Normalize entity order for undirected graph consistency
• Normalize entity pairs for storage • Update API docs for undirected edges
This commit is contained in:
yangdx
parent
3fbd704bf9
commit
bf1897a67e
2 changed files with 41 additions and 9 deletions
|
|
@ -299,6 +299,7 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
|
|||
entity_data (dict): Entity properties including:
|
||||
- description (str): Textual description of the entity
|
||||
- entity_type (str): Category/type of the entity (e.g., PERSON, ORGANIZATION, LOCATION)
|
||||
- source_id (str): Related chunk_id from which the description originates
|
||||
- Additional custom properties as needed
|
||||
|
||||
Response Schema:
|
||||
|
|
@ -309,6 +310,7 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
|
|||
"entity_name": "Tesla",
|
||||
"description": "Electric vehicle manufacturer",
|
||||
"entity_type": "ORGANIZATION",
|
||||
"source_id": "chunk-123<SEP>chunk-456"
|
||||
... (other entity properties)
|
||||
}
|
||||
}
|
||||
|
|
@ -361,10 +363,11 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
|
|||
"""
|
||||
Create a new relationship between two entities in the knowledge graph
|
||||
|
||||
This endpoint establishes a directed relationship between two existing entities.
|
||||
Both the source and target entities must already exist in the knowledge graph.
|
||||
The system automatically generates vector embeddings for the relationship to
|
||||
enable semantic search and graph traversal.
|
||||
This endpoint establishes an undirected relationship between two existing entities.
|
||||
The provided source/target order is accepted for convenience, but the backend
|
||||
stored edge is undirected and may be returned with the entities swapped.
|
||||
Both entities must already exist in the knowledge graph. The system automatically
|
||||
generates vector embeddings for the relationship to enable semantic search and graph traversal.
|
||||
|
||||
Prerequisites:
|
||||
- Both source_entity and target_entity must exist in the knowledge graph
|
||||
|
|
@ -376,6 +379,7 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
|
|||
relation_data (dict): Relationship properties including:
|
||||
- description (str): Textual description of the relationship
|
||||
- keywords (str): Comma-separated keywords describing the relationship type
|
||||
- source_id (str): Related chunk_id from which the description originates
|
||||
- weight (float): Relationship strength/importance (default: 1.0)
|
||||
- Additional custom properties as needed
|
||||
|
||||
|
|
@ -388,6 +392,7 @@ def create_graph_routes(rag, api_key: Optional[str] = None):
|
|||
"tgt_id": "Tesla",
|
||||
"description": "Elon Musk is the CEO of Tesla",
|
||||
"keywords": "CEO, founder",
|
||||
"source_id": "chunk-123<SEP>chunk-456"
|
||||
"weight": 1.0,
|
||||
... (other relationship properties)
|
||||
}
|
||||
|
|
|
|||
|
|
@ -100,6 +100,10 @@ async def adelete_by_relation(
|
|||
# Use graph database lock to ensure atomic graph and vector db operations
|
||||
async with graph_db_lock:
|
||||
try:
|
||||
# Normalize entity order for undirected graph (ensures consistent key generation)
|
||||
if source_entity > target_entity:
|
||||
source_entity, target_entity = target_entity, source_entity
|
||||
|
||||
# Check if the relation exists
|
||||
edge_exists = await chunk_entity_relation_graph.has_edge(
|
||||
source_entity, target_entity
|
||||
|
|
@ -878,6 +882,10 @@ async def acreate_relation(
|
|||
source_entity, target_entity, edge_data
|
||||
)
|
||||
|
||||
# Normalize entity order for undirected relation vector (ensures consistent key generation)
|
||||
if source_entity > target_entity:
|
||||
source_entity, target_entity = target_entity, source_entity
|
||||
|
||||
# Prepare content for embedding
|
||||
description = edge_data.get("description", "")
|
||||
keywords = edge_data.get("keywords", "")
|
||||
|
|
@ -1121,19 +1129,27 @@ async def amerge_entities(
|
|||
tgt = rel_data["tgt"]
|
||||
edge_data = rel_data["data"]
|
||||
|
||||
# Normalize entity order for consistent vector storage
|
||||
normalized_src, normalized_tgt = sorted([src, tgt])
|
||||
|
||||
description = edge_data.get("description", "")
|
||||
keywords = edge_data.get("keywords", "")
|
||||
source_id = edge_data.get("source_id", "")
|
||||
weight = float(edge_data.get("weight", 1.0))
|
||||
|
||||
content = f"{keywords}\t{src}\n{tgt}\n{description}"
|
||||
relation_id = compute_mdhash_id(src + tgt, prefix="rel-")
|
||||
# Use normalized order for content and relation ID
|
||||
content = (
|
||||
f"{keywords}\t{normalized_src}\n{normalized_tgt}\n{description}"
|
||||
)
|
||||
relation_id = compute_mdhash_id(
|
||||
normalized_src + normalized_tgt, prefix="rel-"
|
||||
)
|
||||
|
||||
relation_data_for_vdb = {
|
||||
relation_id: {
|
||||
"content": content,
|
||||
"src_id": src,
|
||||
"tgt_id": tgt,
|
||||
"src_id": normalized_src,
|
||||
"tgt_id": normalized_tgt,
|
||||
"source_id": source_id,
|
||||
"description": description,
|
||||
"keywords": keywords,
|
||||
|
|
@ -1340,7 +1356,18 @@ async def get_relation_info(
|
|||
tgt_entity: str,
|
||||
include_vector_data: bool = False,
|
||||
) -> dict[str, str | None | dict[str, str]]:
|
||||
"""Get detailed information of a relationship"""
|
||||
"""
|
||||
Get detailed information of a relationship between two entities.
|
||||
Relationship is unidirectional, swap src_entity and tgt_entity does not change the relationship.
|
||||
|
||||
Args:
|
||||
src_entity: Source entity name
|
||||
tgt_entity: Target entity name
|
||||
include_vector_data: Whether to include vector database information
|
||||
|
||||
Returns:
|
||||
Dictionary containing relationship information
|
||||
"""
|
||||
|
||||
# Get information from the graph
|
||||
edge_data = await chunk_entity_relation_graph.get_edge(src_entity, tgt_entity)
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue