veadk-python 0.2.6__py3-none-any.whl → 0.2.8__py3-none-any.whl

veadk/knowledgebase/backends/opensearch_backend.py

@@ -0,0 +1,136 @@
+# Copyright (c) 2025 Beijing Volcano Engine Technology Co., Ltd. and/or its affiliates.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import re
+
+from llama_index.core import (
+    Document,
+    SimpleDirectoryReader,
+    StorageContext,
+    VectorStoreIndex,
+)
+from llama_index.core.schema import BaseNode
+from llama_index.embeddings.openai_like import OpenAILikeEmbedding
+from pydantic import Field
+from typing_extensions import Any, override
+
+import veadk.config  # noqa E401
+from veadk.configs.database_configs import OpensearchConfig
+from veadk.configs.model_configs import EmbeddingModelConfig, NormalEmbeddingModelConfig
+from veadk.knowledgebase.backends.base_backend import BaseKnowledgebaseBackend
+from veadk.knowledgebase.backends.utils import get_llama_index_splitter
+
+try:
+    from llama_index.vector_stores.opensearch import (
+        OpensearchVectorClient,
+        OpensearchVectorStore,
+    )
+except ImportError:
+    raise ImportError(
+        "Please install VeADK extensions\npip install veadk-python[extensions]"
+    )
+
+
+class OpensearchKnowledgeBackend(BaseKnowledgebaseBackend):
+    opensearch_config: OpensearchConfig = Field(default_factory=OpensearchConfig)
+    """Opensearch client configs"""
+
+    embedding_config: EmbeddingModelConfig | NormalEmbeddingModelConfig = Field(
+        default_factory=EmbeddingModelConfig
+    )
+    """Embedding model configs"""
+
+    def model_post_init(self, __context: Any) -> None:
+        self.precheck_index_naming()
+        self._opensearch_client = OpensearchVectorClient(
+            endpoint=self.opensearch_config.host,
+            port=self.opensearch_config.port,
+            http_auth=(
+                self.opensearch_config.username,
+                self.opensearch_config.password,
+            ),
+            use_ssl=True,
+            verify_certs=False,
+            dim=self.embedding_config.dim,
+            index=self.index,  # collection name
+        )
+
+        self._vector_store = OpensearchVectorStore(client=self._opensearch_client)
+
+        self._storage_context = StorageContext.from_defaults(
+            vector_store=self._vector_store
+        )
+
+        self._embed_model = OpenAILikeEmbedding(
+            model_name=self.embedding_config.name,
+            api_key=self.embedding_config.api_key,
+            api_base=self.embedding_config.api_base,
+        )
+
+        self._vector_index = VectorStoreIndex.from_documents(
+            documents=[],
+            storage_context=self._storage_context,
+            embed_model=self._embed_model,
+        )
+
+    @override
+    def precheck_index_naming(self) -> None:
+        if not (
+            isinstance(self.index, str)
+            and not self.index.startswith(("_", "-"))
+            and self.index.islower()
+            and re.match(r"^[a-z0-9_\-.]+$", self.index)
+        ):
+            raise ValueError(
+                "The index name does not conform to the naming rules of OpenSearch"
+            )
+
+    @override
+    def add_from_directory(self, directory: str) -> bool:
+        documents = SimpleDirectoryReader(input_dir=directory).load_data()
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def add_from_files(self, files: list[str]) -> bool:
+        documents = SimpleDirectoryReader(input_files=files).load_data()
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def add_from_text(self, text: str | list[str]) -> bool:
+        if isinstance(text, str):
+            documents = [Document(text=text)]
+        else:
+            documents = [Document(text=t) for t in text]
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def search(self, query: str, top_k: int = 5) -> list[str]:
+        _retriever = self._vector_index.as_retriever(similarity_top_k=top_k)
+        retrieved_nodes = _retriever.retrieve(query)
+        return [node.text for node in retrieved_nodes]
+
+    def _split_documents(self, documents: list[Document]) -> list[BaseNode]:
+        """Split document into chunks"""
+        nodes = []
+        for document in documents:
+            splitter = get_llama_index_splitter(document.metadata.get("file_path", ""))
+            _nodes = splitter.get_nodes_from_documents([document])
+            nodes.extend(_nodes)
+        return nodes

veadk/knowledgebase/backends/redis_backend.py

@@ -0,0 +1,144 @@
+# Copyright (c) 2025 Beijing Volcano Engine Technology Co., Ltd. and/or its affiliates.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from llama_index.core import (
+    Document,
+    SimpleDirectoryReader,
+    StorageContext,
+    VectorStoreIndex,
+)
+from llama_index.core.schema import BaseNode
+from llama_index.embeddings.openai_like import OpenAILikeEmbedding
+from pydantic import Field
+from typing_extensions import Any, override
+
+import veadk.config  # noqa E401
+from veadk.configs.database_configs import RedisConfig
+from veadk.configs.model_configs import EmbeddingModelConfig, NormalEmbeddingModelConfig
+from veadk.knowledgebase.backends.base_backend import BaseKnowledgebaseBackend
+from veadk.knowledgebase.backends.utils import get_llama_index_splitter
+
+try:
+    from llama_index.vector_stores.redis import RedisVectorStore
+    from llama_index.vector_stores.redis.schema import (
+        RedisIndexInfo,
+        RedisVectorStoreSchema,
+    )
+    from redis import Redis
+    from redisvl.schema.fields import BaseVectorFieldAttributes
+except ImportError:
+    raise ImportError(
+        "Please install VeADK extensions\npip install veadk-python[extensions]"
+    )
+
+
+class RedisKnowledgeBackend(BaseKnowledgebaseBackend):
+    redis_config: RedisConfig = Field(default_factory=RedisConfig)
+    """Redis client configs"""
+
+    embedding_config: EmbeddingModelConfig | NormalEmbeddingModelConfig = Field(
+        default_factory=EmbeddingModelConfig
+    )
+    """Embedding model configs"""
+
+    def model_post_init(self, __context: Any) -> None:
+        # We will use `from_url` to init Redis client once the
+        # AK/SK -> STS token is ready.
+        # self._redis_client = Redis.from_url(url=...)
+
+        self._redis_client = Redis(
+            host=self.redis_config.host,
+            port=self.redis_config.port,
+            db=self.redis_config.db,
+            password=self.redis_config.password,
+        )
+
+        self._embed_model = OpenAILikeEmbedding(
+            model_name=self.embedding_config.name,
+            api_key=self.embedding_config.api_key,
+            api_base=self.embedding_config.api_base,
+        )
+
+        self._schema = RedisVectorStoreSchema(
+            index=RedisIndexInfo(name=self.index),
+        )
+        if "vector" in self._schema.fields:
+            vector_field = self._schema.fields["vector"]
+            if (
+                vector_field
+                and vector_field.attrs
+                and isinstance(vector_field.attrs, BaseVectorFieldAttributes)
+            ):
+                vector_field.attrs.dims = self.embedding_config.dim
+
+        self._vector_store = RedisVectorStore(
+            schema=self._schema,
+            redis_client=self._redis_client,
+            overwrite=True,
+            collection_name=self.index,
+        )
+
+        self._storage_context = StorageContext.from_defaults(
+            vector_store=self._vector_store
+        )
+
+        self._vector_index = VectorStoreIndex.from_documents(
+            documents=[],
+            storage_context=self._storage_context,
+            embed_model=self._embed_model,
+        )
+
+    @override
+    def precheck_index_naming(self) -> None:
+        # Checking is not needed
+        pass
+
+    @override
+    def add_from_directory(self, directory: str) -> bool:
+        documents = SimpleDirectoryReader(input_dir=directory).load_data()
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def add_from_files(self, files: list[str]) -> bool:
+        documents = SimpleDirectoryReader(input_files=files).load_data()
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def add_from_text(self, text: str | list[str]) -> bool:
+        if isinstance(text, str):
+            documents = [Document(text=text)]
+        else:
+            documents = [Document(text=t) for t in text]
+        nodes = self._split_documents(documents)
+        self._vector_index.insert_nodes(nodes)
+        return True
+
+    @override
+    def search(self, query: str, top_k: int = 5) -> list[str]:
+        _retriever = self._vector_index.as_retriever(similarity_top_k=top_k)
+        retrieved_nodes = _retriever.retrieve(query)
+        return [node.text for node in retrieved_nodes]
+
+    def _split_documents(self, documents: list[Document]) -> list[BaseNode]:
+        """Split document into chunks"""
+        nodes = []
+        for document in documents:
+            splitter = get_llama_index_splitter(document.metadata.get("file_path", ""))
+            _nodes = splitter.get_nodes_from_documents([document])
+            nodes.extend(_nodes)
+        return nodes

veadk/knowledgebase/backends/utils.py

@@ -0,0 +1,91 @@
+# Copyright (c) 2025 Beijing Volcano Engine Technology Co., Ltd. and/or its affiliates.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import json
+from pathlib import Path
+from typing import Literal
+
+from llama_index.core.node_parser import (
+    CodeSplitter,
+    HTMLNodeParser,
+    MarkdownNodeParser,
+    SentenceSplitter,
+)
+from volcengine.auth.SignerV4 import SignerV4
+from volcengine.base.Request import Request
+from volcengine.Credentials import Credentials
+
+
+def get_llama_index_splitter(
+    file_path: str,
+) -> CodeSplitter | MarkdownNodeParser | HTMLNodeParser | SentenceSplitter:
+    suffix = Path(file_path).suffix.lower()
+
+    if suffix in [".py", ".js", ".java", ".cpp"]:
+        return CodeSplitter(language=suffix.strip("."))
+    elif suffix in [".md"]:
+        return MarkdownNodeParser()
+    elif suffix in [".html", ".htm"]:
+        return HTMLNodeParser()
+    else:
+        return SentenceSplitter(chunk_size=512, chunk_overlap=50)
+
+
+def build_vikingdb_knowledgebase_request(
+    path: str,
+    volcengine_access_key: str,
+    volcengine_secret_key: str,
+    method: Literal["GET", "POST", "PUT", "DELETE"] = "POST",
+    region: str = "cn-beijing",
+    params=None,
+    data=None,
+    doseq=0,
+) -> Request:
+    if params:
+        for key in params:
+            if (
+                type(params[key]) is int
+                or type(params[key]) is float
+                or type(params[key]) is bool
+            ):
+                params[key] = str(params[key])
+            elif type(params[key]) is list:
+                if not doseq:
+                    params[key] = ",".join(params[key])
+
+    r = Request()
+    r.set_shema("https")
+    r.set_method(method)
+    r.set_connection_timeout(10)
+    r.set_socket_timeout(10)
+
+    mheaders = {
+        "Accept": "application/json",
+        "Content-Type": "application/json",
+    }
+    r.set_headers(mheaders)
+
+    if params:
+        r.set_query(params)
+
+    r.set_path(path)
+
+    if data is not None:
+        r.set_body(json.dumps(data))
+
+    credentials = Credentials(
+        volcengine_access_key, volcengine_secret_key, "air", region
+    )
+    SignerV4.sign(r, credentials)
+    return r