docs(Eino): update "fan-in and merge" (#1292)

Author: wolfogre

content/en/docs/eino/core_modules/chain_and_graph_orchestration/orchestration_design_principles.md

@@ -334,9 +334,18 @@ This principle also applies to Chunks in the StreamReader.
 
 ### **Fan-In and Merging**
 
-**Fan-in**: Outputs from multiple predecessor nodes converge to a successor node and together serve as the input for the successor node. It is necessary to clearly define how the outputs of multiple predecessor nodes are **merged**. Eino's choice is as follows. First, it requires that the **actual types** of the outputs of multiple predecessor nodes must be the same Map type, and the keys among them must not be repeated. Second:
+**Fan-in**: Outputs from multiple predecessor nodes converge to a successor node and together serve as the input for the successor node. It is necessary to clearly define how the outputs of multiple predecessor nodes are **merged**. Eino's choice is as follows. First, it requires that the **actual types** of the outputs of multiple predecessor nodes must be the same and of a mergeable type.
 
-- In non-streaming scenarios, after merging, it becomes a single map that contains all key-value pairs from all upstream sources.
+Mergeable types can be:
+
+- Map type, and the keys are not repeated among each other.
+- Any type, and a merge function is registered through `compose.RegisterValuesMergeFunc`.
+
+Second,
+
+- In non-streaming scenarios, 
+  - if the input type has a registered merge function, it will be merged into a single value using the registered function.
+  - if the input type is a Map, it will be merged into a single Map containing all key-value pairs from all upstream sources.
 - In streaming scenarios, multiple upstream StreamReaders of the same type are merged into one StreamReader. When actually receiving data from the merged StreamReader, the effect is to read fairly from multiple upstream StreamReaders.
 
 When adding a node (AddNode), you can add the WithOutputKey option to convert the output of the node into a Map:

content/zh/docs/eino/core_modules/chain_and_graph_orchestration/orchestration_design_principles.md

@@ -340,10 +340,19 @@ Eino 的 Graph 中的数据在 Node、Branch、Handler 间流转时，一律是
 
 ### 扇入与合并
 
-**扇入**：多个上游的数据汇入到下游，一起作为下游的输入。需要明确定义多个上游的输出，如何**合并（Merge）**起来。Eino 的选择是，首先要求多个上游输出的**实际类型**必须相同且为 Map，且相互间 key 不可重复。其次：
+**扇入**：多个上游的数据汇入到下游，一起作为下游的输入。需要明确定义多个上游的输出，如何**合并（Merge）**起来。Eino 的选择是，首先要求多个上游输出的**实际类型**必须相同且为可合并的类型。
 
-- 在非流式场景下，合并后成为一个 Map，包含所有上游的所有键值对。
-- 在流式场景下，将类型相同的多个上游 StreamReader 合并为一个 StreamReader。实际 Recv 时效果为从多个上游 StreamReader 中公平读取。
+可合并的类型可以是：
+
+- Map 类型，且相互间 key 不重复。
+- 任意类型，且通过 `compose.RegisterValuesMergeFunc` 注册了合并函数。
+
+其次：
+
+- 在非流式场景下，
+  - 如果输入的类型有注册过合并函数，则使用注册的合并函数合并为一个值。
+  - 如果输入的类型是 Map，则合并成为一个 Map，包含所有上游的所有键值对。
+- 在流式场景下，将类型相同的多个上游 StreamReader 合并为一个 StreamReader。实际 Recv 时从效果为从多个上游 StreamReader 中公平读取。
 
 在 AddNode 时，可以通过添加 WithOutputKey 这个 Option 来把节点的输出转成 Map：