Crate graphify_serve 

MCP server for graph queries.

Provides graph traversal and scoring functions used by the query engine and MCP protocol server. Port of Python query tools.

Modules

mcp

MCP (Model Context Protocol) server implementation.

search

In-memory inverted index for fast node lookup.

Enums

ServeError

Errors from the server.

SummaryLevel

Abstraction level for graph summaries.

Functions

all_simple_paths

Find all simple paths between source and target up to max_length edges.

bfs

BFS traversal from start nodes up to a maximum depth.

dfs

DFS traversal from start nodes up to a maximum depth.

dijkstra_path

Dijkstra shortest path using edge weights.

graph_stats

Get basic statistics about the graph.

load_graph

Load a knowledge graph from a JSON file.

score_nodes

Score nodes by relevance to search terms using a pre-built SearchIndex.

smart_summary

Generate a multi-level graph summary within a token budget.

start_server

Start the MCP server over stdio (JSON-RPC 2.0).

subgraph_to_text

Convert a subgraph (set of nodes and edges) to a text representation suitable for LLM context windows.

Type Aliases

EdgeDetail

Edge detail in a weighted path: (from_id, to_id, cost, relation).