Expand governed Dataflow editor and node library

src/govoplan_dataflow/backend/service.py

@@ -5,6 +5,12 @@ from dataclasses import dataclass
 from sqlalchemy import select
 from sqlalchemy.orm import Session
 
+from govoplan_core.auth import ApiPrincipal
+from govoplan_core.core.tabular_sources import (
+    TabularReadRequest,
+    TabularSourceError,
+    tabular_source_provider,
+)
 from govoplan_core.db.base import utcnow
 from govoplan_dataflow.backend.db.models import (
     DataflowPipeline,
@@ -14,6 +20,7 @@ from govoplan_dataflow.backend.db.models import (
 from govoplan_dataflow.backend.executor import (
     EXECUTOR_VERSION,
     PipelineExecutionError,
+    ResolvedSource,
     execute_preview,
 )
 from govoplan_dataflow.backend.graph import canonical_graph_payload, definition_hash, validate_graph
@@ -333,6 +340,8 @@ def preview_pipeline(
     tenant_id: str,
     actor_id: str | None,
     payload: PipelinePreviewRequest,
+    principal: ApiPrincipal | None = None,
+    registry: object | None = None,
 ) -> PipelinePreviewResponse:
     pipeline: DataflowPipeline | None = None
     revision: DataflowPipelineRevision | None = None
@@ -360,6 +369,8 @@ def preview_pipeline(
                 truncated=False,
                 diagnostics=validated.diagnostics,
                 node_diagnostics=[],
+                source_fingerprints=[],
+                input_row_count=0,
                 definition_hash="",
                 executor_version=EXECUTOR_VERSION,
             )
@@ -370,7 +381,47 @@ def preview_pipeline(
     started_at = utcnow()
     run: DataflowRun | None = None
     try:
-        result = execute_preview(graph, row_limit=payload.row_limit)
+        provider = tabular_source_provider(registry)
+
+        def resolve_source(node: GraphNode, limit: int) -> ResolvedSource:
+            if provider is None:
+                raise PipelineExecutionError(
+                    "Connector-backed preview requires the Connectors tabular-source capability.",
+                    node_id=node.id,
+                )
+            if principal is None:
+                raise PipelineExecutionError(
+                    "Connector-backed preview requires a tenant API principal.",
+                    node_id=node.id,
+                )
+            try:
+                resolved = provider.read_source(
+                    session,
+                    principal,
+                    request=TabularReadRequest(
+                        source_ref=str(node.config["source_ref"]),
+                        limit=limit,
+                        expected_fingerprint=_clean_optional(
+                            node.config.get("expected_fingerprint")
+                        ),
+                    ),
+                )
+            except TabularSourceError as exc:
+                raise PipelineExecutionError(str(exc), node_id=node.id) from exc
+            return ResolvedSource(
+                rows=tuple(dict(row) for row in resolved.rows),
+                source_ref=resolved.source.ref,
+                provider=resolved.source.provider,
+                fingerprint=resolved.source.fingerprint,
+                total_rows=resolved.total_rows,
+                truncated=resolved.truncated,
+            )
+
+        result = execute_preview(
+            graph,
+            row_limit=payload.row_limit,
+            source_resolver=resolve_source,
+        )
         status = "succeeded"
         error = None
         diagnostics = result.diagnostics
@@ -385,6 +436,7 @@ def preview_pipeline(
         status = "failed"
         error = str(exc)
         diagnostics = [
+            *exc.diagnostics,
             DataflowDiagnostic(
                 severity="error",
                 code="preview.execution",
@@ -396,9 +448,9 @@ def preview_pipeline(
         rows = []
         total_rows = 0
         truncated = False
-        node_diagnostics = []
-        source_fingerprints = []
-        input_row_count = 0
+        node_diagnostics = list(exc.node_diagnostics)
+        source_fingerprints = list(exc.source_fingerprints)
+        input_row_count = exc.input_row_count
 
     if pipeline is not None and revision is not None:
         run = DataflowRun(
@@ -433,6 +485,8 @@ def preview_pipeline(
         truncated=truncated,
         diagnostics=diagnostics,
         node_diagnostics=node_diagnostics,
+        source_fingerprints=source_fingerprints,
+        input_row_count=input_row_count,
         definition_hash=graph_hash,
         executor_version=EXECUTOR_VERSION,
     )