Check if project config exists.

Count total points in collection - test compatibility method.

Load configuration from multiple sources with precedence. Maintains backward compatibility with old signature. Args: settings_file:

Extract text from tree-sitter node.

Get the global logger instance.

Legacy search interface for test compatibility.

Index an entire project with automatic incremental detection.

Load unified configuration from all sources. Precedence (highest to lowest): 1. Explicit overrides 2. Environment variables

Legacy method for backward compatibility.

Recursively find all nodes matching given types.

Load previous indexing state.

Format a change value with +/- indicator.

Create an 'imports' relationship.

Create deterministic chunk ID following existing pattern.

Get collection-specific state file with simple naming.

Create a 'contains' relationship.

Update state file by merging new files with existing state, or do full rebuild.

Save debug information to last updated log file (keeps 3 files).

Register a new parser.

Find all files matching project patterns.

Get current state of files.

Get SHA256 hash of file contents.

Calculate SHA256 hash of file contents.

Create embedder from configuration (IndexerConfig or dict).

Start file watching using IndexingEventHandler.

Create a successful processing result.

Clean up relations that reference non-existent entities. Uses a single atomic query to get a consistent snapshot of the database, avo

Create file entity using EntityFactory pattern.

Extract individual items from a content array.

Unified scroll method for retrieving points from a collection. Args: collection_name: Name of the collection to scroll

Enhanced scope filtering - exclude function-local and block-scoped variables. Only include module-level variables (like Python parser

Check if this is a Markdown file.

Check if collection exists.

Load configuration from legacy settings.txt format.

Accurate token count using tiktoken with fallback.

Get the value of an attribute from an HTML element.

Check if this is a Python file.

Create vector store from configuration (IndexerConfig or dict).

Helper function to check if there are no errors in the log output.

Stop file watching and cleanup.

Verify that a specific entity is indexed and searchable. Args: qdrant_store: QdrantStore instance dummy_embedder: Embedder f

Helper to create variable entity with consistent structure.

Parse Python file using Tree-sitter and Jedi.

Wait for Qdrant eventual consistency by retrying searches until entities are properly deleted. Args: search_func: Function that retu

Accurate token estimation using tiktoken.

Check if the parse tree contains syntax errors.

Preprocess text into tokens.

Create a file entity with standard observations.

Generate deterministic ID from content (same as base.py).

Load service configuration from file.

Calculate estimated cost for token count.

Extract function name from various function node types.

Process a file system event by adding it to the coalescer.

Check if collection exists.

Create an 'inherits' relationship.

Generate embedding for a single text.

Get all JSONL files for a project.

Get the appropriate parser for a file.

Load project configuration.

Parse the file and extract entities and relations.

Parse Markdown file to extract documentation entities.

Parse a file using the appropriate parser.

Parse a single JSONL file into a conversation.

Parse content into tree-sitter AST.

Search for similar entities/relations. Args: collection_name: Name of the collection to search query: Search query te

Stop the background service.

Convert chat chunk to vector storage payload.

Truncate text to fit within token limits.

Check and enforce rate limits.

Execute embedding operation with retry logic.

Process all pending events.

Check if an import is internal to the project by checking if the module file exists.

Determine meaningfulness using AST structural heuristics.

Load previous run statistics from state file.

Unified Git+Meta setup for ALL indexing flows.

Store entities, relations, and implementation chunks with Git+Meta content deduplication.

Clear collection data. By default, preserves manually-added memories. Args: collection_name: Name of the collection p

Generate SHA256 hash of content

Create a new Qdrant collection.

Get the dimensionality of the embedding vector.

Generate embeddings for multiple texts with enhanced token-based batching.

Generate embedding for a single text.

Ensure collection exists, create if necessary.

Get detailed statistics for a collection.

Search for similar vectors.

Setup global logging configuration.

Start the background service.

Insert or update points in the collection with improved reliability.

Validate state file JSON structure and file contents.

Calculate estimated cost for token count.

Create common indexer components (project, logger, config, embedder, vector_store, indexer). Args: project_path: Path to the project root

Detect the MCP collection name for this project.

Detect the project root directory using Claude-first weighted scoring.

Accurate token estimation using tiktoken (cl100k_base).

Extract function calls from source code using regex.

Extract meaningful text content from a content item.

Extract Python function and class names.

Extract keys from a YAML mapping under a parent key.

Extract a value from YAML by key path (e.g., 'metadata.name').