natural-pdf 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

natural_pdf/search/searchable_mixin.py

@@ -1,34 +1,48 @@
+import hashlib  # For hashing content
 import logging
-from typing import Optional, List, Dict, Any, Iterable, TYPE_CHECKING, Union, Type, Generator
 from abc import ABC, abstractmethod
-import hashlib # For hashing content
+from typing import TYPE_CHECKING, Any, Dict, Generator, Iterable, List, Optional, Type, Union
 
 # Now import the flag from the canonical source - this import should always work
 from .haystack_utils import HAS_HAYSTACK_EXTRAS
+
 DEFAULT_SEARCH_COLLECTION_NAME = "default_collection"
 
 # Avoid runtime import errors if extras not installed
 try:
     # Import protocols and options first
+    from . import get_search_service
+    from .search_options import SearchOptions, TextSearchOptions
     from .search_service_protocol import (
-        SearchServiceProtocol, Indexable, 
-        IndexConfigurationError, IndexExistsError
+        Indexable,
+        IndexConfigurationError,
+        IndexExistsError,
+        SearchServiceProtocol,
     )
-    from .search_options import SearchOptions, TextSearchOptions
-    from . import get_search_service
-    
-    if TYPE_CHECKING: # Keep type hints working
-        from natural_pdf.elements.region import Region # Example indexable type
+
+    if TYPE_CHECKING:  # Keep type hints working
+        from natural_pdf.elements.region import Region  # Example indexable type
 except ImportError:
     # Define dummies if extras missing
-    SearchServiceProtocol, Indexable, IndexConfigurationError, IndexExistsError = object, object, RuntimeError, RuntimeError
+    SearchServiceProtocol, Indexable, IndexConfigurationError, IndexExistsError = (
+        object,
+        object,
+        RuntimeError,
+        RuntimeError,
+    )
     SearchOptions, TextSearchOptions = object, object
     DEFAULT_SEARCH_COLLECTION_NAME = "default_collection"
-    def get_search_service(**kwargs): raise ImportError("Search dependencies missing.")
-    class Region: pass # Dummy for type hint
+
+    def get_search_service(**kwargs):
+        raise ImportError("Search dependencies missing.")
+
+    class Region:
+        pass  # Dummy for type hint
+
 
 logger = logging.getLogger(__name__)
 
+
 class SearchableMixin(ABC):
     """
     Mixin class providing search functionality (initialization, indexing, searching, syncing).
@@ -36,6 +50,7 @@ class SearchableMixin(ABC):
     Requires the inheriting class to implement `get_indexable_items`.
     Assumes the inheriting class has a `_search_service` attribute initialized to None.
     """
+
     # Ensure inheriting class initializes this
     _search_service: Optional[SearchServiceProtocol] = None
 
@@ -53,12 +68,12 @@ class SearchableMixin(ABC):
         *,
         persist: Optional[bool] = None,
         collection_name: Optional[str] = None,
-        embedding_model: Optional[str] = None, # Allow overriding embedding model
-        index: bool = False, # Changed from index_now
+        embedding_model: Optional[str] = None,  # Allow overriding embedding model
+        index: bool = False,  # Changed from index_now
         force_reindex: bool = False,
         embedder_device: Optional[str] = None,
-        **kwargs # Pass other args to get_search_service
-    ) -> 'SearchableMixin': # Return self for chaining
+        **kwargs,  # Pass other args to get_search_service
+    ) -> "SearchableMixin":  # Return self for chaining
         """
         Initializes and configures the search service for this instance.
 
@@ -90,7 +105,9 @@ class SearchableMixin(ABC):
         """
         if service:
             # Attach provided service
-            logger.info(f"Attaching provided SearchService instance (Collection: '{getattr(service, 'collection_name', '<Unknown>')}').")
+            logger.info(
+                f"Attaching provided SearchService instance (Collection: '{getattr(service, 'collection_name', '<Unknown>')}')."
+            )
             # TODO: Add stricter type check? isinstance(service, SearchServiceProtocol) requires runtime_checkable
             self._search_service = service
         else:
@@ -101,24 +118,45 @@ class SearchableMixin(ABC):
                 raise ValueError("A collection_name must be provided when persist=True.")
             elif not effective_persist and not effective_collection_name:
                 effective_collection_name = DEFAULT_SEARCH_COLLECTION_NAME
-                logger.info(f"Using default collection name '{DEFAULT_SEARCH_COLLECTION_NAME}' for in-memory service.")
+                logger.info(
+                    f"Using default collection name '{DEFAULT_SEARCH_COLLECTION_NAME}' for in-memory service."
+                )
 
-            logger.info(f"Creating new SearchService: name='{effective_collection_name}', persist={effective_persist}, model={embedding_model or 'default'}")
+            logger.info(
+                f"Creating new SearchService: name='{effective_collection_name}', persist={effective_persist}, model={embedding_model or 'default'}"
+            )
             try:
-                service_args = {"collection_name": effective_collection_name, "persist": effective_persist, **kwargs}
-                if embedding_model: service_args['embedding_model'] = embedding_model
+                service_args = {
+                    "collection_name": effective_collection_name,
+                    "persist": effective_persist,
+                    **kwargs,
+                }
+                if embedding_model:
+                    service_args["embedding_model"] = embedding_model
                 self._search_service = get_search_service(**service_args)
+            except ImportError as ie:  # Catch the specific ImportError first
+                logger.error(f"Failed to create SearchService due to missing dependency: {ie}")
+                raise ie  # Re-raise the original ImportError
             except Exception as e:
-                logger.error(f"Failed to create SearchService: {e}", exc_info=True)
-                raise RuntimeError("Could not create SearchService instance.") from e
-
-        # --- Optional Immediate Indexing (with safety check for persistent) --- 
+                logger.error(
+                    f"Failed to create SearchService due to unexpected error: {e}", exc_info=True
+                )
+                # Keep the RuntimeError for other unexpected creation errors
+                raise RuntimeError(
+                    "Could not create SearchService instance due to an unexpected error."
+                ) from e
+
+        # --- Optional Immediate Indexing (with safety check for persistent) ---
         if index:
-            if not self._search_service: # Should not happen if logic above is correct
-                raise RuntimeError("Cannot index: Search service not available after initialization attempt.")
+            if not self._search_service:  # Should not happen if logic above is correct
+                raise RuntimeError(
+                    "Cannot index: Search service not available after initialization attempt."
+                )
 
-            is_persistent = getattr(self._search_service, '_persist', False) # Check if service is persistent
-            collection_name = getattr(self._search_service, 'collection_name', '<Unknown>')
+            is_persistent = getattr(
+                self._search_service, "_persist", False
+            )  # Check if service is persistent
+            collection_name = getattr(self._search_service, "collection_name", "<Unknown>")
 
             if is_persistent and not force_reindex:
                 # Check existence only if persistent and not forcing reindex
@@ -131,14 +169,20 @@ class SearchableMixin(ABC):
                     )
                 else:
                     # Index doesn't exist, safe to proceed
-                    logger.info(f"Persistent index '{collection_name}' does not exist. Proceeding with initial indexing.")
+                    logger.info(
+                        f"Persistent index '{collection_name}' does not exist. Proceeding with initial indexing."
+                    )
             elif is_persistent and force_reindex:
-                logger.warning(f"Proceeding with index=True and force_reindex=True for persistent index '{collection_name}'. Existing data will be deleted.")
+                logger.warning(
+                    f"Proceeding with index=True and force_reindex=True for persistent index '{collection_name}'. Existing data will be deleted."
+                )
             # else: # Not persistent, safe to proceed without existence check
             #     logger.debug("Proceeding with index=True for non-persistent index.")
 
             # Proceed with indexing if checks passed or not applicable
-            logger.info(f"index=True: Proceeding to index collection immediately after search initialization.")
+            logger.info(
+                f"index=True: Proceeding to index collection immediately after search initialization."
+            )
             self._perform_indexing(force_reindex=force_reindex, embedder_device=embedder_device)
 
         return self
@@ -148,42 +192,53 @@ class SearchableMixin(ABC):
         if not self._search_service:
             raise RuntimeError("Search service not initialized. Call init_search first.")
 
-        collection_name = getattr(self._search_service, 'collection_name', '<Unknown>')
-        logger.info(f"Starting internal indexing process into SearchService collection '{collection_name}'...")
+        collection_name = getattr(self._search_service, "collection_name", "<Unknown>")
+        logger.info(
+            f"Starting internal indexing process into SearchService collection '{collection_name}'..."
+        )
 
         # Use the abstract method to get items
         try:
-            indexable_items = list(self.get_indexable_items()) # Consume iterator
+            indexable_items = list(self.get_indexable_items())  # Consume iterator
         except Exception as e:
-             logger.error(f"Error calling get_indexable_items: {e}", exc_info=True)
-             raise RuntimeError("Failed to retrieve indexable items for indexing.") from e
+            logger.error(f"Error calling get_indexable_items: {e}", exc_info=True)
+            raise RuntimeError("Failed to retrieve indexable items for indexing.") from e
 
         if not indexable_items:
-            logger.warning("No indexable items provided by get_indexable_items(). Skipping index call.")
+            logger.warning(
+                "No indexable items provided by get_indexable_items(). Skipping index call."
+            )
             return
 
         logger.info(f"Prepared {len(indexable_items)} indexable items for indexing.")
         try:
-             logger.debug(f"Calling index() on SearchService for collection '{collection_name}' (force_reindex={force_reindex}).")
-             self._search_service.index(
-                 documents=indexable_items,
-                 embedder_device=embedder_device,
-                 force_reindex=force_reindex,
-             )
-             logger.info(f"Successfully completed indexing into SearchService collection '{collection_name}'.")
+            logger.debug(
+                f"Calling index() on SearchService for collection '{collection_name}' (force_reindex={force_reindex})."
+            )
+            self._search_service.index(
+                documents=indexable_items,
+                embedder_device=embedder_device,
+                force_reindex=force_reindex,
+            )
+            logger.info(
+                f"Successfully completed indexing into SearchService collection '{collection_name}'."
+            )
         except IndexConfigurationError as ice:
-             logger.error(f"Indexing failed due to configuration error in collection '{collection_name}': {ice}", exc_info=True)
-             raise # Re-raise specific error
-        except Exception as e: # Catch other indexing errors from the service
-             logger.error(f"Indexing failed for collection '{collection_name}': {e}", exc_info=True)
-             raise RuntimeError(f"Indexing failed for collection '{collection_name}'.") from e
-
+            logger.error(
+                f"Indexing failed due to configuration error in collection '{collection_name}': {ice}",
+                exc_info=True,
+            )
+            raise  # Re-raise specific error
+        except Exception as e:  # Catch other indexing errors from the service
+            logger.error(f"Indexing failed for collection '{collection_name}': {e}", exc_info=True)
+            raise RuntimeError(f"Indexing failed for collection '{collection_name}'.") from e
 
     def index_for_search(
-        self, *, # Make args keyword-only
+        self,
+        *,  # Make args keyword-only
         embedder_device: Optional[str] = None,
         force_reindex: bool = False,
-    ) -> 'SearchableMixin':
+    ) -> "SearchableMixin":
         """
         Ensures the search service is initialized (using default if needed)
         and indexes the items provided by `get_indexable_items`.
@@ -201,8 +256,10 @@ class SearchableMixin(ABC):
         """
         # --- Ensure Service is Initialized (Use Default if Needed) ---
         if not self._search_service:
-            logger.info("Search service not initialized prior to index_for_search. Initializing default in-memory service.")
-            self.init_search() # Call init with defaults
+            logger.info(
+                "Search service not initialized prior to index_for_search. Initializing default in-memory service."
+            )
+            self.init_search()  # Call init with defaults
 
         # --- Perform Indexing ---
         self._perform_indexing(force_reindex=force_reindex, embedder_device=embedder_device)
@@ -210,10 +267,10 @@ class SearchableMixin(ABC):
 
     def find_relevant(
         self,
-        query: Any, # Query type depends on service capabilities
-        *, # Make options/service keyword-only
+        query: Any,  # Query type depends on service capabilities
+        *,  # Make options/service keyword-only
         options: Optional[SearchOptions] = None,
-        search_service: Optional[SearchServiceProtocol] = None # Allow override
+        search_service: Optional[SearchServiceProtocol] = None,  # Allow override
     ) -> List[Dict[str, Any]]:
         """
         Finds relevant items using the configured or provided search service.
@@ -240,8 +297,10 @@ class SearchableMixin(ABC):
                 "or provide an explicit 'search_service' instance to find_relevant()."
             )
 
-        collection_name = getattr(effective_service, 'collection_name', '<Unknown>')
-        logger.info(f"Searching collection '{collection_name}' via {type(effective_service).__name__}...")
+        collection_name = getattr(effective_service, "collection_name", "<Unknown>")
+        logger.info(
+            f"Searching collection '{collection_name}' via {type(effective_service).__name__}..."
+        )
 
         # --- Prepare Query and Options ---
         query_input = query
@@ -263,11 +322,15 @@ class SearchableMixin(ABC):
                 query=query_input,
                 options=effective_options,
             )
-            logger.info(f"SearchService returned {len(results)} results from collection '{collection_name}'.")
+            logger.info(
+                f"SearchService returned {len(results)} results from collection '{collection_name}'."
+            )
             return results
         except FileNotFoundError as fnf:
-            logger.error(f"Search failed: Collection '{collection_name}' not found by service. Error: {fnf}")
-            raise # Re-raise specific error
+            logger.error(
+                f"Search failed: Collection '{collection_name}' not found by service. Error: {fnf}"
+            )
+            raise  # Re-raise specific error
         except Exception as e:
             logger.error(f"Search failed for collection '{collection_name}': {e}", exc_info=True)
             # Consider wrapping in a SearchError?
@@ -276,11 +339,11 @@ class SearchableMixin(ABC):
     # --- NEW Sync Method ---
     def sync_index(
         self,
-        strategy: str = 'full', # 'full' (add/update/delete) or 'upsert_only'
+        strategy: str = "full",  # 'full' (add/update/delete) or 'upsert_only'
         dry_run: bool = False,
-        batch_size: int = 100, # For batching deletes/updates if needed
-        embedder_device: Optional[str] = None, # Pass embedder device if needed for updates
-        **kwargs: Any # Allow passing extra args to get_search_service
+        batch_size: int = 100,  # For batching deletes/updates if needed
+        embedder_device: Optional[str] = None,  # Pass embedder device if needed for updates
+        **kwargs: Any,  # Allow passing extra args to get_search_service
     ) -> Dict[str, int]:
         """
         Synchronizes the search index with the current state of indexable items.
@@ -307,91 +370,107 @@ class SearchableMixin(ABC):
             RuntimeError: For backend errors during synchronization.
         """
         if not self._search_service:
-             raise RuntimeError("Search service not configured. Call init_search first.")
+            raise RuntimeError("Search service not configured. Call init_search first.")
 
-        collection_name = getattr(self._search_service, 'collection_name', '<Unknown>')
-        logger.info(f"Starting index synchronization for collection '{collection_name}' (Strategy: {strategy}, Dry run: {dry_run})...")
-        summary = {'added': 0, 'updated': 0, 'deleted': 0, 'skipped': 0}
+        collection_name = getattr(self._search_service, "collection_name", "<Unknown>")
+        logger.info(
+            f"Starting index synchronization for collection '{collection_name}' (Strategy: {strategy}, Dry run: {dry_run})..."
+        )
+        summary = {"added": 0, "updated": 0, "deleted": 0, "skipped": 0}
 
         # --- Check Service Capabilities for 'full' sync ---
-        if strategy == 'full':
-             required_methods = ['list_documents', 'delete_documents']
-             missing_methods = [m for m in required_methods if not hasattr(self._search_service, m)]
-             if missing_methods:
-                 raise NotImplementedError(
-                     f"The configured search service ({type(self._search_service).__name__}) "
-                     f"is missing required methods for 'full' sync strategy: {', '.join(missing_methods)}"
-                 )
+        if strategy == "full":
+            required_methods = ["list_documents", "delete_documents"]
+            missing_methods = [m for m in required_methods if not hasattr(self._search_service, m)]
+            if missing_methods:
+                raise NotImplementedError(
+                    f"The configured search service ({type(self._search_service).__name__}) "
+                    f"is missing required methods for 'full' sync strategy: {', '.join(missing_methods)}"
+                )
 
         # --- 1. Get Desired State (from current collection) ---
-        desired_state: Dict[str, Indexable] = {} # {id: item}
-        desired_hashes: Dict[str, Optional[str]] = {} # {id: hash or None}
+        desired_state: Dict[str, Indexable] = {}  # {id: item}
+        desired_hashes: Dict[str, Optional[str]] = {}  # {id: hash or None}
         try:
-             for item in self.get_indexable_items():
-                 item_id = item.get_id()
-                 if not item_id:
-                     logger.warning(f"Skipping item with no ID: {item}")
-                     summary['skipped'] += 1
-                     continue
-                 if item_id in desired_state:
-                      logger.warning(f"Duplicate ID '{item_id}' found in get_indexable_items(). Skipping subsequent item.")
-                      summary['skipped'] += 1
-                      continue
-                 desired_state[item_id] = item
-                 # Try to get hash, store None if unavailable or fails
-                 try:
-                      desired_hashes[item_id] = item.get_content_hash()
-                 except (AttributeError, NotImplementedError):
-                      logger.debug(f"get_content_hash not available for item ID '{item_id}' ({type(item).__name__}). Sync update check will be ID-based.")
-                      desired_hashes[item_id] = None
-                 except Exception as e:
-                      logger.warning(f"Error getting content hash for item ID '{item_id}': {e}. Sync update check will be ID-based.", exc_info=False)
-                      desired_hashes[item_id] = None
+            for item in self.get_indexable_items():
+                item_id = item.get_id()
+                if not item_id:
+                    logger.warning(f"Skipping item with no ID: {item}")
+                    summary["skipped"] += 1
+                    continue
+                if item_id in desired_state:
+                    logger.warning(
+                        f"Duplicate ID '{item_id}' found in get_indexable_items(). Skipping subsequent item."
+                    )
+                    summary["skipped"] += 1
+                    continue
+                desired_state[item_id] = item
+                # Try to get hash, store None if unavailable or fails
+                try:
+                    desired_hashes[item_id] = item.get_content_hash()
+                except (AttributeError, NotImplementedError):
+                    logger.debug(
+                        f"get_content_hash not available for item ID '{item_id}' ({type(item).__name__}). Sync update check will be ID-based."
+                    )
+                    desired_hashes[item_id] = None
+                except Exception as e:
+                    logger.warning(
+                        f"Error getting content hash for item ID '{item_id}': {e}. Sync update check will be ID-based.",
+                        exc_info=False,
+                    )
+                    desired_hashes[item_id] = None
 
         except Exception as e:
-             logger.error(f"Error iterating through get_indexable_items: {e}", exc_info=True)
-             raise RuntimeError("Failed to get current indexable items.") from e
+            logger.error(f"Error iterating through get_indexable_items: {e}", exc_info=True)
+            raise RuntimeError("Failed to get current indexable items.") from e
 
         logger.info(f"Desired state contains {len(desired_state)} indexable items.")
 
         # --- 2. Handle Different Strategies ---
-        if strategy == 'upsert_only':
+        if strategy == "upsert_only":
             # Simple case: just index everything, let the service handle upserts
             items_to_index = list(desired_state.values())
-            summary['added'] = len(items_to_index) # Approximate count
-            logger.info(f"Strategy 'upsert_only': Prepared {len(items_to_index)} items for indexing/upserting.")
+            summary["added"] = len(items_to_index)  # Approximate count
+            logger.info(
+                f"Strategy 'upsert_only': Prepared {len(items_to_index)} items for indexing/upserting."
+            )
             if not dry_run and items_to_index:
-                 logger.debug("Calling service.index for upsert...")
-                 # Call index directly, force_reindex=False implies upsert
-                 self._search_service.index(
-                     documents=items_to_index,
-                     force_reindex=False,
-                     embedder_device=embedder_device
-                 )
+                logger.debug("Calling service.index for upsert...")
+                # Call index directly, force_reindex=False implies upsert
+                self._search_service.index(
+                    documents=items_to_index, force_reindex=False, embedder_device=embedder_device
+                )
             elif dry_run:
                 logger.info("[Dry Run] Would index/upsert %d items.", len(items_to_index))
 
-        elif strategy == 'full':
+        elif strategy == "full":
             # Complex case: Add/Update/Delete
             # 2a. Get Current Index State
             try:
-                 logger.debug("Listing documents currently in the index...")
-                 # Assumes list_documents takes filters and include_metadata
-                 # Fetch all documents with metadata
-                 current_docs = self._search_service.list_documents(include_metadata=True)
-                 current_state: Dict[str, Dict] = {} # {id: {'meta': {...}, ...}}
-                 duplicates = 0
-                 for doc in current_docs:
-                     doc_id = doc.get('id')
-                     if not doc_id: continue # Skip docs without ID from service
-                     if doc_id in current_state: duplicates +=1
-                     current_state[doc_id] = doc
-                 logger.info(f"Found {len(current_state)} documents currently in the index (encountered {duplicates} duplicate IDs).")
-                 if duplicates > 0: logger.warning(f"Found {duplicates} duplicate IDs in the index. Using the last encountered version for comparison.")
+                logger.debug("Listing documents currently in the index...")
+                # Assumes list_documents takes filters and include_metadata
+                # Fetch all documents with metadata
+                current_docs = self._search_service.list_documents(include_metadata=True)
+                current_state: Dict[str, Dict] = {}  # {id: {'meta': {...}, ...}}
+                duplicates = 0
+                for doc in current_docs:
+                    doc_id = doc.get("id")
+                    if not doc_id:
+                        continue  # Skip docs without ID from service
+                    if doc_id in current_state:
+                        duplicates += 1
+                    current_state[doc_id] = doc
+                logger.info(
+                    f"Found {len(current_state)} documents currently in the index (encountered {duplicates} duplicate IDs)."
+                )
+                if duplicates > 0:
+                    logger.warning(
+                        f"Found {duplicates} duplicate IDs in the index. Using the last encountered version for comparison."
+                    )
 
             except Exception as e:
-                 logger.error(f"Failed to list documents from search service: {e}", exc_info=True)
-                 raise RuntimeError("Could not retrieve current index state for sync.") from e
+                logger.error(f"Failed to list documents from search service: {e}", exc_info=True)
+                raise RuntimeError("Could not retrieve current index state for sync.") from e
 
             # 2b. Compare States and Plan Actions
             ids_in_desired = set(desired_state.keys())
@@ -403,62 +482,68 @@ class SearchableMixin(ABC):
 
             items_to_update = []
             for item_id in ids_to_check_update:
-                 desired_hash = desired_hashes.get(item_id)
-                 current_meta = current_state[item_id].get('meta', {})
-                 current_hash = current_meta.get('content_hash') # Assuming hash stored in meta
-
-                 # Check if hash exists and differs, or if hash is missing (force update)
-                 if desired_hash is None or current_hash is None or desired_hash != current_hash:
-                      if desired_hash != current_hash:
-                          logger.debug(f"Content hash changed for ID {item_id}. Scheduling for update.")
-                      else:
-                          logger.debug(f"Hash missing for ID {item_id}. Scheduling for update.")
-                      items_to_update.append(desired_state[item_id])
-                 # Else: hashes match, no update needed
+                desired_hash = desired_hashes.get(item_id)
+                current_meta = current_state[item_id].get("meta", {})
+                current_hash = current_meta.get("content_hash")  # Assuming hash stored in meta
+
+                # Check if hash exists and differs, or if hash is missing (force update)
+                if desired_hash is None or current_hash is None or desired_hash != current_hash:
+                    if desired_hash != current_hash:
+                        logger.debug(
+                            f"Content hash changed for ID {item_id}. Scheduling for update."
+                        )
+                    else:
+                        logger.debug(f"Hash missing for ID {item_id}. Scheduling for update.")
+                    items_to_update.append(desired_state[item_id])
+                # Else: hashes match, no update needed
 
             items_to_add = [desired_state[id_] for id_ in ids_to_add]
-            items_to_index = items_to_add + items_to_update # Combine adds and updates for single index call
+            items_to_index = (
+                items_to_add + items_to_update
+            )  # Combine adds and updates for single index call
 
-            summary['added'] = len(items_to_add)
-            summary['updated'] = len(items_to_update)
-            summary['deleted'] = len(ids_to_delete)
+            summary["added"] = len(items_to_add)
+            summary["updated"] = len(items_to_update)
+            summary["deleted"] = len(ids_to_delete)
 
-            logger.info(f"Sync Plan: Add={summary['added']}, Update={summary['updated']}, Delete={summary['deleted']}")
+            logger.info(
+                f"Sync Plan: Add={summary['added']}, Update={summary['updated']}, Delete={summary['deleted']}"
+            )
 
             # 2c. Execute Actions (if not dry_run)
             if not dry_run:
-                 # Execute Deletes
-                 if ids_to_delete:
-                      logger.info(f"Deleting {len(ids_to_delete)} items from index...")
-                      try:
-                           # Assuming delete_documents takes list of IDs
-                           # Implement batching if needed
-                           self._search_service.delete_documents(ids=list(ids_to_delete))
-                           logger.info("Deletion successful.")
-                      except Exception as e:
-                           logger.error(f"Failed to delete documents: {e}", exc_info=True)
-                           # Decide whether to continue or raise
-                           raise RuntimeError("Failed during deletion phase of sync.") from e
-
-                 # Execute Adds/Updates
-                 if items_to_index:
-                      logger.info(f"Indexing/Updating {len(items_to_index)} items...")
-                      try:
-                           # Upsert logic handled by service's index method with force_reindex=False
-                           self._search_service.index(
-                               documents=items_to_index,
-                               force_reindex=False,
-                               embedder_device=embedder_device
-                           )
-                           logger.info("Add/Update successful.")
-                      except Exception as e:
-                           logger.error(f"Failed to index/update documents: {e}", exc_info=True)
-                           raise RuntimeError("Failed during add/update phase of sync.") from e
-                 logger.info("Sync actions completed.")
+                # Execute Deletes
+                if ids_to_delete:
+                    logger.info(f"Deleting {len(ids_to_delete)} items from index...")
+                    try:
+                        # Assuming delete_documents takes list of IDs
+                        # Implement batching if needed
+                        self._search_service.delete_documents(ids=list(ids_to_delete))
+                        logger.info("Deletion successful.")
+                    except Exception as e:
+                        logger.error(f"Failed to delete documents: {e}", exc_info=True)
+                        # Decide whether to continue or raise
+                        raise RuntimeError("Failed during deletion phase of sync.") from e
+
+                # Execute Adds/Updates
+                if items_to_index:
+                    logger.info(f"Indexing/Updating {len(items_to_index)} items...")
+                    try:
+                        # Upsert logic handled by service's index method with force_reindex=False
+                        self._search_service.index(
+                            documents=items_to_index,
+                            force_reindex=False,
+                            embedder_device=embedder_device,
+                        )
+                        logger.info("Add/Update successful.")
+                    except Exception as e:
+                        logger.error(f"Failed to index/update documents: {e}", exc_info=True)
+                        raise RuntimeError("Failed during add/update phase of sync.") from e
+                logger.info("Sync actions completed.")
             else:
-                 logger.info("[Dry Run] No changes applied to the index.")
+                logger.info("[Dry Run] No changes applied to the index.")
 
         else:
             raise ValueError(f"Unknown sync strategy: '{strategy}'. Use 'full' or 'upsert_only'.")
 
-        return summary 
+        return summary

natural_pdf/selectors/__init__.py

@@ -1,4 +1,5 @@
 """
 Selector module for natural-pdf.
 """
-from natural_pdf.selectors.parser import parse_selector, selector_to_filter_func
+
+from natural_pdf.selectors.parser import parse_selector, selector_to_filter_func