Define runtime workspace selection contract

CONTRACT.md

@@ -0,0 +1,91 @@
+# hotdata-runtime Contract
+
+`hotdata-runtime` is the framework-agnostic runtime contract for Hotdata integrations.
+
+## Scope
+
+This package provides shared primitives for:
+
+- Environment and workspace resolution
+- Query execution and polling
+- Normalized tabular result handling
+- Basic workspace health checks
+
+## Public Runtime Contract
+
+The supported import surface is:
+
+- `HotdataClient`
+- `QueryResult`
+- `from_env`
+- `workspace_health_lines`
+- `default_api_key`
+- `default_host`
+- `default_session_id`
+- `explicit_workspace_id`
+- `list_workspaces`
+- `normalize_host`
+- `pick_workspace`
+- `resolve_workspace_selection`
+- `ResultSummary`
+- `RunHistoryItem`
+- `WorkspaceSelection`
+
+Adapters should import from `hotdata_runtime` and treat this surface as the stable API.
+
+## Semantic Guarantees
+
+### `HotdataClient`
+
+- Represents runtime context: API key, host, workspace, optional session.
+- `from_env()` resolves runtime context from env vars and selected workspace.
+- `execute_sql(sql)` returns `QueryResult` or raises `RuntimeError`/`TimeoutError`.
+- `get_result(result_id)` returns a ready `QueryResult` and waits for readiness when needed.
+- `connections()` returns the connections API wrapper for adapter UI/status features.
+- `query_runs()` returns the query-runs API wrapper for adapter history views.
+- `results()` returns the results API wrapper for adapter result pickers.
+- `list_recent_results(...)` returns normalized `ResultSummary` entries.
+- `list_run_history(...)` returns normalized `RunHistoryItem` entries.
+- `list_qualified_table_names(...)` returns sorted fully qualified table names.
+- `columns_for_qualified(qualified, connection_id=...)` resolves table columns, and
+  adapters should pass `connection_id` when known.
+
+### `QueryResult`
+
+- Canonical tabular result model with `columns`, `rows`, and `row_count`.
+- Carries server identifiers and execution metadata when available.
+- `to_pandas()` converts to a DataFrame with stable column ordering.
+- `to_records(max_rows=...)` returns row dicts keyed by column names.
+- `metadata_dict()` returns normalized result metadata for adapter rendering.
+
+### Env Resolution
+
+- `default_api_key()` reads `HOTDATA_API_KEY` then `HOTDATA_TOKEN`.
+- `default_host()` reads `HOTDATA_API_URL` (default: `https://api.hotdata.dev`) and normalizes it.
+- `default_session_id()` reads `HOTDATA_SANDBOX`.
+- `pick_workspace()` prefers explicit env workspace, then active workspace, then first workspace.
+- `resolve_workspace_selection()` is the canonical workspace selection algorithm. It returns `WorkspaceSelection` with selected workspace id, selection source, and discovered workspaces when auto-selected.
+
+## Adapter Responsibilities
+
+Framework packages (Jupyter, Marimo, LangChain, LangGraph, LlamaIndex, Streamlit) own:
+
+- Framework-native lifecycle and state management
+- Rendering/UI concerns
+- Tool/agent wrappers and callback integration
+
+They should not duplicate runtime env/workspace/query semantics.
+
+## Runtime Non-Goals
+
+`hotdata-runtime` does not define framework UI primitives and does not require framework dependencies.
+
+## Versioning Policy
+
+- Backward-incompatible contract changes require a major version bump.
+- Additive contract changes are minor versions.
+- Bug fixes that preserve contract semantics are patch versions.
+
+## Enforcement
+
+Contract stability is enforced by tests that verify the public export surface and key behavioral invariants.

README.md

@@ -2,6 +2,8 @@
 
 Shared runtime primitives for Hotdata integrations: workspace/session semantics, execution context, query state, run history, and replayable result handles. Framework packages (Marimo, Jupyter, Streamlit, LangGraph) depend on this package.
 
+Runtime boundary and guarantees are defined in `CONTRACT.md`.
+
 Install:
 
 ```bash

hotdata_runtime/__init__.py

@@ -2,7 +2,12 @@
 
 from importlib.metadata import PackageNotFoundError, version
 
-from hotdata_runtime.client import HotdataClient, from_env
+from hotdata_runtime.client import (
+    HotdataClient,
+    ResultSummary,
+    RunHistoryItem,
+    from_env,
+)
 from hotdata_runtime.env import (
     default_api_key,
     default_host,
@@ -11,6 +16,8 @@
     list_workspaces,
     normalize_host,
     pick_workspace,
+    resolve_workspace_selection,
+    WorkspaceSelection,
 )
 from hotdata_runtime.health import workspace_health_lines
 from hotdata_runtime.result import QueryResult
@@ -33,4 +40,8 @@
     "list_workspaces",
     "normalize_host",
     "pick_workspace",
+    "resolve_workspace_selection",
+    "ResultSummary",
+    "RunHistoryItem",
+    "WorkspaceSelection",
 ]

hotdata_runtime/client.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+from dataclasses import asdict, dataclass
 import time
 from typing import Any, Iterator
 
@@ -25,6 +26,29 @@
 from hotdata_runtime.result import QueryResult
 
 _TERMINAL = frozenset({"succeeded", "failed", "cancelled"})
+_RESULT_FAILURE = frozenset({"failed", "cancelled"})
+
+
+@dataclass(frozen=True)
+class ResultSummary:
+    result_id: str
+    status: str
+    created_at: str | None
+
+    def to_dict(self) -> dict[str, Any]:
+        return asdict(self)
+
+
+@dataclass(frozen=True)
+class RunHistoryItem:
+    query_run_id: str
+    status: str
+    created_at: str | None
+    execution_time_ms: int | None
+    result_id: str | None
+
+    def to_dict(self) -> dict[str, Any]:
+        return asdict(self)
 
 
 class HotdataClient:
@@ -108,6 +132,39 @@ def query_runs(self) -> QueryRunsApi:
     def results(self) -> ResultsApi:
         return self._results_api()
 
+    def list_recent_results(
+        self,
+        *,
+        limit: int = 50,
+        offset: int = 0,
+    ) -> list[ResultSummary]:
+        listing = self.results().list_results(limit=limit, offset=offset)
+        return [
+            ResultSummary(
+                result_id=r.id,
+                status=r.status,
+                created_at=r.created_at,
+            )
+            for r in listing.results
+        ]
+
+    def list_run_history(
+        self,
+        *,
+        limit: int = 20,
+    ) -> list[RunHistoryItem]:
+        listing = self.query_runs().list_query_runs(limit=limit)
+        return [
+            RunHistoryItem(
+                query_run_id=r.id,
+                status=r.status,
+                created_at=r.created_at,
+                execution_time_ms=r.execution_time_ms,
+                result_id=r.result_id,
+            )
+            for r in listing.query_runs
+        ]
+
     def iter_tables(
         self,
         *,
@@ -143,9 +200,26 @@ def list_qualified_table_names(
 
     def connection_id_by_name(self) -> dict[str, str]:
         listing = self.connections().list_connections()
-        return {c.name: c.id for c in listing.connections}
+        id_map: dict[str, str] = {}
+        duplicate_names: set[str] = set()
+        for c in listing.connections:
+            if c.name in id_map and id_map[c.name] != c.id:
+                duplicate_names.add(c.name)
+            id_map[c.name] = c.id
+        if duplicate_names:
+            names = ", ".join(sorted(duplicate_names))
+            raise RuntimeError(
+                f"Duplicate connection names found: {names}. "
+                "Use an explicit connection_id."
+            )
+        return id_map
 
-    def columns_for_qualified(self, qualified: str) -> list[TableInfo]:
+    def columns_for_qualified(
+        self,
+        qualified: str,
+        *,
+        connection_id: str | None = None,
+    ) -> list[TableInfo]:
         parts = qualified.split(".")
         if len(parts) < 3:
             raise ValueError(
@@ -156,10 +230,12 @@ def columns_for_qualified(self, qualified: str) -> list[TableInfo]:
             parts[1],
             ".".join(parts[2:]),
         )
-        id_map = self.connection_id_by_name()
-        conn_id = id_map.get(conn_name)
-        if not conn_id:
-            raise KeyError(f"Unknown connection {conn_name!r}")
+        conn_id = connection_id
+        if conn_id is None:
+            id_map = self.connection_id_by_name()
+            conn_id = id_map.get(conn_name)
+            if not conn_id:
+                raise KeyError(f"Unknown connection {conn_name!r}")
         resp = self._information_schema().information_schema(
             connection_id=conn_id,
             var_schema=schema_name,
@@ -206,9 +282,9 @@ def _wait_result_ready(
             last = results.get_result(result_id)
             if last.status == "ready":
                 return last
-            if last.status == "failed":
+            if last.status in _RESULT_FAILURE:
                 raise RuntimeError(
-                    last.error_message or "Result persistence failed"
+                    last.error_message or f"Result {last.status}"
                 )
             time.sleep(interval_s)
         raise TimeoutError(

hotdata_runtime/env.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import os
+from dataclasses import dataclass
 from urllib.parse import urlparse
 
 from hotdata import ApiClient, Configuration
@@ -50,13 +51,35 @@ def list_workspaces(api_key: str, host: str, session_id: str | None):
     return listing.workspaces
 
 
-def pick_workspace(api_key: str, host: str, session_id: str | None) -> str:
+@dataclass(frozen=True)
+class WorkspaceSelection:
+    workspace_id: str
+    source: str
+    workspaces: list
+
+
+def resolve_workspace_selection(
+    api_key: str, host: str, session_id: str | None
+) -> WorkspaceSelection:
     explicit = explicit_workspace_id()
     if explicit:
-        return explicit
+        return WorkspaceSelection(
+            workspace_id=explicit,
+            source="explicit_env",
+            workspaces=[],
+        )
     workspaces = list_workspaces(api_key, host, session_id)
     if not workspaces:
         raise RuntimeError("No Hotdata workspaces found for this API key.")
     active = [w for w in workspaces if w.active]
     chosen = active[0] if active else workspaces[0]
-    return chosen.public_id
+    return WorkspaceSelection(
+        workspace_id=chosen.public_id,
+        source="active" if active else "first",
+        workspaces=workspaces,
+    )
+
+
+def pick_workspace(api_key: str, host: str, session_id: str | None) -> str:
+    selection = resolve_workspace_selection(api_key, host, session_id)
+    return selection.workspace_id

hotdata_runtime/result.py

@@ -20,6 +20,25 @@ class QueryResult:
     warning: str | None = None
     error_message: str | None = None
 
+    def to_records(
+        self,
+        *,
+        max_rows: int | None = None,
+    ) -> list[dict[str, Any]]:
+        rows = self.rows if max_rows is None else self.rows[:max_rows]
+        return [dict(zip(self.columns, row)) for row in rows]
+
+    def metadata_dict(self) -> dict[str, Any]:
+        return {
+            "row_count": self.row_count,
+            "column_count": len(self.columns),
+            "result_id": self.result_id,
+            "query_run_id": self.query_run_id,
+            "execution_time_ms": self.execution_time_ms,
+            "warning": self.warning,
+            "error_message": self.error_message,
+        }
+
     def to_pandas(self):  # type: ignore[no-untyped-def]
         import pandas as pd