zlei6/ragflow
1
0
Fork 0
mirror of https://github.com/infiniflow/ragflow.git synced 2026-08-05 07:10:29 +08:00
Code Issues Packages Projects Releases Wiki Activity

refa: simplify RAPTOR tree clustering and configuration (#17614)

Browse Source
This commit is contained in:
buua436
2026-08-03 17:46:50 +08:00
committed by GitHub
parent e290a0d476
commit 3e7cfbe052
32 changed files with 318 additions and 1356 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
rag/utils/raptor_utils.py
View File

@@ -25,11 +25,6 @@ from typing import Optional
import xxhash
RAPTOR_TREE_BUILDER = "raptor"
PSI_TREE_BUILDER = "psi"
SUPPORTED_TREE_BUILDERS = {RAPTOR_TREE_BUILDER, PSI_TREE_BUILDER}
GMM_CLUSTERING_METHOD = "gmm"
AHC_CLUSTERING_METHOD = "ahc"
SUPPORTED_CLUSTERING_METHODS = {GMM_CLUSTERING_METHOD, AHC_CLUSTERING_METHOD}
# File extensions for structured data types
EXCEL_EXTENSIONS = {".xls", ".xlsx", ".xlsm", ".xlsb"}
@@ -37,26 +32,6 @@ CSV_EXTENSIONS = {".csv", ".tsv"}
STRUCTURED_EXTENSIONS = EXCEL_EXTENSIONS | CSV_EXTENSIONS
def get_raptor_tree_builder(raptor_config: dict | None) -> str:
"""Return the configured RAPTOR tree builder with legacy ext fallback."""
raptor_config = raptor_config or {}
ext = raptor_config.get("ext") or {}
tree_builder = ext.get("tree_builder") or raptor_config.get("tree_builder") or RAPTOR_TREE_BUILDER
if tree_builder not in SUPPORTED_TREE_BUILDERS:
raise ValueError(f"Unsupported RAPTOR tree builder: {tree_builder}")
return tree_builder
def get_raptor_clustering_method(raptor_config: dict | None) -> str:
"""Return the configured RAPTOR clustering method with legacy ext fallback."""
raptor_config = raptor_config or {}
ext = raptor_config.get("ext") or {}
clustering_method = ext.get("clustering_method") or raptor_config.get("clustering_method") or GMM_CLUSTERING_METHOD
if clustering_method not in SUPPORTED_CLUSTERING_METHODS:
raise ValueError(f"Unsupported RAPTOR clustering method: {clustering_method}")
return clustering_method
def _as_extra_dict(extra) -> dict:
"""Normalize a chunk extra payload into a dictionary."""
if isinstance(extra, dict):