godber/EmbeddingBuddy
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 7 Wiki Activity

opensearch load improvements
Some checks failed
Security Scan / dependency-check (pull_request) Successful in 44s
Details
Test Suite / lint (pull_request) Failing after 32s
Details
Security Scan / security (pull_request) Successful in 45s
Details
Test Suite / test (3.11) (pull_request) Successful in 1m31s
Details
Test Suite / build (pull_request) Has been skipped
Details

Browse Source
This commit is contained in:
Austin Godber
2025-08-14 14:30:52 -07:00
parent 9cf2f0e6fa
commit 09e3c86f0a
4 changed files with 375 additions and 232 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/embeddingbuddy/models/field_mapper.py
View File

@@ -39,7 +39,6 @@ class FieldMapper:
vector_fields = [vf["name"] for vf in field_analysis.get("vector_fields", [])]
text_fields = field_analysis.get("text_fields", [])
keyword_fields = field_analysis.get("keyword_fields", [])
numeric_fields = field_analysis.get("numeric_fields", [])
# Helper function to create ordered suggestions
def create_ordered_suggestions(primary_candidates, all_available_fields):
@@ -57,8 +56,16 @@ class FieldMapper:
suggestions = {}
# Embedding field suggestions (vector fields first, then all fields)
# Embedding field suggestions (vector fields first, then name-based candidates, then all fields)
embedding_candidates = vector_fields.copy()
# Add fields that likely contain embeddings based on name
embedding_name_candidates = [f for f in all_fields if any(
keyword in f.lower() for keyword in ["embedding", "embeddings", "vector", "vectors", "embed"]
)]
# Add name-based candidates that aren't already in vector_fields
for candidate in embedding_name_candidates:
if candidate not in embedding_candidates:
embedding_candidates.append(candidate)
suggestions["embedding"] = create_ordered_suggestions(embedding_candidates, all_fields)
# Text field suggestions (text fields first, then all fields)