duckdb-sqlalchemy 1.5.2.2__tar.gz → 1.5.4__tar.gz

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/.github/workflows/pythonapp.yaml

@@ -66,4 +66,5 @@ jobs:
           token: ${{ secrets.CODECOV_TOKEN }}
           env_vars: PYTHON
           fail_ci_if_error: true
+          use_pypi: true
           verbose: true

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/.pre-commit-config.yaml

@@ -23,10 +23,10 @@ repos:
         language: python
         files: ^duckdb_sqlalchemy/(?!tests/).*\.py$
         additional_dependencies:
-          - "ty==0.0.34"
+          - "ty==0.0.54"
   - repo: https://github.com/astral-sh/ruff-pre-commit
     # Ruff version.
-    rev: 'v0.15.11'
+    rev: 'v0.15.20'
     hooks:
       - id: ruff
         args:

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/CHANGELOG.md

@@ -8,6 +8,47 @@ preserved from the upstream project for historical context.
 
 ## Unreleased
 
+## [1.5.4](https://github.com/leonardovida/duckdb-sqlalchemy/compare/v1.5.3...v1.5.4) (2026-06-26)
+
+### Features
+
+- expose MotherDuck `md_user_info().c.region` and Flight run `config` columns in the SQLAlchemy table-function helpers
+- document one-off `md_run_flight(config=...)` overrides and selecting effective Flight run config
+- route `motherduck_enable_server_side_temp_tables` through the MotherDuck startup URL parameter bucket
+
+### Bug Fixes
+
+- restore SQLAlchemy 2.0.0+ compatibility by using a reflection fallback on early 2.0 releases
+- reject empty MotherDuck Flight config keys before compiling Flight helper calls
+- reject invalid MotherDuck Flight config map keys and NUL-containing values before compiling Flight helper calls
+
+### Tooling
+
+- extend the validated compatibility matrix to DuckDB 1.5.4 and SQLAlchemy 2.0.51
+- bump the local `ty` pin to `0.0.54` and the Ruff pre-commit hook to `0.15.20` after validating the current checks against the latest non-breaking releases
+- refresh the locked development dependency set to the latest non-breaking transitive releases within existing version caps
+
+## [1.5.3](https://github.com/leonardovida/duckdb-sqlalchemy/compare/v1.5.2.2...v1.5.3) (2026-06-02)
+
+### Features
+
+- add `quack_query()` for stateless DuckDB Quack remote queries on DuckDB 1.5.3+
+- add `pragma_storage_info()` for SQLAlchemy access to DuckDB storage metadata
+- add MotherDuck job table-function helpers for job metadata, versions, runs, logs, and job lifecycle operations
+- add MotherDuck Dive table-function helpers for Dives-as-code create, update, fetch, versioning, and cleanup workflows
+
+### Bug Fixes
+
+- quote validated table-function named parameters so reserved names such as `limit` and `offset` work with SQLAlchemy helpers
+
+### Tooling
+
+- focus the nox compatibility matrix on active DuckDB release lines while keeping the runtime dependency floor unchanged
+- extend the DuckDB compatibility test matrix to include DuckDB 1.5.3
+- update the SQLAlchemy compatibility matrix to include SQLAlchemy 2.0.50
+- refresh the locked development dependency set to the latest non-breaking releases within existing version caps
+- bump the local `ty` pin to `0.0.38` after validating the current checks against the latest non-breaking release
+
 ## [1.5.2.2](https://github.com/leonardovida/duckdb-sqlalchemy/compare/v1.5.2.1...v1.5.2.2) (2026-05-11)
 
 ### Features

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: duckdb-sqlalchemy
-Version: 1.5.2.2
+Version: 1.5.4
 Summary: DuckDB SQLAlchemy dialect for DuckDB and MotherDuck
 Project-URL: Bug Tracker, https://github.com/leonardovida/duckdb-sqlalchemy/issues
 Project-URL: Changelog, https://github.com/leonardovida/duckdb-sqlalchemy/releases
@@ -27,7 +27,8 @@ Classifier: Topic :: Database :: Front-Ends
 Requires-Python: <4,>=3.9
 Requires-Dist: duckdb>=0.5.0
 Requires-Dist: packaging>=21
-Requires-Dist: sqlalchemy>=2.0.45
+Requires-Dist: sqlalchemy>=2.0.0; python_version < '3.14'
+Requires-Dist: sqlalchemy>=2.0.45; python_version >= '3.14'
 Provides-Extra: dev
 Requires-Dist: fsspec<2027.0.0,>=2025.2.0; extra == 'dev'
 Requires-Dist: github-action-utils<2.0.0,>=1.1.0; extra == 'dev'
@@ -46,7 +47,7 @@ Requires-Dist: pytest-snapshot<1.0.0,>=0.9.0; extra == 'dev'
 Requires-Dist: pytest<10.0.0,>=8.0.0; extra == 'dev'
 Requires-Dist: pytz>=2024.2; (python_version >= '3.12') and extra == 'dev'
 Requires-Dist: toml<0.11.0,>=0.10.2; extra == 'dev'
-Requires-Dist: ty==0.0.34; extra == 'dev'
+Requires-Dist: ty==0.0.54; extra == 'dev'
 Provides-Extra: devtools
 Requires-Dist: nox<2027.0.0,>=2026.4.10; extra == 'devtools'
 Requires-Dist: pdbpp<0.13.0,>=0.11.0; extra == 'devtools'
@@ -106,8 +107,8 @@ Current direction in this repository:
 | Component | Supported versions |
 | --- | --- |
 | Python | 3.9+ |
-| SQLAlchemy | 2.0.45+ (CI-tested: 2.0.45, 2.0.49) |
-| DuckDB | 0.5.0+ (CI-tested currently: 1.1.3 to 1.5.2) |
+| SQLAlchemy | 2.0.0+ on Python <3.14; 2.0.45+ on Python 3.14+ (CI-tested: 2.0.0, 2.0.51) |
+| DuckDB | 0.5.0+ (CI-tested currently: 1.3.0 to 1.5.4) |
 
 ## Install
 

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/README.md

@@ -51,8 +51,8 @@ Current direction in this repository:
 | Component | Supported versions |
 | --- | --- |
 | Python | 3.9+ |
-| SQLAlchemy | 2.0.45+ (CI-tested: 2.0.45, 2.0.49) |
-| DuckDB | 0.5.0+ (CI-tested currently: 1.1.3 to 1.5.2) |
+| SQLAlchemy | 2.0.0+ on Python <3.14; 2.0.45+ on Python 3.14+ (CI-tested: 2.0.0, 2.0.51) |
+| DuckDB | 0.5.0+ (CI-tested currently: 1.3.0 to 1.5.4) |
 
 ## Install
 

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/docs/configuration.md

@@ -50,7 +50,8 @@ engine = create_engine("duckdb:///analytics.db", connect_args={"read_only": True
 - Unknown but syntactically valid keys are forwarded to DuckDB and may fail with
   DuckDB's native "unrecognized configuration parameter" error.
 - MotherDuck path-query options (`attach_mode`, `session_name`, `access_mode`,
-  and related keys) are handled specially for `md:` URLs. See
+  `motherduck_enable_server_side_temp_tables`, and related keys) are handled
+  specially for `md:` URLs. See
   [motherduck.md](motherduck) and [connection-urls.md](connection-urls).
 
 ## Preload extensions

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/docs/connection-urls.md

@@ -108,6 +108,15 @@ engine = create_engine(f"duckdb:///md:my_db?motherduck_token={escaped}")
 
 See [motherduck.md](motherduck) for MotherDuck-specific examples and options.
 
+## Quack remotes
+
+DuckDB 1.5.3 ships Quack as a core extension for remote DuckDB access. Quack is
+normally used from SQL with `ATTACH 'quack:host' AS name (...)` or the
+`quack_query` table function; keep the SQLAlchemy engine connected to a local
+DuckDB database and attach the remote from that session.
+
+See [olap.md](olap) for `quack_query` and `ATTACH` examples.
+
 ## Pool defaults
 
 Pooling behavior is described in [configuration.md](configuration)

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/docs/motherduck.md

@@ -70,6 +70,7 @@ Parameters that are treated as part of the database string:
 - `dbinstance_inactivity_ttl` (preferred; `motherduck_dbinstance_inactivity_ttl`
   remains supported but is deprecated)
 - `saas_mode` (`pgendpoint` is supported when you need PG endpoint-compatible routing)
+- `motherduck_enable_server_side_temp_tables`
 - `cache_buster`
 
 For backward compatibility the dialect also accepts `session_hint`,

{duckdb_sqlalchemy-1.5.2.2 → duckdb_sqlalchemy-1.5.4}/docs/olap.md

@@ -50,25 +50,131 @@ parquet = table_function(
 stmt = select(parquet.c.event_id, parquet.c.ts)
 ```
 
-## MotherDuck user metadata
+## Storage metadata
 
-MotherDuck exposes `md_user_info()` for the current user and organization.
-The helper names the released columns so they are available through SQLAlchemy:
+Use `pragma_storage_info` to inspect DuckDB table storage through SQLAlchemy:
 
 ```python
 from sqlalchemy import select
-from duckdb_sqlalchemy import md_user_info
+from duckdb_sqlalchemy import pragma_storage_info
+
+storage = pragma_storage_info("events")
+stmt = select(storage.c.column_name, storage.c.segment_type, storage.c.compression)
+```
+
+The helper names DuckDB's released columns by default. On engines that support
+the optional segment-info argument, pass `include_segment_info=True`.
+
+## Quack remote queries
+
+DuckDB 1.5.3 ships Quack as a core extension. Use `quack_query` for stateless
+remote queries and name the returned columns for SQLAlchemy:
+
+```python
+from sqlalchemy import select
+from duckdb_sqlalchemy import quack_query
+
+remote = quack_query(
+    "quack:localhost",
+    "SELECT 42 AS answer",
+    columns=["answer"],
+    token="MY_QUACK_TOKEN_01234567890ABCDEF",
+)
+stmt = select(remote.c.answer)
+```
+
+## MotherDuck metadata
+
+MotherDuck exposes table functions for account and Dive metadata. The helpers
+name the released columns so they are available through SQLAlchemy:
+
+```python
+from sqlalchemy import select
+from duckdb_sqlalchemy import md_access_tokens, md_list_dives, md_user_info
 
 user_info = md_user_info()
-stmt = select(
+user_stmt = select(
     user_info.c.user_id,
     user_info.c.username,
     user_info.c.org_id,
     user_info.c.org_name,
     user_info.c.org_type,
+    user_info.c.region,
 )
+
+dives = md_list_dives()
+dives_stmt = select(dives.c.id, dives.c.title, dives.c.required_resources)
+
+tokens = md_access_tokens()
+tokens_stmt = select(tokens.c.token_name, tokens.c.token_type, tokens.c.expire_at)
 ```
 
+## MotherDuck Dives as code
+
+Dive helpers expose the SQL functions MotherDuck provides for managing Dives
+from a SQL client or deployment script:
+
+```python
+from sqlalchemy import select
+from duckdb_sqlalchemy import (
+    md_create_dive,
+    md_get_dive,
+    md_list_dive_versions,
+    md_update_dive_content,
+)
+
+created = md_create_dive(
+    title="Sales overview",
+    content="export default function Dive() { return null }",
+    api_version=1,
+    required_resources=[{"url": "md:analytics", "alias": "analytics"}],
+)
+create_stmt = select(created.c.id, created.c.version_id)
+
+dive = md_get_dive(id="00000000-0000-0000-0000-000000000000")
+dive_stmt = select(dive.c.title, dive.c.content)
+
+updated = md_update_dive_content(
+    id="00000000-0000-0000-0000-000000000000",
+    content="export default function Dive() { return null }",
+)
+update_stmt = select(updated.c.version, updated.c.storage_url)
+
+versions = md_list_dive_versions(id="00000000-0000-0000-0000-000000000000")
+versions_stmt = select(versions.c.version, versions.c.created_at)
+```
+
+## MotherDuck Flights
+
+MotherDuck also exposes preview table functions for Flight metadata. The
+read-only helpers are useful for listing Flights, runs, logs, and versions:
+
+```python
+from sqlalchemy import select
+from duckdb_sqlalchemy import md_flight_runs, md_flight_versions, md_flights
+
+flights = md_flights(limit=10)
+flights_stmt = select(flights.c.flight_id, flights.c.flight_name, flights.c.status)
+
+runs = md_flight_runs(flight_id="00000000-0000-0000-0000-000000000000", limit=10)
+runs_stmt = select(runs.c.run_number, runs.c.status, runs.c.config, runs.c.started_at)
+
+versions = md_flight_versions(flight_id="00000000-0000-0000-0000-000000000000")
+versions_stmt = select(versions.c.flight_version, versions.c.requirements_txt)
+```
+
+Mutating Flight functions are available as helpers too:
+`md_create_flight`, `md_update_flight`, `md_delete_flight`, `md_run_flight`,
+and `md_cancel_flight_run`. They only execute when the SQLAlchemy statement is
+run. `md_run_flight` accepts a `config` map to override a Flight's config for
+that run only; run result helpers expose the effective `config` column. Config
+map keys must be strings, non-empty, and cannot contain `=` or NUL bytes.
+Config values must be strings or `None`, and cannot contain NUL bytes.
+
+The older `md_*job*` helper names remain as deprecated compatibility aliases.
+They compile through the current Flight functions while preserving legacy
+`job_*` column access where possible.
+
 ## Arrow results
 
 For large reads, you can request Arrow tables directly:
@@ -162,6 +268,18 @@ conn.execute(text("ATTACH 'analytics.duckdb' AS analytics"))
 rows = conn.execute(text("SELECT * FROM analytics.events LIMIT 10")).fetchall()
 ```
 
+Quack remotes can be attached the same way when a Quack server is available:
+
+```python
+conn.execute(
+    text(
+        "ATTACH 'quack:localhost' AS remote_db "
+        "(TOKEN 'MY_QUACK_TOKEN_01234567890ABCDEF')"
+    )
+)
+rows = conn.execute(text("SELECT * FROM remote_db.events LIMIT 10")).fetchall()
+```
+
 ## Notes
 
 - Column naming for table functions requires SQLAlchemy >= 1.4 (uses `table_valued`).