gardel/lightrag
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Added search mode and min degree filtering for NetworkX

Browse Source 
- Implemented exact and inclusive search modes
- Added min degree filtering for nodes
- Updated API to parse label for search options
This commit is contained in:
yangdx
2025-03-04 16:08:05 +08:00
parent 735231d851
commit 002948d342
3 changed files with 70 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
lightrag/api/routers/graph_routes.py
View File

@@ -34,6 +34,11 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
2. Followed by nodes directly connected to the matching nodes
3. Finally, the degree of the nodes
Maximum number of nodes is limited to env MAX_GRAPH_NODES(default: 1000)
Control search mode by label content:
1. only label-name : exact search with the label name (selecting from the label list return previously)
2. label-name follow by '>n' : exact search of nodes with degree more than n
3. label-name follow by* : inclusive search of nodes with degree more than n
4. label-name follow by '>n*' : inclusive search
Args:
label (str): Label to get knowledge graph for
@@ -42,6 +47,37 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
Returns:
Dict[str, List[str]]: Knowledge graph for label
"""
return await rag.get_knowledge_graph(node_label=label, max_depth=max_depth)
# Parse label to extract search mode and min degree if specified
search_mode = "exact" # Default search mode
min_degree = 0 # Default minimum degree
original_label = label
# First check if label ends with *
if label.endswith("*"):
search_mode = "inclusive" # Always set to inclusive if ends with *
label = label[:-1].strip() # Remove trailing *
# Try to parse >n if it exists
if ">" in label:
try:
degree_pos = label.rfind(">")
degree_str = label[degree_pos + 1:].strip()
min_degree = int(degree_str) + 1
label = label[:degree_pos].strip()
except ValueError:
# If degree parsing fails, just remove * and keep the rest as label
label = original_label[:-1].strip()
# If no *, check for >n pattern
elif ">" in label:
try:
degree_pos = label.rfind(">")
degree_str = label[degree_pos + 1:].strip()
min_degree = int(degree_str) + 1
label = label[:degree_pos].strip()
except ValueError:
# If degree parsing fails, treat the whole string as label
label = original_label
return await rag.get_knowledge_graph(node_label=label, max_depth=max_depth, search_mode=search_mode, min_degree=min_degree)
return router

21
lightrag/kg/networkx_impl.py
View File

@@ -232,7 +232,7 @@ class NetworkXStorage(BaseGraphStorage):
return sorted(list(labels))
async def get_knowledge_graph(
self, node_label: str, max_depth: int = 5
self, node_label: str, max_depth: int = 5, search_mode: str = "exact", min_degree: int = 0
) -> KnowledgeGraph:
"""
Retrieve a connected subgraph of nodes where the label includes the specified `node_label`.
@@ -245,6 +245,8 @@ class NetworkXStorage(BaseGraphStorage):
Args:
node_label: Label of the starting node
max_depth: Maximum depth of the subgraph
search_mode (str, optional): Search mode, either "exact" or "inclusive". Defaults to "exact".
min_degree (int, optional): Minimum degree of nodes to include. Defaults to 0.
Returns:
KnowledgeGraph object containing nodes and edges
@@ -262,11 +264,16 @@ class NetworkXStorage(BaseGraphStorage):
graph.copy()
) # Create a copy to avoid modifying the original graph
else:
# Find nodes with matching node id (partial match)
# Find nodes with matching node id based on search_mode
nodes_to_explore = []
for n, attr in graph.nodes(data=True):
if node_label in str(n): # Use partial matching
nodes_to_explore.append(n)
node_str = str(n)
if search_mode == "exact":
if node_label == node_str: # Use exact matching
nodes_to_explore.append(n)
else: # inclusive mode
if node_label in node_str: # Use partial matching
nodes_to_explore.append(n)
if not nodes_to_explore:
logger.warning(f"No nodes found with label {node_label}")
@@ -277,6 +284,12 @@ class NetworkXStorage(BaseGraphStorage):
for start_node in nodes_to_explore:
node_subgraph = nx.ego_graph(graph, start_node, radius=max_depth)
combined_subgraph = nx.compose(combined_subgraph, node_subgraph)
# Filter nodes based on min_degree
if min_degree > 0:
nodes_to_keep = [node for node, degree in combined_subgraph.degree() if degree >= min_degree]
combined_subgraph = combined_subgraph.subgraph(nodes_to_keep)
subgraph = combined_subgraph
# Check if number of nodes exceeds max_graph_nodes

18
lightrag/lightrag.py
View File

@@ -504,10 +504,24 @@ class LightRAG:
return text
async def get_knowledge_graph(
self, node_label: str, max_depth: int
self, node_label: str, max_depth: int, search_mode: str = "exact", min_degree: int = 0
) -> KnowledgeGraph:
"""Get knowledge graph for a given label
Args:
node_label (str): Label to get knowledge graph for
max_depth (int): Maximum depth of graph
search_mode (str, optional): Search mode, either "exact" or "inclusive". Defaults to "exact".
min_degree (int, optional): Minimum degree of nodes to include. Defaults to 0.
Returns:
KnowledgeGraph: Knowledge graph containing nodes and edges
"""
return await self.chunk_entity_relation_graph.get_knowledge_graph(
node_label=node_label, max_depth=max_depth
node_label=node_label,
max_depth=max_depth,
search_mode=search_mode,
min_degree=min_degree
)
def _get_storage_class(self, storage_name: str) -> Callable[..., Any]: