Find files that have changed since last indexing.

Fit BM25 model on corpus.

Get all entity names from vector store for global entity awareness.

Get all relations from the collection. Returns: List of relation points from the collection.

Generate cache key for text.

Get the tag name of an HTML element.

Get files that need processing for incremental indexing.

Handle deleted files by removing their entities and orphaned relations.

Parse timestamp from various formats.

Update the last cleanup timestamp in state file.

Upsert points in batches.

Claude Code Memory Indexer - Universal semantic indexing for codebases.

Combine this result with another result.

Create a 'calls' relationship.

Override to handle relation chunk creation.

Delete a collection.

Generate deterministic embeddings based on text hash.

Generate embeddings for multiple texts.

Create a failed processing result.

Find entities for file grouped by chunk type for targeted replacement. Args: collection_name: Name of the collection to search

Create a RelationChunk from a Relation.

Get information about a collection.

Get overall database statistics.

Get usage statistics.

Process entity metadata with Git+Meta deduplication.

Generate embeddings with error handling and cost tracking.

Register an embedder class.

Run indexing with specific file paths, bypassing file discovery. This function accepts specific files to process, eliminating the expensive f

Insert or update points in the collection.

Add result to cache with size management.

Atomically write JSON data to a file using temp file + rename pattern. Args: file_path: Target file path data: Dictio

Build Qdrant filter from conditions.

Check and enforce rate limits.

Create a callback function for batch processing during streaming.

Create a meaningful name for a content entity.

Create a content-only result without structural parsing.

Embed a single batch of texts.

Embed a single batch of texts.

Extract method names from class body.

Extract selectors from a CSS rule (same logic as CSSParser).

Extract docstring/JSDoc from function or class node with deep AST traversal.

Extract variables from array destructuring patterns. Handles: [first, second], [head, ...tail]

Extract selectors from a CSS rule.

Extract string argument from a function call.

Extract individual variable names from different pattern types. Handles: identifier, object_pattern, array_pattern, rest_pattern.

Format entity for enhanced BM25 searchability with 6-component structure.

Format memory entries for HTML display.

Generate sparse vector for a single text using proper IDF-based term weighting.

Extract class name from class declaration.

Get entity content for hashing

Get consistent entity ID for comparison

Determine comprehensive health status based on collection statistics.

Get QdrantStore instance

Get set of files that currently have entities in the vector database.

Print individual collection statistics.

Process a batch of files with progressive disclosure support. Returns: Tuple of (entities, relations, implementation_chunks, erro

Process a batch of file changes with phantom deletion detection.

Check if Qdrant is available.

Internal scraping logic

Check if orphan cleanup should run based on timer interval.

Determine if a file should use batch processing during parsing.

Start watcher for a single project with project-specific config.

Remove HTML tags, JavaScript, and CSS from text while preserving readable content.

Comprehensive code analysis in a single pass. Args: code_info: Formatted code info string Returns:

Summarize multiple conversations with rate limiting.

Check if content hash already exists in storage

Universal deduplication logic using content hashes.

Clear collection data. Args: collection_name: Name of the collection preserve_manual: If True (default), preserve man

Claude Code Memory Indexer - Universal semantic indexing for codebases.

Create a vector point from an EntityChunk for progressive disclosure.

Create a class entity with enhanced observations.

Create a new collection.

Create default project configuration.

Create a default settings.txt file template.

Create a function entity with enhanced observations.

Delete a collection.

Delete points by their IDs.

Delete points by their IDs.

Generate embedding for single text - interface compatibility.

Extract deletion information from verbose CLI output.

Extract observations for function entities with Jedi enrichment.

Extract processed filenames from verbose CLI output.

Format a chat conversation as a markdown entry.

Generate deterministic ID from content.

Generate HTML report for a conversation. Args: conversation_input: Conversation ID, chat file path, or ChatConversation object

Get Qdrant client information.

Define code-indexed entity types to exclude.

Get information about a collection.

Generate hash of content for change detection

Get the default log file path, optionally per collection and project.

Get file exclusion patterns.

Get file inclusion patterns.

Define manual entity types based on common patterns.

Get maximum token limit for input text.

Get information about the embedding model.

Get all supported file extensions.

Check if the project is a git repository.

Check if Memory Guard is disabled globally.

Check if a collection name is a production collection that should never be deleted.

List all collections.