chore: import upstream snapshot with attribution

2026-07-13 13:25:13 +08:00
commit ec2b666284
2231 changed files with 491535 additions and 0 deletions
@@ -0,0 +1,43 @@
+# ADK Workflow Routing Sample
+
+## Overview
+
+This sample demonstrates how to use routing in **ADK Workflows**.
+
+It takes user input and uses an LLM node to categorize it as a **question**, a **statement**, or **other**. Based on the classification, it appropriately routes the execution to a specialized agent or function to handle that specific type of input.
+
+In ADK Workflows, **routing** allows conditionally executing different execution paths based on the output of a previous node.
+
+## Sample Inputs
+
+- `What is the capital of France?`
+
+- `The weather is very nice today.`
+
+- `Translate bonjour to english`
+
+## Graph
+
+```mermaid
+graph TD
+    START --> process_input
+    process_input --> classify_input
+    classify_input --> route_on_category
+    route_on_category -->|question| answer_question
+    route_on_category -->|statement| comment_on_statement
+    route_on_category -->|other| handle_other
+```
+
+## How To
+
+1. A node (agent or function) yields an `Event` with a specific route name:
+
+   ```python
+   yield Event(route="your_route_name")
+   ```
+
+1. In the `Workflow` edges definition, conditional edges are constructed using a routing map dict as the second element of the edge tuple:
+
+   ```python
+   (source_node, {"your_route_name": target_node})
+   ```