enhance tool input parsing and add agentic rag tutorial (opensearch-project#4023)

* enhance tool input parsing and add agentic rag tutorial

Signed-off-by: Yaliang Wu <ylwu@amazon.com>

* remove tutorial; add unit test

Signed-off-by: Yaliang Wu <ylwu@amazon.com>

* fix unit test

Signed-off-by: Yaliang Wu <ylwu@amazon.com>

---------

Signed-off-by: Yaliang Wu <ylwu@amazon.com>

Author: ylwu-amzn

ml-algorithms/src/main/java/org/opensearch/ml/engine/algorithms/agent/AgentUtils.java

@@ -932,7 +932,14 @@ public static Map<String, String> constructToolParams(
                 toolParams.putAll(params);
             }
         } else {
-            toolParams.put("input", actionInput);
+            if (toolParams.containsKey("input")) {
+                String input = toolParams.get("input");
+                StringSubstitutor substitutor = new StringSubstitutor(toolParams, "${parameters.", "}");
+                input = substitutor.replace(input);
+                toolParams.put("input", input);
+            } else {
+                toolParams.put("input", actionInput);
+            }
         }
         return toolParams;
     }

ml-algorithms/src/test/java/org/opensearch/ml/engine/algorithms/agent/AgentUtilsTest.java

@@ -76,6 +76,7 @@
 import org.opensearch.ml.common.output.model.ModelTensorOutput;
 import org.opensearch.ml.common.output.model.ModelTensors;
 import org.opensearch.ml.common.spi.tools.Tool;
+import org.opensearch.ml.common.utils.StringUtils;
 import org.opensearch.ml.engine.MLEngineClassLoader;
 import org.opensearch.ml.engine.MLStaticMockBase;
 import org.opensearch.ml.engine.algorithms.remote.McpConnectorExecutor;
@@ -1665,4 +1666,105 @@ public void testCreateTool_ToolNotFound() {
 
         assertThrows(IllegalArgumentException.class, () -> AgentUtils.createTool(toolFactories, new HashMap<>(), toolSpec, "test_tenant"));
     }
+
+    @Test
+    public void testConstructToolParams_ToolInputSubstitution() {
+        String question = "What is the population?";
+        String actionInput = "{\"question\": \"Seattle 2025 population\"}";
+
+        Map<String, Object> queryObj = new HashMap<>();
+        queryObj
+            .put(
+                "query",
+                Map
+                    .of(
+                        "neural",
+                        Map
+                            .of(
+                                "population_description_embedding",
+                                Map.of("query_text", "${parameters.question}", "model_id", "embedding_model_id")
+                            )
+                    )
+            );
+        queryObj.put("size", 2);
+        queryObj.put("_source", "population_description");
+
+        Map<String, Tool> tools = Map.of("SearchIndexTool", tool1);
+        Map<String, MLToolSpec> toolSpecMap = Map
+            .of(
+                "SearchIndexTool",
+                MLToolSpec
+                    .builder()
+                    .type("SearchIndexTool")
+                    .parameters(
+                        Map
+                            .of(
+                                "index",
+                                "test_population_data",
+                                "input",
+                                "{\"index\": \"${parameters.index}\", \"query\": " + StringUtils.toJson(queryObj) + "}"
+                            )
+                    )
+                    .build()
+            );
+
+        AtomicReference<String> lastActionInput = new AtomicReference<>();
+        String action = "SearchIndexTool";
+
+        // Execute
+        Map<String, String> toolParams = AgentUtils.constructToolParams(tools, toolSpecMap, question, lastActionInput, action, actionInput);
+
+        // Verify
+        Assert.assertTrue(toolParams.get("input").contains("\"query_text\":\"Seattle 2025 population\""));
+        assertEquals("Seattle 2025 population", toolParams.get("question"));
+    }
+
+    @Test
+    public void testConstructToolParams_NoInputKey() {
+        // Setup
+        String question = "What is the population?";
+        String actionInput = "{\"question\": \"Seattle 2025 population\"}";
+
+        Map<String, Tool> tools = Map.of("SearchTool", tool1);
+        Map<String, MLToolSpec> toolSpecMap = Map
+            .of("SearchTool", MLToolSpec.builder().type("SearchTool").parameters(Map.of("key1", "value1")).build());
+
+        AtomicReference<String> lastActionInput = new AtomicReference<>();
+        String action = "SearchTool";
+
+        // Execute
+        Map<String, String> toolParams = AgentUtils.constructToolParams(tools, toolSpecMap, question, lastActionInput, action, actionInput);
+
+        // Verify - should fall back to actionInput
+        assertEquals(actionInput, toolParams.get("input"));
+        assertEquals(actionInput, toolParams.get(LLM_GEN_INPUT));
+    }
+
+    @Test
+    public void testConstructToolParams_InputWithMultipleSubstitutions() {
+        // Setup
+        String question = "What is the population?";
+        String actionInput = "{\"city\": \"Seattle\", \"year\": \"2025\"}";
+
+        Map<String, Tool> tools = Map.of("SearchTool", tool1);
+        Map<String, MLToolSpec> toolSpecMap = Map
+            .of(
+                "SearchTool",
+                MLToolSpec
+                    .builder()
+                    .type("SearchTool")
+                    .parameters(Map.of("input", "Find population of ${parameters.city} in ${parameters.year}"))
+                    .build()
+            );
+
+        AtomicReference<String> lastActionInput = new AtomicReference<>();
+        String action = "SearchTool";
+
+        // Execute
+        Map<String, String> toolParams = AgentUtils.constructToolParams(tools, toolSpecMap, question, lastActionInput, action, actionInput);
+
+        // Verify
+        assertEquals("Find population of Seattle in 2025", toolParams.get("input"));
+        assertEquals(actionInput, toolParams.get(LLM_GEN_INPUT));
+    }
 }