Skip to content

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

fix(vectordb): resolve timestamp format and collection creation issues (#154)

kkkwjxd89d869176a0
diff --git a/openviking/core/context.py b/openviking/core/context.py
index 24d836872..ba816085b 100644
--- a/openviking/core/context.py
+++ b/openviking/core/context.py
@@ -2,11 +2,12 @@
# SPDX-License-Identifier: Apache-2.0
"""Unified context class for OpenViking."""
-from datetime import datetime
+from datetime import datetime, timezone
from enum import Enum
from typing import Any, Dict, List, Optional
from uuid import uuid4
+from openviking.utils.time_utils import format_iso8601
from openviking_cli.session.user_id import UserIdentifier
@@ -70,7 +71,7 @@ class Context:
self.abstract = abstract
self.context_type = context_type or self._derive_context_type()
self.category = category or self._derive_category()
- self.created_at = created_at or datetime.now()
+ self.created_at = created_at or datetime.now(timezone.utc)
self.updated_at = updated_at or self.created_at
self.active_count = active_count
self.related_uri = related_uri or []
@@ -122,10 +123,13 @@ class Context:
def update_activity(self):
"""Update activity statistics."""
self.active_count += 1
- self.updated_at = datetime.now()
+ self.updated_at = datetime.now(timezone.utc)
def to_dict(self) -> Dict[str, Any]:
"""Convert context to dictionary format for storage."""
+ created_at_str = format_iso8601(self.created_at) if self.created_at else None
+ updated_at_str = format_iso8601(self.updated_at) if self.updated_at else None
+
data = {
"id": self.id,
"uri": self.uri,
@@ -134,8 +138,8 @@ class Context:
"abstract": self.abstract,
"context_type": self.context_type,
"category": self.category,
- "created_at": self.created_at.isoformat(),
- "updated_at": self.updated_at.isoformat() if self.updated_at else None,
+ "created_at": created_at_str,
+ "updated_at": updated_at_str,
"active_count": self.active_count,
"vector": self.vector,
"meta": self.meta,
diff --git a/openviking/message/message.py b/openviking/message/message.py
index af9cfe2a4..90f5fdd3d 100644
--- a/openviking/message/message.py
+++ b/openviking/message/message.py
@@ -7,10 +7,11 @@ Message = role + parts, supports serialization to JSONL.
import json
from dataclasses import dataclass
-from datetime import datetime
+from datetime import datetime, timezone
from typing import List, Literal, Optional
from openviking.message.part import ContextPart, Part, TextPart, ToolPart
+from openviking.utils.time_utils import format_iso8601
@dataclass
@@ -32,11 +33,13 @@ class Message:
def to_dict(self) -> dict:
"""Serialize to JSONL."""
+ created_at_val = self.created_at or datetime.now(timezone.utc)
+ created_at_str = format_iso8601(created_at_val)
return {
"id": self.id,
"role": self.role,
"parts": [self._part_to_dict(p) for p in self.parts],
- "created_at": (self.created_at or datetime.now()).isoformat(),
+ "created_at": created_at_str,
}
def _part_to_dict(self, part: Part) -> dict:
@@ -108,7 +111,7 @@ class Message:
id=msg_id or f"msg_{uuid4().hex}",
role="user",
parts=[TextPart(text=content)],
- created_at=datetime.now(),
+ created_at=datetime.now(timezone.utc),
)
@classmethod
@@ -151,7 +154,7 @@ class Message:
id=msg_id or f"msg_{uuid4().hex}",
role="assistant",
parts=parts,
- created_at=datetime.now(),
+ created_at=datetime.now(timezone.utc),
)
def get_context_parts(self) -> List[ContextPart]:
diff --git a/openviking/models/vlm/base.py b/openviking/models/vlm/base.py
index 25d63aa01..2efeb192e 100644
--- a/openviking/models/vlm/base.py
+++ b/openviking/models/vlm/base.py
@@ -6,6 +6,8 @@ from abc import ABC, abstractmethod
from pathlib import Path
from typing import Any, Dict, List, Union
+from openviking.utils.time_utils import format_iso8601
+
from .token_usage import TokenUsageTracker
@@ -94,7 +96,7 @@ class VLMBase(ABC):
"total_prompt_tokens": total_usage.prompt_tokens,
"total_completion_tokens": total_usage.completion_tokens,
"total_tokens": total_usage.total_tokens,
- "last_updated": total_usage.last_updated.isoformat(),
+ "last_updated": format_iso8601(total_usage.last_updated),
}
def reset_token_usage(self) -> None:
diff --git a/openviking/models/vlm/token_usage.py b/openviking/models/vlm/token_usage.py
index 3df8db6ab..8b3219c14 100644
--- a/openviking/models/vlm/token_usage.py
+++ b/openviking/models/vlm/token_usage.py
@@ -6,6 +6,8 @@ from dataclasses import dataclass, field
from datetime import datetime
from typing import Dict, Optional
+from openviking.utils.time_utils import format_iso8601
+
@dataclass
class TokenUsage:
@@ -45,7 +47,7 @@ class TokenUsage:
"prompt_tokens": self.prompt_tokens,
"completion_tokens": self.completion_tokens,
"total_tokens": self.total_tokens,
- "last_updated": self.last_updated.isoformat(),
+ "last_updated": format_iso8601(self.last_updated),
}
def __str__(self) -> str:
diff --git a/openviking/session/session.py b/openviking/session/session.py
index 6dcbe6759..30f9490df 100644
--- a/openviking/session/session.py
+++ b/openviking/session/session.py
@@ -13,6 +13,7 @@ from typing import TYPE_CHECKING, Any, Dict, List, Optional
from uuid import uuid4
from openviking.message import Message, Part
+from openviking.utils.time_utils import get_current_timestamp
from openviking_cli.session.user_id import UserIdentifier
from openviking_cli.utils import get_logger, run_async
from openviking_cli.utils.config import get_openviking_config
@@ -57,7 +58,7 @@ class Usage:
input: str = ""
output: str = ""
success: bool = True
- timestamp: str = field(default_factory=lambda: datetime.now().isoformat())
+ timestamp: str = field(default_factory=get_current_timestamp)
class Session:
@@ -498,7 +499,7 @@ class Session:
"input": tool_part.tool_input,
"output": output,
"status": status,
- "time": {"created": datetime.now().isoformat()},
+ "time": {"created": get_current_timestamp()},
}
run_async(
self._viking_fs.write_file(
diff --git a/openviking/storage/collection_schemas.py b/openviking/storage/collection_schemas.py
index 0ee61069b..83f55d5d1 100644
--- a/openviking/storage/collection_schemas.py
+++ b/openviking/storage/collection_schemas.py
@@ -57,6 +57,18 @@ class CollectionSchemas:
{"FieldName": "tags", "FieldType": "string"},
{"FieldName": "abstract", "FieldType": "string"},
],
+ "ScalarIndex": [
+ "uri",
+ "type",
+ "context_type",
+ "created_at",
+ "updated_at",
+ "active_count",
+ "parent_uri",
+ "is_leaf",
+ "name",
+ "tags",
+ ],
}
diff --git a/openviking/storage/vectordb/collection/volcengine_collection.py b/openviking/storage/vectordb/collection/volcengine_collection.py
index 41a67ffd6..15f34fbc8 100644
--- a/openviking/storage/vectordb/collection/volcengine_collection.py
+++ b/openviking/storage/vectordb/collection/volcengine_collection.py
@@ -59,6 +59,8 @@ def get_or_create_volcengine_collection(config: Dict[str, Any], meta_data: Dict[
logger.error(f"Failed to create collection: {e}")
raise e
+ return VolcengineCollection(ak, sk, region, host, meta_data)
+
class VolcengineCollection(ICollection):
def __init__(
@@ -277,7 +279,6 @@ class VolcengineCollection(ICollection):
]
if "ids_not_exist" in data:
result.ids_not_exist = data.get("ids_not_exist", [])
- # print(result)
return result
def _parse_search_result(self, data: Dict[str, Any]) -> SearchResult:
diff --git a/openviking/storage/viking_fs.py b/openviking/storage/viking_fs.py
index b68c64ffc..d7a9f6476 100644
--- a/openviking/storage/viking_fs.py
+++ b/openviking/storage/viking_fs.py
@@ -15,13 +15,13 @@ Responsibilities:
import asyncio
import json
from dataclasses import dataclass, field
-from datetime import datetime
from pathlib import PurePath
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Union
from pyagfs import AGFSClient
from openviking.storage.vikingdb_interface import VikingDBInterface
+from openviking.utils.time_utils import get_current_timestamp
from openviking_cli.utils.logger import get_logger
if TYPE_CHECKING:
@@ -40,7 +40,7 @@ class RelationEntry:
id: str
uris: List[str]
reason: str = ""
- created_at: str = field(default_factory=lambda: datetime.now().isoformat())
+ created_at: str = field(default_factory=get_current_timestamp)
def to_dict(self) -> Dict[str, Any]:
return {
@@ -292,10 +292,10 @@ class VikingFS:
Returns:
FindResult
"""
+ from openviking.retrieve.hierarchical_retriever import HierarchicalRetriever
from openviking_cli.retrieve import (
ContextType,
FindResult,
- HierarchicalRetriever,
TypedQuery,
)
diff --git a/openviking/storage/viking_vector_index_backend.py b/openviking/storage/viking_vector_index_backend.py
index 5854b3e10..4f1b7efd4 100644
--- a/openviking/storage/viking_vector_index_backend.py
+++ b/openviking/storage/viking_vector_index_backend.py
@@ -225,8 +225,11 @@ class VikingVectorIndexBackend(VikingDBInterface):
logger.debug(f"Collection '{name}' already exists")
return False
- # Extract configuration from schema
- collection_meta = schema
+ collection_meta = schema.copy()
+
+ scalar_index_fields = []
+ if "ScalarIndex" in collection_meta:
+ scalar_index_fields = collection_meta.pop("ScalarIndex")
# Ensure CollectionName is set
if "CollectionName" not in collection_meta:
@@ -245,21 +248,23 @@ class VikingVectorIndexBackend(VikingDBInterface):
# Create collection using vectordb project
collection = self.project.create_collection(name, collection_meta)
- # Build scalar index fields list from Fields
- scalar_index_fields = []
- exclude_types = {"vector", "sparse_vector", "abstract"}
-
- for field in collection_meta.get("Fields", []):
- field_name = field.get("FieldName")
- field_type = field.get("FieldType")
- is_primary_key = field.get("IsPrimaryKey", False)
- # Index all non-vector and non-primary-key fields by default
- if field_name and field_type not in exclude_types and not is_primary_key:
- scalar_index_fields.append(field_name)
+ # Filter date_time fields for volcengine and vikingdb backends
+ if self._mode in ["volcengine", "vikingdb"]:
+ date_time_fields = {
+ field.get("FieldName")
+ for field in collection_meta.get("Fields", [])
+ if field.get("FieldType") == "date_time"
+ }
+ scalar_index_fields = [
+ field for field in scalar_index_fields if field not in date_time_fields
+ ]
# Create default index for the collection
use_sparse = self.sparse_weight > 0.0
index_type = "flat_hybrid" if use_sparse else "flat"
+ if self._mode in ["volcengine", "vikingdb"]:
+ index_type = "hnsw_hybrid" if use_sparse else "hnsw"
+
index_meta = {
"IndexName": self.DEFAULT_INDEX_NAME,
"VectorIndex": {
diff --git a/openviking/utils/__init__.py b/openviking/utils/__init__.py
index c2ecc203b..a86f76ff9 100644
--- a/openviking/utils/__init__.py
+++ b/openviking/utils/__init__.py
@@ -2,6 +2,7 @@
# SPDX-License-Identifier: Apache-2.0
"""Utility functions and helpers."""
+from openviking.utils.time_utils import get_current_timestamp
from openviking_cli.utils.async_utils import run_async
from openviking_cli.utils.llm import StructuredLLM, parse_json_from_response, parse_json_to_model
from openviking_cli.utils.logger import default_logger, get_logger
@@ -11,6 +12,7 @@ __all__ = [
"VikingURI",
"get_logger",
"default_logger",
+ "get_current_timestamp",
"StructuredLLM",
"parse_json_from_response",
"parse_json_to_model",
diff --git a/openviking/utils/time_utils.py b/openviking/utils/time_utils.py
new file mode 100644
index 000000000..db6df6ecf
--- /dev/null
+++ b/openviking/utils/time_utils.py
@@ -0,0 +1,26 @@
+from datetime import datetime, timezone
+
+
+def format_iso8601(dt: datetime) -> str:
+ """
+ Format datetime object to ISO 8601 format compatible with VikingDB.
+
+ Format: yyyy-MM-ddTHH:mm:ss.SSSZ (UTC)
+ """
+ # Ensure dt is timezone-aware and in UTC
+ if dt.tzinfo is None:
+ dt = dt.replace(tzinfo=timezone.utc)
+ else:
+ dt = dt.astimezone(timezone.utc)
+
+ return dt.isoformat(timespec="milliseconds").replace("+00:00", "Z")
+
+
+def get_current_timestamp() -> str:
+ """
+ Get current timestamp in ISO 8601 format compatible with VikingDB.
+
+ Format: yyyy-MM-ddTHH:mm:ss.SSSZ (UTC)
+ """
+ now = datetime.now(timezone.utc)
+ return format_iso8601(now)