Fix/Update Qdrant: set field_schema to KEYWORD & add encoding="utf-8" for file opening

wren-ai-service/eval/preparation.py

@@ -89,7 +89,7 @@ def get_database_names(path: Path):
 
 
 def get_tables_by_db(path: Path, key: str):
-    with open(path, "rb") as f:
+    with open(path, "rb",encoding="utf-8") as f:
         json_data = orjson.loads(f.read())
 
     return {item[key]: item for item in json_data}
@@ -210,7 +210,7 @@ def build_mdl_relationships(tables_info):
 
 
 def get_ground_truths_by_db(path: Path, key: str):
-    with open(path, "rb") as f:
+    with open(path, "rb",encoding="utf-8") as f:
         json_data = orjson.loads(f.read())
 
     results = defaultdict(list)

wren-ai-service/src/pipelines/generation/chart_adjustment.py

@@ -165,7 +165,7 @@ def __init__(
             "post_processor": ChartGenerationPostProcessor(),
         }
 
-        with open("src/pipelines/generation/utils/vega-lite-schema-v5.json", "r") as f:
+        with open("src/pipelines/generation/utils/vega-lite-schema-v5.json", "r",encoding="utf-8") as f:
             _vega_schema = orjson.loads(f.read())
 
         self._configs = {

wren-ai-service/src/pipelines/generation/chart_generation.py

@@ -138,7 +138,7 @@ def __init__(
             "post_processor": ChartGenerationPostProcessor(),
         }
 
-        with open("src/pipelines/generation/utils/vega-lite-schema-v5.json", "r") as f:
+        with open("src/pipelines/generation/utils/vega-lite-schema-v5.json", "r",encoding="utf-8") as f:
             _vega_schema = orjson.loads(f.read())
 
         self._configs = {

wren-ai-service/src/providers/document_store/qdrant.py

@@ -1,7 +1,7 @@
 import logging
 import os
 from typing import Any, Dict, List, Optional
-
+from qdrant_client.http import models
 import numpy as np
 import qdrant_client
 from haystack import Document, component
@@ -159,7 +159,8 @@ def __init__(
         # to improve the indexing performance
         # see https://qdrant.tech/documentation/guides/multiple-partitions/?q=mul#calibrate-performance
         self.client.create_payload_index(
-            collection_name=index, field_name="project_id", field_schema="keyword"
+            collection_name=index, field_name="project_id",
+              field_schema=models.PayloadSchemaType.KEYWORD,
         )
 
     async def _query_by_embedding(