
-kodit/infrastructure/cloning/folder/factory.py,sha256=qpFXxZuwYbdO17FKurWtm84ahwerDpcaLubClEZrbtk,3955
+kodit/infrastructure/cloning/folder/factory.py,sha256=vl1hwnYA7lczjotn2fahJQAt7IK96CSArx8cSaRFKeY,4242
 kodit/infrastructure/cloning/folder/working_copy.py,sha256=FPhwzuPj40yGoYvwcm9VG8mv8MbJxwfby_N5JS-_daA,1154
 kodit/infrastructure/cloning/git/__init__.py,sha256=20ePcp0qE6BuLsjsv4KYB1DzKhMIMsPXwEqIEZtjTJs,34
-kodit/infrastructure/cloning/git/factory.py,sha256=pjIvl3NIhXNNbRcI86wedz8CP-jV_Z3OVpGl4BUGCmo,4585
-kodit/infrastructure/cloning/git/working_copy.py,sha256=DMW_p7WWGoSeyDI9g55ItwsRomZSotXWRrlopqwszaQ,1115
+kodit/infrastructure/cloning/git/factory.py,sha256=cY0cxapp0NCvjMRpzesW_qRzbWbh-tMKIeAj0Eodyhw,5409
+kodit/infrastructure/cloning/git/working_copy.py,sha256=r_Uu6NYhRQLAQce6k4KThXLtGPqkzr6CgWx7AJ89gN4,1462
 kodit/infrastructure/embedding/__init__.py,sha256=F-8nLlWAerYJ0MOIA4tbXHLan8bW5rRR84vzxx6tRKI,39
 kodit/infrastructure/embedding/embedding_factory.py,sha256=1AypjhWJGxvLnZt1SEH_FHPk9P0Vkt9fXdSGzFPp2ow,3432
 kodit/infrastructure/embedding/local_vector_search_repository.py,sha256=UO8A3Eb_djFVrWKKSukAo4u7k8djDD1SlOPHk2pP9ps,3921
@@ -58,13 +59,13 @@ kodit/infrastructure/enrichment/local_enrichment_provider.py,sha256=8CATNtgMHgBR
 kodit/infrastructure/enrichment/null_enrichment_provider.py,sha256=5Ksyxl3qDLxUjmOeIdHZ0UAIULy7RcbLXJoT7_CNXoQ,775
 kodit/infrastructure/enrichment/openai_enrichment_provider.py,sha256=fenq4HiJ2UkrzsE2D0A0qpmro38z9mKaIzKKU5v7hnY,3189
 kodit/infrastructure/git/__init__.py,sha256=0iMosFzudj4_xNIMe2SRbV6l5bWqkjnUsZoFsoZFuM8,33
-kodit/infrastructure/git/git_utils.py,sha256=lOujEx41UuWfYSnFWbY4HC2tK5utytyzNkW1e5IPCr0,543
+kodit/infrastructure/git/git_utils.py,sha256=2DH6cyTjDRwFfL5Bzt1y2w0DwHZNypbC6R0Gm_A3hhg,2476
 kodit/infrastructure/ignore/__init__.py,sha256=VzFv8XOzHmsu0MEGnWVSF6KsgqLBmvHlRqAkT1Xb1MY,36
 kodit/infrastructure/ignore/ignore_pattern_provider.py,sha256=9m2XCsgW87UBTfzHr6Z0Ns6WpzwkLir3zyBY3PwsgXk,2225
 kodit/infrastructure/indexing/__init__.py,sha256=7UPRa2jwCAsa0Orsp6PqXSF8iIXJVzXHMFmrKkI9yH8,38
 kodit/infrastructure/indexing/fusion_service.py,sha256=mXUUcx3-8e75mWkxXMfl30HIoFXrTNHzB1w90MmEbak,1806
-kodit/infrastructure/indexing/index_repository.py,sha256=iHK9wcC8893Q6vY8OhW9UrN0gj1_GVzjTVlbewyZdxI,9094
-kodit/infrastructure/indexing/indexing_factory.py,sha256=BbavU9DJZapAe0V9H6Zsk7J6oPY48SJhGP3nrCPvzoQ,3914
+kodit/infrastructure/indexing/index_repository.py,sha256=4aSCBE_Gn9ihOx_kXOpUTTIv6_Q71-VRFHEBgpWaAEw,8906
+kodit/infrastructure/indexing/indexing_factory.py,sha256=KHA8c0XR9QrgqSR6gRUQk9wp6md97_oA1lwZFzoJAtk,3964
 kodit/infrastructure/snippet_extraction/__init__.py,sha256=v6KqrRDjSj0nt87m7UwRGx2GN_fz_14VWq9Q0uABR_s,54
 kodit/infrastructure/snippet_extraction/language_detection_service.py,sha256=Lo9xPLVia-70yP9gzyH4cQcBQzsp7WXjGOa5NBggScg,1158
 kodit/infrastructure/snippet_extraction/snippet_extraction_factory.py,sha256=LGbm614KCPNsM9K8r1z-E763NyAMIZA9ETJ_C61EknA,2759
@@ -76,10 +77,10 @@ kodit/infrastructure/snippet_extraction/languages/javascript.scm,sha256=Ini5TsVN
 kodit/infrastructure/snippet_extraction/languages/python.scm,sha256=ee85R9PBzwye3IMTE7-iVoKWd_ViU3EJISTyrFGrVeo,429
 kodit/infrastructure/snippet_extraction/languages/typescript.scm,sha256=U-ujbbv4tylbUBj9wuhL-e5cW6hmgPCNs4xrIX3r_hE,448
 kodit/infrastructure/sqlalchemy/__init__.py,sha256=UXPMSF_hgWaqr86cawRVqM8XdVNumQyyK5B8B97GnlA,33
-kodit/infrastructure/sqlalchemy/embedding_repository.py,sha256=vdjn3E5dFriFkceK8E8QMZzZk83etSf3NQOyS_LCeGY,7548
-kodit/infrastructure/sqlalchemy/file_repository.py,sha256=9O6ysCG_ldZFzDtpcxaTJQ8xya1A2q3CU2NafQo7GS0,2027
-kodit/infrastructure/sqlalchemy/repository.py,sha256=xWVAO9Bx9NKh3fSM-D1KGehT-6MNIYW2YQQcUcTE2xk,4369
-kodit/infrastructure/sqlalchemy/snippet_repository.py,sha256=zMqc5KxD0rP6r2BG2qLVAf3rh3IONkhZkI_pXg-r3a8,2137
+kodit/infrastructure/sqlalchemy/embedding_repository.py,sha256=dCCRV5rD8T5xBksPKvdr-z4F72WKo4-dH9mXazDByXQ,7476
+kodit/infrastructure/sqlalchemy/file_repository.py,sha256=9_kXHJ1YiWA1ingpvBNq8cuxkMu59PHwl_m9_Ttnq2o,2353
+kodit/infrastructure/sqlalchemy/repository.py,sha256=EpZnOjR3wfPEqIauWw_KczpkSqBQPTq5sIyCpJCuW2w,4565
+kodit/infrastructure/sqlalchemy/snippet_repository.py,sha256=RYhBnyvvBo-6obrUkds6BhEjchs4HYQL8k9x0Cy7BtM,2430
 kodit/infrastructure/ui/__init__.py,sha256=CzbLOBwIZ6B6iAHEd1L8cIBydCj-n_kobxJAhz2I9_Y,32
 kodit/infrastructure/ui/progress.py,sha256=BaAeMEgXlSSb0c_t_NPxnThIktkzzCS9kegb5ExULJs,4791
 kodit/infrastructure/ui/spinner.py,sha256=GcP115qtR0VEnGfMEtsGoAUpRzVGUSfiUXfoJJERngA,2357
@@ -92,8 +93,8 @@ kodit/migrations/versions/85155663351e_initial.py,sha256=Cg7zlF871o9ShV5rQMQ1v7h
 kodit/migrations/versions/9e53ea8bb3b0_add_authors.py,sha256=a32Zm8KUQyiiLkjKNPYdaJDgjW6VsV-GhaLnPnK_fpI,3884
 kodit/migrations/versions/__init__.py,sha256=9-lHzptItTzq_fomdIRBegQNm4Znx6pVjwD4MiqRIdo,36
 kodit/migrations/versions/c3f5137d30f5_index_all_the_things.py,sha256=rI8LmjF-I2OMxZ2nOIF_NRmqOLXe45hL_iz_nx97DTQ,1680
-kodit-0.2.5.dist-info/METADATA,sha256=UuPIpWyvjccCITCITmMcJhUwIxRd211UvDmVXi8MbY4,5867
-kodit-0.2.5.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-kodit-0.2.5.dist-info/entry_points.txt,sha256=hoTn-1aKyTItjnY91fnO-rV5uaWQLQ-Vi7V5et2IbHY,40
-kodit-0.2.5.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-kodit-0.2.5.dist-info/RECORD,,
+kodit-0.2.7.dist-info/METADATA,sha256=G5rdRgHtm6V-p1tt0h7g7pZJKnQ5mAnCgDXVFRJG7Fg,5867
+kodit-0.2.7.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+kodit-0.2.7.dist-info/entry_points.txt,sha256=hoTn-1aKyTItjnY91fnO-rV5uaWQLQ-Vi7V5et2IbHY,40
+kodit-0.2.7.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+kodit-0.2.7.dist-info/RECORD,,