chore: Update SDC Agents integration to v4.3.3

- Bump sdc-agents dependency from >=4.2.0 to >=4.3.3
- Update toolset descriptions for 4.3.x improvements (catalog-first
  discovery, structured unmatched columns, mapping context persistence,
  sidecar metadata support)
- Enhance demo agent to show full pipeline (5 toolsets) instead of
  minimal 2-toolset example

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: twcook

contributing/samples/sdc_agents_demo/agent.py

@@ -12,11 +12,11 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-"""Minimal SDC Agents demo -- catalog search + data introspection.
+"""SDC Agents demo -- full data governance pipeline.
 
-Composes CatalogToolset and IntrospectToolset into a single LlmAgent
-that can discover published SDC4 schemas and introspect datasource
-structure.
+Composes five SDC Agents toolsets into a single LlmAgent that can
+introspect datasources, discover matching catalog components, map
+columns to schemas, and assemble validated data models.
 
 Prerequisites:
     pip install google-adk-community[sdc-agents]
@@ -31,6 +31,9 @@
 from google.adk_community.sdc_agents import (
     CatalogToolset,
     IntrospectToolset,
+    MappingToolset,
+    AssemblyToolset,
+    ValidationToolset,
     SDCAgentsConfig,
 )
 
@@ -53,16 +56,24 @@
     name="sdc_demo_agent",
     model="gemini-2.0-flash",
     description=(
-        "Discovers SDC4 schemas and introspects datasource structure."
+        "Full data governance pipeline: introspect, discover, map,"
+        " and assemble SDC4 data models."
     ),
     instruction=(
-        "You help data engineers govern their data.\n"
-        "1. Use catalog tools to search for published SDC4 schemas\n"
-        "2. Use introspect tools to analyze datasource columns and types\n"
-        "3. Summarize findings clearly"
+        "You help data engineers govern their data. Follow this workflow:\n"
+        "1. Introspect the datasource to discover columns and types\n"
+        "2. Search the SDC4 catalog for matching published schemas\n"
+        "3. Discover catalog components that match the datasource structure\n"
+        "4. Map unmatched columns to schema components by similarity\n"
+        "5. Propose a cluster hierarchy for the data model\n"
+        "6. Assemble the final data model via the Assembly API\n"
+        "7. Validate the generated artifacts"
     ),
     tools=[
-        CatalogToolset(config=config),
         IntrospectToolset(config=config),
+        CatalogToolset(config=config),
+        MappingToolset(config=config),
+        AssemblyToolset(config=config),
+        ValidationToolset(config=config),
     ],
 )

pyproject.toml

@@ -46,7 +46,7 @@ test = [
     "pytest-asyncio>=1.2.0",
 ]
 sdc-agents = [
-    "sdc-agents>=4.2.0",
+    "sdc-agents>=4.3.3",
 ]
 
 

src/google/adk_community/sdc_agents/README.md

@@ -49,10 +49,10 @@ agent = LlmAgent(
 
 | Toolset | Description |
 |---------|-------------|
-| **CatalogToolset** | Discover published SDC4 schemas, download artifacts (XSD, RDF, JSON-LD), check wallet balance |
-| **IntrospectToolset** | Analyze datasource structure -- infer column types and constraints from SQL, CSV, JSON, MongoDB |
-| **MappingToolset** | Match datasource columns to schema components by type compatibility and name similarity |
-| **AssemblyToolset** | Compose data models from catalog components -- reuse existing or mint new, with wallet billing |
+| **CatalogToolset** | Discover published SDC4 schemas, download artifacts (XSD, RDF, JSON-LD) |
+| **IntrospectToolset** | Analyze datasource structure -- infer column types and constraints from SQL, CSV, JSON, MongoDB with sidecar metadata support |
+| **MappingToolset** | Match datasource columns to schema components by type compatibility and name similarity, persist mapping configs with schema and datasource context |
+| **AssemblyToolset** | Compose data models from catalog components -- reuse existing or mint new, with catalog-first discovery and structured unmatched column reporting |
 | **GeneratorToolset** | Generate validated XML instances, batch processing, and preview |
 | **ValidationToolset** | Validate XML instances against schemas, digitally sign via VaaS API |
 | **DistributionToolset** | Deliver RDF triples to Fuseki, Neo4j, GraphDB, or REST endpoints |

src/google/adk_community/sdc_agents/__init__.py

@@ -20,6 +20,8 @@
 
 Install: pip install google-adk-community[sdc-agents]
 Docs: https://github.com/SemanticDataCharter/SDC_Agents
+
+Requires sdc-agents >= 4.3.3.
 """
 
 from sdc_agents.common.config import load_config