Added infinity rank_feature support (#9044)

### What problem does this PR solve?

Added infinity rank_feature support

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

api/apps/kb_app.py

@@ -14,7 +14,6 @@
 #  limitations under the License.
 #
 import json
-import os
 
 from flask import request
 from flask_login import login_required, current_user
@@ -106,13 +105,6 @@ def update():
             return get_data_error_result(
                 message="Can't find this knowledgebase!")
 
-        if req.get("parser_id", "") == "tag" and os.environ.get('DOC_ENGINE', "elasticsearch") == "infinity":
-            return get_json_result(
-                data=False,
-                message='The chunking method Tag has not been supported by Infinity yet.',
-                code=settings.RetCode.OPERATING_ERROR
-            )
-
         if req["name"].lower() != kb.name.lower() \
                 and len(
             KnowledgebaseService.query(name=req["name"], tenant_id=current_user.id, status=StatusEnum.VALID.value)) >= 1:
@@ -124,9 +116,6 @@ def update():
             return get_data_error_result()
 
         if kb.pagerank != req.get("pagerank", 0):
-            if os.environ.get("DOC_ENGINE", "elasticsearch") != "elasticsearch":
-                return get_data_error_result(message="'pagerank' can only be set when doc_engine is elasticsearch")
-            
             if req.get("pagerank", 0) > 0:
                 settings.docStoreConn.update({"kb_id": kb.id}, {PAGERANK_FLD: req["pagerank"]},
                                          search.index_name(kb.tenant_id), kb.id)

conf/infinity_mapping.json

@@ -30,7 +30,7 @@
 	"knowledge_graph_kwd": {"type": "varchar", "default": ""},
 	"entities_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"pagerank_fea": {"type": "integer", "default":  0},
-	"tag_feas": {"type": "varchar", "default":  ""},
+	"tag_feas": {"type": "varchar", "default": "", "analyzer": "rankfeatures"},
 
 	"from_entity_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"to_entity_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},

docker/docker-compose-base.yml

@@ -77,7 +77,7 @@ services:
     container_name: ragflow-infinity
     profiles:
       - infinity
-    image: infiniflow/infinity:v0.6.0-dev4
+    image: infiniflow/infinity:v0.6.0-dev5
     volumes:
       - infinity_data:/var/infinity
       - ./infinity_conf.toml:/infinity_conf.toml

docker/infinity_conf.toml

@@ -17,7 +17,7 @@ log_file_max_size        = "100MB"
 log_file_rotate_count    = 10
 
 # trace/debug/info/warning/error/critical 6 log levels, default: info
-log_level               = "info"
+log_level               = "trace"
 
 [storage]
 persistence_dir         = "/var/infinity/persistence"
@@ -47,7 +47,7 @@ mem_index_capacity       = 65536
 buffer_manager_size      = "8GB"
 lru_num                  = 7
 temp_dir                 = "/var/infinity/tmp"
-result_cache             = "on"
+result_cache             = "off"
 memindex_memory_quota    = "1GB"
 
 [wal]

helm/values.yaml

@@ -113,7 +113,7 @@ ragflow:
 infinity:
   image:
     repository: infiniflow/infinity
-    tag: v0.6.0-dev4
+    tag: v0.6.0-dev5
   storage:
     className:
     capacity: 5Gi

rag/nlp/query.py

@@ -274,4 +274,4 @@ class FulltextQueryer:
                 keywords.append(f"{tk}^{w}")
 
         return MatchTextExpr(self.query_fields, " ".join(keywords), 100,
-                             {"minimum_should_match": min(3, len(keywords) / 10)})
+                             {"minimum_should_match": min(3, len(keywords) // 10)})

rag/nlp/search.py

@@ -111,7 +111,7 @@ class Dealer:
                 q_vec = matchDense.embedding_data
                 src.append(f"q_{len(q_vec)}_vec")
 
-                fusionExpr = FusionExpr("weighted_sum", topk, {"weights": "0.05, 0.95"})
+                fusionExpr = FusionExpr("weighted_sum", topk, {"weights": "0.05,0.95"})
                 matchExprs = [matchText, matchDense, fusionExpr]
 
                 res = self.dataStore.search(src, highlightFields, filters, matchExprs, orderBy, offset, limit,

rag/utils/infinity_conn.py

@@ -26,7 +26,7 @@ from infinity.index import IndexInfo, IndexType
 from infinity.connection_pool import ConnectionPool
 from infinity.errors import ErrorCode
 from rag import settings
-from rag.settings import PAGERANK_FLD
+from rag.settings import PAGERANK_FLD, TAG_FLD
 from rag.utils import singleton
 import pandas as pd
 from api.utils.file_utils import get_project_base_directory
@@ -311,7 +311,7 @@ class InfinityConnection(DocStoreConnection):
         df_list = list()
         table_list = list()
         output = selectFields.copy()
-        for essential_field in ["id"]:
+        for essential_field in ["id"] + aggFields:
             if essential_field not in output:
                 output.append(essential_field)
         score_func = ""
@@ -333,15 +333,29 @@ class InfinityConnection(DocStoreConnection):
             if PAGERANK_FLD not in output:
                 output.append(PAGERANK_FLD)
         output = [f for f in output if f != "_score"]
+        if limit <= 0:
+            # ElasticSearch default limit is 10000
+            limit = 10000
 
         # Prepare expressions common to all tables
         filter_cond = None
         filter_fulltext = ""
         if condition:
+            table_found = False
             for indexName in indexNames:
-                table_name = f"{indexName}_{knowledgebaseIds[0]}"
-                filter_cond = equivalent_condition_to_str(condition, db_instance.get_table(table_name))
-                break
+                for kb_id in knowledgebaseIds:
+                    table_name = f"{indexName}_{kb_id}"
+                    try:
+                        filter_cond = equivalent_condition_to_str(condition, db_instance.get_table(table_name))
+                        table_found = True
+                        break
+                    except Exception:
+                        pass
+                if table_found:
+                    break
+            if not table_found:
+                logger.error(f"No valid tables found for indexNames {indexNames} and knowledgebaseIds {knowledgebaseIds}")
+                return pd.DataFrame(), 0
 
         for matchExpr in matchExprs:
             if isinstance(matchExpr, MatchTextExpr):
@@ -355,6 +369,18 @@ class InfinityConnection(DocStoreConnection):
                 if isinstance(minimum_should_match, float):
                     str_minimum_should_match = str(int(minimum_should_match * 100)) + "%"
                     matchExpr.extra_options["minimum_should_match"] = str_minimum_should_match
+
+                # Add rank_feature support
+                if rank_feature and "rank_features" not in matchExpr.extra_options:
+                    # Convert rank_feature dict to Infinity's rank_features string format
+                    # Format: "field^feature_name^weight,field^feature_name^weight"
+                    rank_features_list = []
+                    for feature_name, weight in rank_feature.items():
+                        # Use TAG_FLD as the field containing rank features
+                        rank_features_list.append(f"{TAG_FLD}^{feature_name}^{weight}")
+                    if rank_features_list:
+                        matchExpr.extra_options["rank_features"] = ",".join(rank_features_list)
+
                 for k, v in matchExpr.extra_options.items():
                     if not isinstance(v, str):
                         matchExpr.extra_options[k] = str(v)
@@ -416,7 +442,7 @@ class InfinityConnection(DocStoreConnection):
                                 matchExpr.method, matchExpr.topn, matchExpr.fusion_params
                             )
                 else:
-                    if len(filter_cond) > 0:
+                    if filter_cond and len(filter_cond) > 0:
                         builder.filter(filter_cond)
                 if orderBy.fields:
                     builder.sort(order_by_expr_list)
@@ -662,6 +688,8 @@ class InfinityConnection(DocStoreConnection):
             k = column.lower()
             if field_keyword(k):
                 res2[column] = res2[column].apply(lambda v:[kwd for kwd in v.split("###") if kwd])
+            elif re.search(r"_feas$", k):
+                res2[column] = res2[column].apply(lambda v: json.loads(v) if v else {})
             elif k == "position_int":
                 def to_position_int(v):
                     if v:
@@ -712,9 +740,46 @@ class InfinityConnection(DocStoreConnection):
 
     def getAggregation(self, res: tuple[pd.DataFrame, int] | pd.DataFrame, fieldnm: str):
         """
-        TODO: Infinity doesn't provide aggregation
+        Manual aggregation for tag fields since Infinity doesn't provide native aggregation
         """
-        return list()
+        from collections import Counter
+
+        # Extract DataFrame from result
+        if isinstance(res, tuple):
+            df, _ = res
+        else:
+            df = res
+
+        if df.empty or fieldnm not in df.columns:
+            return []
+
+        # Aggregate tag counts
+        tag_counter = Counter()
+
+        for value in df[fieldnm]:
+            if pd.isna(value) or not value:
+                continue
+
+            # Handle different tag formats
+            if isinstance(value, str):
+                # Split by ### for tag_kwd field or comma for other formats
+                if fieldnm == "tag_kwd" and "###" in value:
+                    tags = [tag.strip() for tag in value.split("###") if tag.strip()]
+                else:
+                    # Try comma separation as fallback
+                    tags = [tag.strip() for tag in value.split(",") if tag.strip()]
+
+                for tag in tags:
+                    if tag:  # Only count non-empty tags
+                        tag_counter[tag] += 1
+            elif isinstance(value, list):
+                # Handle list format
+                for tag in value:
+                    if tag and isinstance(tag, str):
+                        tag_counter[tag.strip()] += 1
+
+        # Return as list of [tag, count] pairs, sorted by count descending
+        return [[tag, count] for tag, count in tag_counter.most_common()]
 
     """
     SQL

sdk/python/test/test_http_api/test_chunk_management_within_dataset/test_list_chunks.py

@@ -68,8 +68,8 @@ class TestChunksList:
         "params, expected_code, expected_page_size, expected_message",
         [
             ({"page_size": None}, 0, 5, ""),
-            pytest.param({"page_size": 0}, 0, 5, "", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") == "infinity", reason="Infinity does not support page_size=0")),
-            pytest.param({"page_size": 0}, 100, 0, "3013", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") in [None, "opensearch", "elasticsearch"], reason="Infinity does not support page_size=0")),
+            pytest.param({"page_size": 0}, 0, 5, ""),
+            pytest.param({"page_size": 0}, 100, 0, ""),
             ({"page_size": 1}, 0, 1, ""),
             ({"page_size": 6}, 0, 5, ""),
             ({"page_size": "1"}, 0, 1, ""),

test/testcases/test_http_api/test_chunk_management_within_dataset/test_list_chunks.py

@@ -69,8 +69,7 @@ class TestChunksList:
         "params, expected_code, expected_page_size, expected_message",
         [
             ({"page_size": None}, 0, 5, ""),
-            pytest.param({"page_size": 0}, 0, 5, "", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") == "infinity", reason="Infinity does not support page_size=0")),
-            pytest.param({"page_size": 0}, 100, 0, "3013", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") in [None, "opensearch", "elasticsearch"], reason="Infinity does not support page_size=0")),
+            pytest.param({"page_size": 0}, 0, 5, ""),
             ({"page_size": 1}, 0, 1, ""),
             ({"page_size": 6}, 0, 5, ""),
             ({"page_size": "1"}, 0, 1, ""),

test/testcases/test_sdk_api/test_chunk_management_within_dataset/test_list_chunks.py

@@ -50,8 +50,7 @@ class TestChunksList:
         "params, expected_page_size, expected_message",
         [
             ({"page_size": None}, 5, ""),
-            pytest.param({"page_size": 0}, 5, "", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") == "infinity", reason="Infinity does not support page_size=0")),
-            pytest.param({"page_size": 0}, 0, "3013", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") in [None, "opensearch", "elasticsearch"], reason="Infinity does not support page_size=0")),
+            pytest.param({"page_size": 0}, 5, ""),
             ({"page_size": 1}, 1, ""),
             ({"page_size": 6}, 5, ""),
             ({"page_size": "1"}, 1, ""),

test/testcases/test_web_api/test_chunk_app/test_list_chunks.py

@@ -68,8 +68,7 @@ class TestChunksList:
         "params, expected_code, expected_page_size, expected_message",
         [
             ({"size": None}, 100, 0, """TypeError("int() argument must be a string, a bytes-like object or a real number, not 'NoneType'")"""),
-            pytest.param({"size": 0}, 0, 5, "", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") == "infinity", reason="Infinity does not support page_size=0")),
-            pytest.param({"size": 0}, 100, 0, "3013", marks=pytest.mark.skipif(os.getenv("DOC_ENGINE") in [None, "opensearch", "elasticsearch"], reason="Infinity does not support page_size=0")),
+            pytest.param({"size": 0}, 0, 5, ""),
             ({"size": 1}, 0, 1, ""),
             ({"size": 6}, 0, 5, ""),
             ({"size": "1"}, 0, 1, ""),