graphrag-core 0.2.0

Core portable library for GraphRAG - works on native and WASM
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70

# GraphRAG Configuration Template: General Purpose
# ================================================
# Best for: General documents, articles, reports, mixed content
#
# Usage:
#   cp templates/general.toml ./graphrag.toml
#   # or: graphrag init --template general

# Output directory for graph data and embeddings
output_dir = "./graphrag-output"

# Pipeline approach: "semantic" (LLM-based), "algorithmic" (pattern-based), "hybrid"
approach = "hybrid"

# Text chunking settings
chunk_size = 1000
chunk_overlap = 200

# Retrieval settings
top_k_results = 10
similarity_threshold = 0.7

[embeddings]
# Embedding backend: "hash" (fast, offline), "ollama" (local LLM), "candle" (neural)
backend = "hash"
dimension = 384
fallback_to_hash = true
batch_size = 32

[entities]
# Entity extraction settings
min_confidence = 0.7
entity_types = ["PERSON", "ORGANIZATION", "LOCATION", "DATE", "EVENT"]
use_gleaning = false
max_gleaning_rounds = 3

[graph]
max_connections = 10
similarity_threshold = 0.8
extract_relationships = true
relationship_confidence_threshold = 0.5

[graph.traversal]
max_depth = 3
max_paths = 10
use_edge_weights = true
min_relationship_strength = 0.3

[retrieval]
top_k = 10
search_algorithm = "cosine"

[parallel]
enabled = true
num_threads = 0  # 0 = auto-detect
min_batch_size = 10

[ollama]
enabled = false
host = "localhost"
port = 11434
chat_model = "llama3.2:3b"
embedding_model = "nomic-embed-text"
timeout_seconds = 30
enable_caching = true

[auto_save]
enabled = false
interval_seconds = 300
max_versions = 5