已弃用 APIDeprecated API
带有 @Deprecated 注解或 @deprecated Javadoc 标签的类型与成员。Types and members carrying @Deprecated or a @deprecated Javadoc tag.
| API | 类型Kind | 弃用说明Deprecation |
|---|---|---|
io.agentscope.aistio.proto.ConfigType.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.aistio.proto.Downstream.PayloadCase.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.aistio.proto.Upstream.PayloadCase.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.builder.runtime.config.AgentConfigEntry.getSkillRepository | 方法method | 获取技能仓库。 Returns the skill repository. |
io.agentscope.builder.runtime.config.AgentConfigEntry.setSkillRepository | 方法method | 设置技能仓库。 Sets the skill repository. |
io.agentscope.builder.web.coord.TurnLeaseService.acquireOrConflict | 方法method | 执行 `acquireOrConflict` 方法。 use `#acquireOrConflict(String, String, Runnable)` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_CREATED | 字段field | 状态已创建字段。 use `SessionStatuses#CREATED` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_RUNNING | 字段field | 状态运行中字段。 use `SessionStatuses#RUNNING` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_IDLE | 字段field | 状态 IDLE字段。 use `SessionStatuses#IDLE` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_REQUIRES_ACTION | 字段field | 状态 REQUIRES 动作字段。 use `SessionStatuses#REQUIRES_ACTION` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_TERMINATED | 字段field | 状态 TERMINATED字段。 use `SessionStatuses#TERMINATED` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_RESCHEDULED | 字段field | 状态 RESCHEDULED字段。 use `SessionStatuses#RESCHEDULED` |
io.agentscope.builder.web.managed.DataSessionService.STATUS_ARCHIVED | 字段field | 状态 ARCHIVED字段。 use `SessionStatuses#ARCHIVED` |
io.agentscope.builder.web.managed.ManagedSessionChannelBridge.dispatchAndAwaitReply | 方法method | 执行 `dispatchAndAwaitReply` 方法。 use `#dispatchAndAwaitReply(String, String, String, String)` |
io.agentscope.builder.web.managed.MemoryMountService.materialize | 方法method | 执行 `materialize` 方法。 host-disk materialization is no longer used; memory stores are mounted live via
`#createFilesystems`. Retained only so callers mid-migration keep compiling;
delegates to `#resolveMounts`. |
io.agentscope.builder.web.managed.MemoryMountService.writeback | 方法method | 执行 `writeback` 方法。 writeback is a no-op: `MemoryStoreFilesystem` writes directly through to
`MemoryStoreService` on every tool call, so there is nothing left to reconcile
after a turn. |
io.agentscope.controlplane.proto.ConfigType.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.controlplane.proto.Downstream.PayloadCase.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.controlplane.proto.Upstream.PayloadCase.valueOf | 方法method | 设置值的。 Use `#forNumber(int)` instead. |
io.agentscope.core.ReActAgent.interrupt | 方法method | 以协作方式中断当前运行。 Use `#interrupt(String, String)` with explicit userId/sessionId. |
io.agentscope.core.ReActAgent.interrupt | 方法method | 以协作方式中断当前运行。 Use `#interrupt(String, String, Msg)` with explicit userId/sessionId. |
io.agentscope.core.ReActAgent.interrupt | 方法method | 以协作方式中断当前运行。 Use `#interrupt(String, String)` with explicit userId/sessionId. |
io.agentscope.core.ReActAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` (and overloads that
accept `RuntimeContext` via `call()`-driven lifecycle) for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.ReActAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.ReActAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.ReActAgent.getAgentState | 方法method | 获取智能体状态。 Use `#getAgentState(RuntimeContext)` or
`#getAgentState(String, String)` with explicit session identity.
This method delegates to the default session slot. |
io.agentscope.core.ReActAgent.getPermissionEngine | 方法method | 获取权限引擎。 Use `getAgentState(userId, sessionId).getPermissionContext()` instead. |
io.agentscope.core.ReActAgent.getPermissionContext | 方法method | 获取权限上下文。 Use `getAgentState(userId, sessionId).getPermissionContext()` instead. |
io.agentscope.core.ReActAgent.CONFIRM_SINK_KEY | 字段field | 确认 SINK 键字段。 Permission HITL no longer uses a Reactor Sink. Confirm results are now
delivered via a second `agent.call(msgs)` carrying a `ConfirmResult`
payload — see `applyConfirmResults`. This constant is retained as a
compile-time marker and will be removed in a future release. |
io.agentscope.core.ReActAgent.Builder.checkRunning | 方法method | 设置检查运行中。 No longer enforced; per-session serialization handles concurrency. |
io.agentscope.core.ReActAgent.Builder.toolExecutionContext | 方法method | 设置工具执行上下文用于此智能体。 Use `RuntimeContext` with `agent.call(msg, runtimeContext)`
or register POJOs directly via `RuntimeContext.builder().put(Type, value)`. |
io.agentscope.core.ReActAgent.Builder.longTermMemory | 方法method | 设置该配置项。 since 2.0.0. Long-term memory is being redesigned around the upcoming reme
base class. Hooks added through this path still work. |
io.agentscope.core.ReActAgent.Builder.longTermMemoryMode | 方法method | 设置该配置项。 since 2.0.0. See `#longTermMemory`. |
io.agentscope.core.ReActAgent.Builder.longTermMemoryAsyncRecord | 方法method | 设置该配置项。 since 2.0.0. See `#longTermMemory`. |
io.agentscope.core.ReActAgent.Builder.knowledge | 方法method | 挂载一个知识库用于 RAG。 since 2.0.0. RAG is being redesigned; legacy adapters remain functional. |
io.agentscope.core.ReActAgent.Builder.knowledges | 方法method | 设置该配置项。 since 2.0.0. See `#knowledge`. |
io.agentscope.core.ReActAgent.Builder.ragMode | 方法method | 选择知识库的注入模式。 since 2.0.0. See `#knowledge`. |
io.agentscope.core.ReActAgent.Builder.retrieveConfig | 方法method | 设置检索配置。 since 2.0.0. See `#knowledge`. |
io.agentscope.core.ReActAgent.Builder.skillBox | 方法method | 设置该配置项。 since 2.0.0. Skills now flow through `#skillRepository` /
`#skillRepositories`; legacy `io.agentscope.core.skill.SkillBox`
instances are still accepted for source compatibility, but combining a
`skillBox(...)` with `skillRepository(...)` is untested — new code
should prefer `#skillRepository(AgentSkillRepository)`. |
io.agentscope.core.agent.AgentBase.AgentBase | 构造器constructor | 创建智能体基类实例。 Use `#AgentBase(String, String, List)` instead. |
io.agentscope.core.agent.AgentBase.isCheckRunning | 方法method | 判断是否检查运行中。 No longer enforced; per-session serialization handles concurrency. |
io.agentscope.core.agent.AgentBase.interrupt | 方法method | 请求中断当前执行。 Subclasses should implement per-session interrupt via RuntimeContext. |
io.agentscope.core.agent.AgentBase.interrupt | 方法method | 请求中断当前执行。 Subclasses should implement per-session interrupt via RuntimeContext. |
io.agentscope.core.agent.AgentBase.interrupt | 方法method | 请求中断当前执行。 Subclasses should implement per-session interrupt via RuntimeContext. |
io.agentscope.core.agent.AgentBase.checkInterruptedAsync | 方法method | 检查 `checkInterruptedAsync` 条件。 No longer needed; ReActAgent uses per-execution InterruptControl. |
io.agentscope.core.agent.AgentBase.resetInterruptFlag | 方法method | 重置中断开关。 No-op; interrupt state is owned by the execution. |
io.agentscope.core.agent.AgentBase.getInterruptFlag | 方法method | 获取中断开关。 No-op stub. |
io.agentscope.core.agent.AgentBase.getInterruptSource | 方法method | 获取中断来源。 Returns USER. Interrupt source is owned by the execution. |
io.agentscope.core.agent.AgentBase.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(List)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.AgentBase.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.AgentBase.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.Event | 类class | 事件,描述执行过程中的事件。 since 2.0.0. This is the v1 coarse-grained event type, kept as the return type
of the deprecated `stream(...)` API and as the wire format consumed by the
harness, AGUI, A2A, chat-completions-web, and Kotlin extension modules. New code should
use `io.agentscope.core.event.AgentEvent` via `ReActAgent#streamEvents(...)`,
which exposes the fine-grained event hierarchy (28 typed events covering the full agent
lifecycle and HITL flow). The class will be marked `forRemoval = true` once the
dependent internal modules migrate to `AgentEvent`. |
io.agentscope.core.agent.EventSource | 类class | 事件来源,驱动智能体的推理、工具调用与会话执行。 since 2.0.0. Used only by the v1 `Event` stream to attribute events to a
spawned subagent. A subagent-source channel for the new
`io.agentscope.core.event.AgentEvent` stream is on the v2 roadmap; once it lands,
`EventSource` will be marked `forRemoval = true`. |
io.agentscope.core.agent.EventType | 枚举enum | 定义事件类型的枚举。 since 2.0.0. The v1 coarse-grained event discriminator. New code should use
`io.agentscope.core.event.AgentEventType` (and the typed `AgentEvent`
subclasses) exposed by `ReActAgent#streamEvents(...)`. |
io.agentscope.core.agent.StreamableAgent | 接口interface | 定义 Streamable 智能体契约的接口。 since 2.0.0, for removal in a future minor release. Every `stream(...)`
method on this interface returns the coarse-grained `Event` type, which is a v1
carry-over. Use `ReActAgent#streamEvents(...)` instead — it returns
`Flux<io.agentscope.core.event.AgentEvent>`, the fine-grained event hierarchy that
covers the full agent lifecycle (28 typed events including HITL) and is the single
streaming surface going forward. See the v2 changelog section B.4 for migration details. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 设置流。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 设置流。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 设置流。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(Msg)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(Msg)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 设置流。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(List)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(List)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamableAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `ReActAgent#streamEvents(...)` for the
fine-grained `AgentEvent` stream. |
io.agentscope.core.agent.StreamingHook | 类class | 流式钩子,驱动智能体的推理、工具调用与会话执行。 since 2.0.0. Replaced by the `io.agentscope.core.middleware.MiddlewareBase`
system. Use `io.agentscope.core.ReActAgent#streamEvents(java.util.List)` for
fine-grained event streaming. |
io.agentscope.core.agent.user.StreamUserInput | 类class | 流用户输入,驱动智能体的推理、工具调用与会话执行。 Stream-based user input implementation that reads from any InputStream and writes to any
OutputStream. |
io.agentscope.core.agent.user.UserAgent | 类class | 用户智能体,驱动智能体的推理、工具调用与会话执行。 UserAgent class for handling user interaction within the agent framework. |
io.agentscope.core.agent.user.UserInputBase | 接口interface | 定义用户输入基类契约的接口。 Strategy interface for handling user input from different sources. |
io.agentscope.core.agent.user.UserInputData | 类class | 用户输入数据,驱动智能体的推理、工具调用与会话执行。 Data class that holds user input information with dual representation. |
io.agentscope.core.hook.ActingChunkEvent | 类class | 执行分块事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.ActingEvent | 类class | 执行事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.ErrorEvent | 类class | 错误事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.Hook | 接口interface | 定义钩子契约的接口。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead.
Middleware exposes `onAgent`, `onReasoning`, `onActing`,
`onModelCall`, and `onSystemPrompt` hook points and replaces this entire
hook system. |
io.agentscope.core.hook.HookEvent | 类class | 所有钩子事件的基类。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.HookEventType | 枚举enum | 定义钩子事件类型的枚举。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.LegacyHookDispatcher | 类class | Legacy 钩子调度器,驱动智能体的推理、工具调用与会话执行。 since 2.0.0, for removal. Will be deleted together with the Hook system. |
io.agentscope.core.hook.PostActingEvent | 类class | 后置执行事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PostCallEvent | 类class | 后置调用事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PostReasoningEvent | 类class | 后置推理事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PostSummaryEvent | 类class | 后置摘要事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PreActingEvent | 类class | 前置执行事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PreCallEvent | 类class | 前置调用事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PreReasoningEvent | 类class | 前置推理事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.PreSummaryEvent | 类class | 前置摘要事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.ReasoningChunkEvent | 类class | 推理分块事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.ReasoningEvent | 类class | 推理事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.SummaryChunkEvent | 类class | 摘要分块事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.SummaryEvent | 类class | 摘要事件,描述执行过程中的事件。 since 2.0.0. Use `io.agentscope.core.middleware.MiddlewareBase` instead. |
io.agentscope.core.hook.recorder.JsonlTraceExporter | 类class | Jsonl 追踪导出器,驱动智能体的推理、工具调用与会话执行。 since 2.0.0. The hook system is replaced by
`io.agentscope.core.middleware.MiddlewareBase`. |
io.agentscope.core.memory.AgentStateMemoryView | 类class | 智能体状态记忆 View,在多次调用之间保存和恢复会话与记忆数据。 since 2.0.0. New code should read context directly from
`AgentState#getContext()`. |
io.agentscope.core.memory.InMemoryMemory | 类class | In 记忆记忆,在多次调用之间保存和恢复会话与记忆数据。 since 2.0.0. Conversation context lives on
`io.agentscope.core.state.AgentState#getContext()`; this class is retained for
source compatibility with 1.0.x user code. |
io.agentscope.core.memory.LongTermMemory | 接口interface | 定义 Long Term 记忆契约的接口。 since 2.0.0. Long-term memory is removed; conversation context now lives on
`io.agentscope.core.state.AgentState#getContext()`. Integrate any cross-session
persistence at the application layer. |
io.agentscope.core.memory.LongTermMemoryMode | 枚举enum | 定义Long Term 记忆模式相关取值的枚举。 since 2.0.0. Long-term memory is removed. |
io.agentscope.core.memory.LongTermMemoryTools | 类class | Long Term 记忆工具,封装可被智能体调用的能力。 since 2.0.0. Long-term memory is removed. |
io.agentscope.core.memory.Memory | 接口interface | 定义记忆契约的接口。 since 2.0.0. Conversation context is now held on
`io.agentscope.core.state.AgentState#getContext()`. This interface is retained as a
write-only mirror for source compatibility with 1.0.x user code. |
io.agentscope.core.memory.StateBackedMemory | 类class | 状态 Backed 记忆,在多次调用之间保存和恢复会话与记忆数据。 since 2.0.0. Use `AgentState#getContext()` directly. |
io.agentscope.core.memory.StaticLongTermMemoryHook | 类class | Static Long Term 记忆钩子,在多次调用之间保存和恢复会话与记忆数据。 since 2.0.0. Long-term memory and the `Hook` system are removed; reimplement
as a `io.agentscope.core.middleware.MiddlewareBase` if required. |
io.agentscope.core.memory.mem0.Mem0AddRequest.getMemoryType | 方法method | 获取记忆类型。 Not in official API spec |
io.agentscope.core.memory.mem0.Mem0AddRequest.setMemoryType | 方法method | 设置记忆类型。 Not in official API spec |
io.agentscope.core.memory.mem0.Mem0AddRequest.Builder.memoryType | 方法method | 设置记忆类型。 Not in official API spec |
io.agentscope.core.memory.mem0.Mem0SearchRequest.Builder.limit | 方法method | 设置限制。 Use `#topK(Integer)` instead |
io.agentscope.core.rag.GenericRAGHook | 类class | 泛型 RAG 钩子,把私有文档转化为可检索的知识。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.Knowledge | 接口interface | 定义知识库契约的接口。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.KnowledgeRetrievalTools | 类class | 知识库检索工具,封装可被智能体调用的能力。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.RAGMode | 枚举enum | 定义RAG 模式相关取值的枚举。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.integration.ragflow.RAGFlowConfig.Builder.rerankId | 方法method | 设置重排标识。 RAGFlow identifies rerank models by name, not by a numeric id, so no
`Integer` maps to a real model and the retrieval request is rejected. Use
`#rerankId(String)` instead. This overload only keeps existing call sites
compiling and stringifies the value. |
io.agentscope.core.rag.model.Document | 类class | 文档,与模型服务通信并返回生成或向量化结果。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.model.DocumentMetadata | 类class | 文档元数据,与模型服务通信并返回生成或向量化结果。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.model.RetrieveConfig | 类class | 检索配置,保存相关配置项。 since 2.0.0. The rag package is removed; integrate retrieval at the application
layer. |
io.agentscope.core.rag.store.MilvusStore.create | 方法method | 创建并返回实例。 Use `MilvusStore#builder()` instead |
io.agentscope.core.session.mysql.MysqlSession.clearAllSessions | 方法method | 清空所有会话。 Use `#truncateAllSessions()` instead |
io.agentscope.core.session.redis.jedis.JedisSession | 类class | Jedis 会话,在多次调用之间保存和恢复会话与记忆数据。 Use `io.agentscope.core.session.redis.RedisSession` with jedisClient instead.
RedisSession provides a unified session implementation that works with multiple Redis clients.
For basic usage, simple constructors are available.
For advanced configuration, use `RedisSession#builder()`.
This implementation stores session state in Redis with the following key structure:
- Single state: `{prefix}{sessionId}:{stateKey}` - Redis String containing JSON
- List state: `{prefix}{sessionId}:{stateKey}:list` - Redis List containing JSON items
- Session marker: `{prefix}{sessionId}:_keys` - Redis Set tracking all state keys
- Incremental list storage (only appends new items)
- Type-safe state serialization using Jackson
- Automatic session key tracking |
io.agentscope.core.session.redis.redisson.RedissonSession | 类class | Redisson 会话,在多次调用之间保存和恢复会话与记忆数据。 Use `io.agentscope.core.session.redis.RedisSession` with redissonClient instead.
RedisSession provides a unified session implementation that works with multiple Redis clients.
For basic usage, simple constructors are available.
For advanced configuration, use `RedisSession#builder()`
This implementation stores session state in Redis with the following key structure:
- Single state: `{prefix}{sessionId}:{stateKey}` - Redis String containing JSON
- List state: `{prefix}{sessionId}:{stateKey}:list` - Redis List containing JSON items
- Session marker: `{prefix}{sessionId}:_keys` - Redis Set tracking all state keys
- Incremental list storage (only appends new items)
- Type-safe state serialization using Jackson
- Automatic session key tracking |
io.agentscope.core.shutdown.ShutdownSessionBinding | 记录record | 承载 Shutdown 会话绑定数据的记录类型。 since 2.0.0. Use `ShutdownStateSaver` with
`GracefulShutdownManager#bindStateSaver` instead. |
io.agentscope.core.skill.SkillHook | 类class | 技能钩子,让智能体以受控方式调用外部能力。 since 2.0.0. The skill package and the hook system are removed; manage markdown
skill catalogs in application code or via `MiddlewareBase`. |
io.agentscope.core.tool.ContextStore | 接口interface | 定义上下文存储契约的接口。 Part of the `ToolExecutionContext` layer which is superseded by
`io.agentscope.core.agent.RuntimeContext`. |
io.agentscope.core.tool.ToolCallParam.getContext | 方法method | 获取工具执行上下文。 Use `#getRuntimeContext()` instead. |
io.agentscope.core.tool.ToolCallParam.Builder.context | 方法method | 设置工具执行上下文。 Use `#runtimeContext(RuntimeContext)` instead. |
io.agentscope.core.tool.ToolExecutionContext | 类class | 工具执行上下文,让智能体以受控方式调用外部能力。 Use `io.agentscope.core.agent.RuntimeContext` instead.
Register POJOs via `RuntimeContext.builder().put(Type.class, value).build()`
and pass to `agent.call(msg, runtimeContext)`. |
io.agentscope.core.tool.ToolExecutionContextProvider | 接口interface | 定义工具执行上下文提供者契约的接口。 Part of the `ToolExecutionContext` layer which is superseded by
`io.agentscope.core.agent.RuntimeContext`. |
io.agentscope.core.tool.ToolkitConfig.getDefaultContext | 方法method | 获取默认工具执行上下文。 Use `io.agentscope.core.agent.RuntimeContext` instead. |
io.agentscope.core.tool.ToolkitConfig.Builder.defaultContext | 方法method | 设置默认工具执行上下文用于所有工具。 Use `io.agentscope.core.agent.RuntimeContext` instead. |
io.agentscope.core.tool.mcp.McpTool.McpTool | 构造器constructor | 初始化 `McpTool` 实例。 Use `#McpTool(String, String, Map, Map, McpClientWrapper, Map, String,
boolean)` so the MCP server name and read-only hint are propagated. Kept for source
compatibility with callers that pre-date the `ToolBase` integration. |
io.agentscope.core.tool.mcp.McpTool.McpTool | 构造器constructor | 初始化 `McpTool` 实例。 Use `#McpTool(String, String, Map, Map, McpClientWrapper, Map, String,
boolean)` so the MCP server name and read-only hint are propagated. |
io.agentscope.core.tool.mcp.McpTool.McpTool | 构造器constructor | 初始化 `McpTool` 实例。 Use `#McpTool(String, String, Map, Map, McpClientWrapper, Map, String,
boolean)` so the MCP server name and read-only hint are propagated. |
io.agentscope.core.tracing.NoopTracer | 类class | Noop 追踪器,对智能体行为进行评测、追踪与可观测性分析。 since 2.0.0. Use `OtelTracingMiddleware` instead. |
io.agentscope.core.tracing.Tracer | 接口interface | 定义追踪器契约的接口。 since 2.0.0. Use `OtelTracingMiddleware` instead. |
io.agentscope.core.tracing.TracerRegistry | 类class | 追踪器注册表,对智能体行为进行评测、追踪与可观测性分析。 since 2.0.0. Use `OtelTracingMiddleware` instead. |
io.agentscope.core.training.runner.TrainingConfig.Builder.sampleRate | 方法method | 设置该配置项。 Recommend using `#selectionStrategy(TrainingSelectionStrategy)` |
io.agentscope.extensions.channel.dingtalk.DingTalkChannelProperties.DingTalkChannelProperties | 构造器constructor | 初始化 `DingTalkChannelProperties` 实例。 use the canonical constructor (or `#from`) so the reception mode is
explicit |
io.agentscope.extensions.model.openai.formatter.DeepSeekFormatter | 类class | DeepSeek 格式化器,与模型服务通信并返回生成或向量化结果。 use `io.agentscope.extensions.model.openai.compat.deepseek.DeepSeekFormatter`. |
io.agentscope.extensions.model.openai.formatter.DeepSeekMultiAgentFormatter | 类class | DeepSeek Multi 智能体格式化器,与模型服务通信并返回生成或向量化结果。 use `io.agentscope.extensions.model.openai.compat.deepseek.DeepSeekMultiAgentFormatter`. |
io.agentscope.extensions.model.openai.formatter.GLMFormatter | 类class | GLM 格式化器,与模型服务通信并返回生成或向量化结果。 use `io.agentscope.extensions.model.openai.compat.glm.GLMFormatter`. |
io.agentscope.extensions.model.openai.formatter.GLMMultiAgentFormatter | 类class | GLM Multi 智能体格式化器,与模型服务通信并返回生成或向量化结果。 use `io.agentscope.extensions.model.openai.compat.glm.GLMMultiAgentFormatter`. |
io.agentscope.extensions.mysql.MysqlDistributedStore | 类class | MySQL 分布式存储,负责持久化与读取数据。 Use `io.agentscope.extensions.jdbc.JdbcDistributedStore` from the
`agentscope-extensions-jdbc` module instead. The new module provides a unified
multi-database dialect abstraction that supports MySQL, PostgreSQL, H2, and SQLite
with a single aggregated interface. This class is preserved unchanged for backward
compatibility and will be removed in a future major release. |
io.agentscope.extensions.mysql.state.MysqlAgentStateStore.clearAllSessions | 方法method | 清空所有会话。 Use `#truncateAllSessions()` instead |
io.agentscope.extensions.postgresql.PostgresDistributedStore | 类class | Postgres 分布式存储,负责持久化与读取数据。 Use `io.agentscope.extensions.jdbc.JdbcDistributedStore` from the
`agentscope-extensions-jdbc` module instead. The new module provides a unified
multi-database dialect abstraction that supports MySQL, PostgreSQL, H2, and SQLite
with a single aggregated interface. This class is preserved unchanged for backward
compatibility and will be removed in a future major release. |
io.agentscope.extensions.redis.state.jedis.JedisAgentStateStore | 类class | Jedis 智能体状态存储,负责持久化与读取数据。 Use `io.agentscope.extensions.redis.state.RedisAgentStateStore` with jedisClient instead.
RedisAgentStateStore provides a unified session implementation that works with multiple Redis clients.
For basic usage, simple constructors are available.
For advanced configuration, use `RedisAgentStateStore#builder()`.
This implementation stores session state in Redis with the following key structure:
- Single state: `{prefix}{sessionId}:{stateKey}` - Redis String containing JSON
- List state: `{prefix}{sessionId}:{stateKey}:list` - Redis List containing JSON items
- AgentStateStore marker: `{prefix}{sessionId}:_keys` - Redis Set tracking all state keys
- Incremental list storage (only appends new items)
- Type-safe state serialization using Jackson
- Automatic session key tracking |
io.agentscope.extensions.redis.state.redisson.RedissonAgentStateStore | 类class | Redisson 智能体状态存储,负责持久化与读取数据。 Use `io.agentscope.extensions.redis.state.RedisAgentStateStore` with redissonClient instead.
RedisAgentStateStore provides a unified session implementation that works with multiple Redis clients.
For basic usage, simple constructors are available.
For advanced configuration, use `RedisAgentStateStore#builder()`
This implementation stores session state in Redis with the following key structure:
- Single state: `{prefix}{sessionId}:{stateKey}` - Redis String containing JSON
- List state: `{prefix}{sessionId}:{stateKey}:list` - Redis List containing JSON items
- AgentStateStore marker: `{prefix}{sessionId}:_keys` - Redis Set tracking all state keys
- Incremental list storage (only appends new items)
- Type-safe state serialization using Jackson
- Automatic session key tracking |
io.agentscope.harness.agent.HarnessAgent.getAgentState | 方法method | 获取智能体状态。 Use `#getDelegate()``.getAgentState(RuntimeContext)` or
`.getAgentState(String, String)` with explicit session identity. |
io.agentscope.harness.agent.HarnessAgent.call | 方法method | 向智能体发送输入并返回最终回复。 Use `#call(List, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.HarnessAgent.call | 方法method | 向智能体发送输入并返回最终回复。 Use `#call(List, Class, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.HarnessAgent.call | 方法method | 向智能体发送输入并返回最终回复。 Use `#call(List, JsonNode, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` for the fine-grained
`AgentEvent` stream that aligns with Python 2.0's `agent.reply_stream()`. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` for the fine-grained
`AgentEvent` stream. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List)` for the fine-grained
`AgentEvent` stream. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(Msg, RuntimeContext)`. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List, RuntimeContext)`. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List, RuntimeContext)`. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List, RuntimeContext)`. |
io.agentscope.harness.agent.HarnessAgent.stream | 方法method | 执行流相关操作。 since 2.0.0, for removal. Use `#streamEvents(List, RuntimeContext)`. |
io.agentscope.harness.agent.HarnessAgent.streamEvents | 方法method | 执行事件。 Use `#streamEvents(Msg, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.HarnessAgent.streamEvents | 方法method | 执行事件。 Use `#streamEvents(List, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.HarnessAgent.streamEvents | 方法method | 执行事件。 Use `#streamEvents(String, RuntimeContext)` with explicit runtime context. |
io.agentscope.harness.agent.filesystem.ProjectAwareOverlay.ProjectAwareOverlay | 构造器constructor | 初始化 `ProjectAwareOverlay` 实例。 use `#ProjectAwareOverlay(AbstractSandboxFilesystem, AbstractFilesystem,
LocalFilesystem, Path, NamespaceFactory)` so workspace-rooted absolute paths resolve
identically to their relative spelling under namespace scoping |
io.agentscope.harness.agent.filesystem.spec.LocalFilesystemSpec.virtualMode | 方法method | 设置该参数。 use `#mode(LocalFsMode)` for the full three-way selection |
io.agentscope.harness.agent.gateway.HarnessGateway.runWakeup | 方法method | 运行 `runWakeup`。 use `#runWakeup(String, String)` so the wakeup-driven round loads agent
state from the correct `(userId, sessionId)` slot. This overload preserves legacy
anonymous-wakeup behavior. |
io.agentscope.harness.agent.gateway.WakeupDispatcher.WakeupTarget.runWakeup | 方法method | 运行 `runWakeup`。 implementers should override `#runWakeup(String, String)` instead, so
the dispatcher can propagate the owning user id. Required for runtimes that isolate
agent state by `(userId, sessionId)`: without it, a wakeup-driven round loads
state from an anonymous slot and the original conversation is lost. This overload is
retained for backward compatibility and defaults to ignoring the user id. |
io.agentscope.harness.agent.memory.MemoryFlushManager.resolveOffloadPath | 方法method | 解析卸载路径。 Prefer `SessionTranscriptWriter#resolveContextPath`. |
io.agentscope.harness.agent.memory.MemoryFlushManager.offloadMessages | 方法method | 执行卸载消息相关操作。 Prefer `SessionTranscriptWriter#appendMessages`. Kept as a thin
delegate so compaction / overflow-recovery call sites keep compiling during the
migration; new code should use `SessionTranscriptWriter` directly. |
io.agentscope.harness.agent.skill.runtime.SkillLoadTool.SkillLoadTool | 构造器constructor | 初始化 `SkillLoadTool` 实例。 Prefer `#SkillLoadTool()` and bind catalogs through `SkillRuntime#install(SkillCatalog, io.agentscope.core.agent.RuntimeContext,
io.agentscope.core.tool.Toolkit)`. |
io.agentscope.harness.agent.skill.runtime.SkillRuntime.currentCatalog | 方法method | 返回当前目录。 Use `#currentCatalog(RuntimeContext)` so catalog lookup is request-scoped. |
io.agentscope.harness.agent.skill.runtime.SkillRuntime.install | 方法method | 执行 `install` 方法。 Use `#install(SkillCatalog, RuntimeContext, Toolkit)`. |
io.agentscope.harness.agent.subagent.SubagentDeclaration.getMaxIters | 方法method | 获取最大迭代次数。 since Phase A — use `#getSteps()`. Returns the same value; kept for source
compatibility with callers built before the `steps` field existed. |
io.agentscope.harness.agent.subagent.SubagentDeclaration.Builder.maxIters | 方法method | 设置最大迭代次数。 since Phase A — use `#steps(int)`. Equivalent in behaviour. |
io.agentscope.harness.agent.workspace.WorkspaceManager.resolveSessionFile | 方法method | 解析会话文件。 Use `#resolveSessionContextFile(RuntimeContext, String, String)` for the
JSONL format. |
io.agentscope.harness.agent.workspace.WorkspacePathNormalizer.normalize | 方法method | 设置该参数。 Normalize a path to workspace-relative form by stripping the active mode's prefix. |