sdmxlib 0.68.0__tar.gz → 0.69.0__tar.gz

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: sdmxlib
-Version: 0.68.0
+Version: 0.69.0
 Summary: SDMX structural metadata library for Python
 Keywords: sdmx,statistics,metadata,datastructure
 Author: gabrielgellner
@@ -48,11 +48,10 @@ with sl.RestRegistry(sl.Provider.BIS) as reg:
 
     for dim in dsd.dimensions:
         if dim.is_enumerated:
-            cl = dim.representation()
-            print(f"{dim.id}: {cl.id} ({len(cl.codes)} codes)")
+            print(f"{dim.id}: {dim.codelist.id} ({len(dim.code_ids())} codes)")
 ```
 
-```
+```text
 FREQ: CL_FREQ (7 codes)
 REF_AREA: CL_BIS_IF_REF_AREA (44 codes)
 INSTR_ASSET: CL_INSTR_ASSET (12 codes)
@@ -66,6 +65,8 @@ uv add sdmxlib
 # or: pip install sdmxlib
 ```
 
+sdmxlib needs Python 3.14 or later.
+
 ## What it does
 
 - Connects to any SDMX 2.1 or 3.0 REST endpoint; built-in shortcuts for ECB,
@@ -76,25 +77,27 @@ uv add sdmxlib
 - Returns fully resolved domain objects with plain attribute access — no XML,
   no URN chasing
 - Filter collections with lambdas or the `item()` expression API
-- Build SDMX artefacts locally and serialise to SDMX-ML or SDMX-JSON
-- `LocalRegistry` — a DuckDB-backed persistent store for offline access, and
-  a position in a `RestRegistry`'s source chain via `store=`
-- A path query language over both planes: `reg.query(sl.q.dataflow)` for
-  structures, `reg.observations(flow)` for rows
-  with lazy codelist + hierarchy loading, automatic Ref resolution, and
-  direct SQL code maps; supports shared connections and any-language queries
+- Build SDMX artefacts locally — frozen, checked as they are built — and
+  serialise to SDMX-ML 2.1, 3.0 and 3.1 or SDMX-JSON 2.0
+- `LocalRegistry` — a DuckDB-backed persistent store with lazy codelists and
+  hierarchies and direct SQL code maps; pass it as a `RestRegistry`'s
+  `store=` and everything fetched is kept for offline use
+- A path query language over both planes, on a `LocalRegistry` or straight
+  against REST: `reg.query(sl.q.dataflow)` for structures,
+  `reg.obs(flow)` for rows, `reg.sql(...)` as the escape hatch
 - Uniform SKOS-aligned tree API (`roots`/`leaves`/`narrower`/`broader`/
   `*_transitive`/`depth_of` plus bulk projections `depths`/`broader_map`/
   `ancestors_map`/`inherit_along_broader`) on `Codelist`, `Hierarchy`,
   `CategoryScheme`, `MetadataStructure` — and their lazy SQL variants
-  `LazyCodelist`/`LazyHierarchy`
+  `LazyCodelist`/`LazyHierarchy`. A codelist is keyed by code id; the other
+  trees by path (`h["CA.ON"]`), since a node id may repeat across branches
 - Reference metadata via `MetadataStructure` (MSD), `Metadataflow`, and
   `MetadataSet` — attach typed footnotes, methodology notes, or quality
   reports to any identifiable artefact. Definitions ride in
   `StructureMessage`; populated payloads ride in `MetadataMessage`
   (`<mes:GenericMetadata>` in SDMX-ML 3.0; metadata-schema envelope in
   SDMX-JSON 2.0). Both message kinds round-trip through
-  `Registry`, JSON, and XML.
+  `LocalRegistry`, JSON, and XML.
 - Integrates with [Polars](https://pola.rs) for schema generation and data
   validation
 - Fully typed — IDE completion and static analysis work throughout
@@ -104,22 +107,24 @@ uv add sdmxlib
 Two condensed reference files at the repo root are written for LLM tools
 (Claude Code, Cursor, Copilot, etc.):
 
-- [`llms.txt`](llms.txt) — full library overview: install, registries,
-  data fetching, reference metadata, and the domain-type cheat sheet
-- [`llms-dsl.txt`](llms-dsl.txt) — focused query-DSL reference:
-  `sl.item` / `sl.q` / `sl.dim` predicates, the `Query` / `Frame` /
-  `DataQuery` chains, and the uniform SKOS tree API
+- [`llms.txt`](llms.txt) — the whole library in one pass: registries and
+  `get`, navigation, building artefacts, the local store, trees, observation
+  data, reading and writing SDMX, reference metadata, and the type table
+- [`llms-dsl.txt`](llms-dsl.txt) — the path query language in depth:
+  `sl.q`, `Query` / `Select`, projection and aggregation, edges, windows,
+  observation queries (`ObsQuery`), and `reg.sql()`
 
-Point your assistant at these to keep generated code aligned with the
-library's actual surface.
+Every Python block in both is executed by the test suite, so they track the
+library's actual surface. Point your assistant at them.
 
 ## Upgrading
 
 sdmxlib is pre-stability and breaks cleanly rather than carrying two spellings
-of the same idea. If you are coming from **0.48**, read
-[Upgrading](https://pinax-suite.gitlab.io/sdmxlib/upgrading/) — `sl.Registry`
-is now `sl.LocalRegistry`, `get()` takes a URN and a resolution depth, and the
-four query DSLs are one path language.
+of the same idea. [Upgrading](https://pinax-suite.gitlab.io/sdmxlib/upgrading/)
+lists every breaking change since 0.48, with the old and new spelling of each —
+`sl.Registry` is now `sl.LocalRegistry`, `get()` takes a URN and a resolution
+depth, the four query DSLs are one path language, and every model class is
+frozen.
 
 ## Documentation
 
@@ -140,6 +145,6 @@ including:
 ```bash
 just check           # lint + typecheck
 just test            # unit tests
-just test-integration  # live endpoint tests (requires network)
-just ci              # full pre-commit gate
+just test-integration  # end-to-end tests, replayed from recorded cassettes
+just ci              # the full gate: lint, typecheck, unit, docs, integration
 ```

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/README.md

@@ -19,11 +19,10 @@ with sl.RestRegistry(sl.Provider.BIS) as reg:
 
     for dim in dsd.dimensions:
         if dim.is_enumerated:
-            cl = dim.representation()
-            print(f"{dim.id}: {cl.id} ({len(cl.codes)} codes)")
+            print(f"{dim.id}: {dim.codelist.id} ({len(dim.code_ids())} codes)")
 ```
 
-```
+```text
 FREQ: CL_FREQ (7 codes)
 REF_AREA: CL_BIS_IF_REF_AREA (44 codes)
 INSTR_ASSET: CL_INSTR_ASSET (12 codes)
@@ -37,6 +36,8 @@ uv add sdmxlib
 # or: pip install sdmxlib
 ```
 
+sdmxlib needs Python 3.14 or later.
+
 ## What it does
 
 - Connects to any SDMX 2.1 or 3.0 REST endpoint; built-in shortcuts for ECB,
@@ -47,25 +48,27 @@ uv add sdmxlib
 - Returns fully resolved domain objects with plain attribute access — no XML,
   no URN chasing
 - Filter collections with lambdas or the `item()` expression API
-- Build SDMX artefacts locally and serialise to SDMX-ML or SDMX-JSON
-- `LocalRegistry` — a DuckDB-backed persistent store for offline access, and
-  a position in a `RestRegistry`'s source chain via `store=`
-- A path query language over both planes: `reg.query(sl.q.dataflow)` for
-  structures, `reg.observations(flow)` for rows
-  with lazy codelist + hierarchy loading, automatic Ref resolution, and
-  direct SQL code maps; supports shared connections and any-language queries
+- Build SDMX artefacts locally — frozen, checked as they are built — and
+  serialise to SDMX-ML 2.1, 3.0 and 3.1 or SDMX-JSON 2.0
+- `LocalRegistry` — a DuckDB-backed persistent store with lazy codelists and
+  hierarchies and direct SQL code maps; pass it as a `RestRegistry`'s
+  `store=` and everything fetched is kept for offline use
+- A path query language over both planes, on a `LocalRegistry` or straight
+  against REST: `reg.query(sl.q.dataflow)` for structures,
+  `reg.obs(flow)` for rows, `reg.sql(...)` as the escape hatch
 - Uniform SKOS-aligned tree API (`roots`/`leaves`/`narrower`/`broader`/
   `*_transitive`/`depth_of` plus bulk projections `depths`/`broader_map`/
   `ancestors_map`/`inherit_along_broader`) on `Codelist`, `Hierarchy`,
   `CategoryScheme`, `MetadataStructure` — and their lazy SQL variants
-  `LazyCodelist`/`LazyHierarchy`
+  `LazyCodelist`/`LazyHierarchy`. A codelist is keyed by code id; the other
+  trees by path (`h["CA.ON"]`), since a node id may repeat across branches
 - Reference metadata via `MetadataStructure` (MSD), `Metadataflow`, and
   `MetadataSet` — attach typed footnotes, methodology notes, or quality
   reports to any identifiable artefact. Definitions ride in
   `StructureMessage`; populated payloads ride in `MetadataMessage`
   (`<mes:GenericMetadata>` in SDMX-ML 3.0; metadata-schema envelope in
   SDMX-JSON 2.0). Both message kinds round-trip through
-  `Registry`, JSON, and XML.
+  `LocalRegistry`, JSON, and XML.
 - Integrates with [Polars](https://pola.rs) for schema generation and data
   validation
 - Fully typed — IDE completion and static analysis work throughout
@@ -75,22 +78,24 @@ uv add sdmxlib
 Two condensed reference files at the repo root are written for LLM tools
 (Claude Code, Cursor, Copilot, etc.):
 
-- [`llms.txt`](llms.txt) — full library overview: install, registries,
-  data fetching, reference metadata, and the domain-type cheat sheet
-- [`llms-dsl.txt`](llms-dsl.txt) — focused query-DSL reference:
-  `sl.item` / `sl.q` / `sl.dim` predicates, the `Query` / `Frame` /
-  `DataQuery` chains, and the uniform SKOS tree API
+- [`llms.txt`](llms.txt) — the whole library in one pass: registries and
+  `get`, navigation, building artefacts, the local store, trees, observation
+  data, reading and writing SDMX, reference metadata, and the type table
+- [`llms-dsl.txt`](llms-dsl.txt) — the path query language in depth:
+  `sl.q`, `Query` / `Select`, projection and aggregation, edges, windows,
+  observation queries (`ObsQuery`), and `reg.sql()`
 
-Point your assistant at these to keep generated code aligned with the
-library's actual surface.
+Every Python block in both is executed by the test suite, so they track the
+library's actual surface. Point your assistant at them.
 
 ## Upgrading
 
 sdmxlib is pre-stability and breaks cleanly rather than carrying two spellings
-of the same idea. If you are coming from **0.48**, read
-[Upgrading](https://pinax-suite.gitlab.io/sdmxlib/upgrading/) — `sl.Registry`
-is now `sl.LocalRegistry`, `get()` takes a URN and a resolution depth, and the
-four query DSLs are one path language.
+of the same idea. [Upgrading](https://pinax-suite.gitlab.io/sdmxlib/upgrading/)
+lists every breaking change since 0.48, with the old and new spelling of each —
+`sl.Registry` is now `sl.LocalRegistry`, `get()` takes a URN and a resolution
+depth, the four query DSLs are one path language, and every model class is
+frozen.
 
 ## Documentation
 
@@ -111,6 +116,6 @@ including:
 ```bash
 just check           # lint + typecheck
 just test            # unit tests
-just test-integration  # live endpoint tests (requires network)
-just ci              # full pre-commit gate
+just test-integration  # end-to-end tests, replayed from recorded cassettes
+just ci              # the full gate: lint, typecheck, unit, docs, integration
 ```

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sdmxlib"
-version = "0.68.0"
+version = "0.69.0"
 description = "SDMX structural metadata library for Python"
 readme = "README.md"
 requires-python = ">=3.14"

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/pyproject.toml.orig

@@ -1,6 +1,6 @@
 [project]
 name = "sdmxlib"
-version = "0.68.0"
+version = "0.69.0"
 description = "SDMX structural metadata library for Python"
 readme = "README.md"
 license = { text = "Apache-2.0" }
@@ -18,9 +18,8 @@ classifiers = [
 dependencies = [
     "attrs>=26.1.0",
     # >=1.5.3: ducklake extension must support DuckLake catalog format 1.0,
-    # which is what statcan-catalog-builder (duckdb 1.5.3) writes into
-    # published bundles. duckdb 1.5.1's ducklake tops out at format 0.4 and
-    # fails to ATTACH such bundles with "DuckLake catalog version mismatch".
+    # which is what a bundle built with duckdb 1.5.3 carries. duckdb 1.5.1's
+    # ducklake tops out at format 0.4 and fails to ATTACH such bundles with "DuckLake catalog version mismatch".
     #
     # Moved to >=1.5.5 while investigating #400 — two patch releases of fixes,
     # taken on their own merits. It is **not** a fix for #400: 1.5.0, 1.5.1,

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/__init__.py

@@ -1,10 +1,8 @@
 """sdmxlib — SDMX structural metadata library for Python."""
 
 from sdmxlib.api.federated import FederatedRegistry
-from sdmxlib.api.filters import DimFilter, dim
 from sdmxlib.api.policy import RegistryUnavailable
 from sdmxlib.api.providers import Provider
-from sdmxlib.api.query import DataQuery
 from sdmxlib.formats.data_format import DataFormat
 from sdmxlib.formats.gaps import DroppedArtefact, Gap, dropped, unrepresentable
 from sdmxlib.api.client import ResponseTooLargeError
@@ -21,13 +19,32 @@ from sdmxlib import versions
 from sdmxlib.versions import RewrittenVersion, VersionRewrite
 from sdmxlib.query.accessors import ArtefactRow, CodeRow, HCodeRow, ReportedAttributeRow, q
 from sdmxlib.query._compile import CompiledQuery
-from sdmxlib.query._expr import Agg, AggPred, AnyPred, Col, Cols, Cursor, Pred, RowKind, Untyped, Win, WinFn, WinPred
+from sdmxlib.query._expr import (
+    Agg,
+    AggPred,
+    AnyPred,
+    Col,
+    Cols,
+    ColumnKind,
+    Cursor,
+    Num,
+    Pred,
+    RowKind,
+    Str,
+    Untyped,
+    Win,
+    WinFn,
+    WinPred,
+)
 from sdmxlib.query.observations import ObsQuery, ObsRow
-from sdmxlib.query.query import Frame, Grouped, Page, Query, union
+from sdmxlib.query.query import CrossSourceError, Select, Grouped, Page, Query, SliceError, union
+from sdmxlib.query._joins import PathError
+from sdmxlib.query._pushdown import PushdownError
 from sdmxlib.sql import SqlRelation
 from sdmxlib.formats import parse, read, serialize, to_json, to_xml
 from sdmxlib.local import LazyCodelist, LazyHierarchy, QueryPlan, LocalRegistry, SchemaInspector
 from sdmxlib.local.data_store import DuckDBDataStore, DuckLakeVersionError
+from sdmxlib.local.sql import SidecarNotUniqueError
 from sdmxlib.rest import ComponentFilter, parse_components, widen_key
 from sdmxlib.query._common import Descending
 from sdmxlib.model.convert import from_dict, to_dict
@@ -36,6 +53,7 @@ from sdmxlib.model.validation import (
     LossyReadError,
     ReferenceIssue,
     ReferenceValidationError,
+    SdmxConformanceError,
     SdmxParseError,
     ValidationError,
     ValidationIssue,
@@ -110,7 +128,6 @@ from sdmxlib.model import (
     OrganisationUnit,
     OrganisationUnitScheme,
     DataStructure,
-    Dataset,
     Dataflow,
     DataflowBinding,
     DataflowEntry,
@@ -230,6 +247,7 @@ __all__ = [
     "CodelistLike",
     "Col",
     "Cols",
+    "ColumnKind",
     "CompiledQuery",
     "ComponentFilter",
     "ComponentMap",
@@ -241,6 +259,7 @@ __all__ = [
     "ConstraintAttachment",
     "ConstraintType",
     "Contact",
+    "CrossSourceError",
     "CubeRegion",
     "Cursor",
     "DataConstraint",
@@ -251,7 +270,6 @@ __all__ = [
     "DataKeySet",
     "DataProvider",
     "DataProviderScheme",
-    "DataQuery",
     "DataSetReference",
     "DataStore",
     "DataStructure",
@@ -259,10 +277,8 @@ __all__ = [
     "Dataflow",
     "DataflowBinding",
     "DataflowEntry",
-    "Dataset",
     "DatePatternMap",
     "Descending",
-    "DimFilter",
     "Dimension",
     "DriftError",
     "DriftIssue",
@@ -276,7 +292,6 @@ __all__ = [
     "FederatedRegistry",
     "FixedValueMap",
     "FmrRegistry",
-    "Frame",
     "FrequencyFormatMapping",
     "Gap",
     "GeoFeatureSetCode",
@@ -323,6 +338,7 @@ __all__ = [
     "MetadataSet",
     "MetadataStructure",
     "Metadataflow",
+    "Num",
     "ObsQuery",
     "ObsRow",
     "OrganisationSchemeMap",
@@ -330,9 +346,11 @@ __all__ = [
     "OrganisationUnitScheme",
     "Page",
     "Party",
+    "PathError",
     "Pred",
     "Provider",
     "ProvisionAgreement",
+    "PushdownError",
     "Query",
     "QueryPlan",
     "QueryableDataSource",
@@ -361,16 +379,21 @@ __all__ = [
     "RewrittenVersion",
     "RowKind",
     "SchemaInspector",
+    "SdmxConformanceError",
     "SdmxParseError",
     "SdmxRef",
     "SdmxUrn",
     "SdmxWarning",
+    "Select",
     "Sender",
     "SentinelValue",
     "Session",
     "SessionHook",
+    "SidecarNotUniqueError",
+    "SliceError",
     "SourceValue",
     "SqlRelation",
+    "Str",
     "StructureMap",
     "StructureMessage",
     "StructureSet",
@@ -392,7 +415,6 @@ __all__ = [
     "WinPred",
     "WriteSource",
     "artefact_urn",
-    "dim",
     "dropped",
     "from_dict",
     "item",

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/_duckdb.py

@@ -4,8 +4,6 @@ Lives at the package root because both ``storage/`` and ``query/`` need it, and
 neither layer may import the other.
 """
 
-from __future__ import annotations
-
 import duckdb
 
 

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/_singleflight.py

@@ -21,13 +21,8 @@ Example:
     flight.run(key, lambda: fetch_into(store, plan, client))
 """
 
-from __future__ import annotations
-
 import threading
-from typing import TYPE_CHECKING
-
-if TYPE_CHECKING:
-    from collections.abc import Callable, Hashable
+from collections.abc import Callable, Hashable
 
 __all__ = ["SingleFlight"]
 

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/_time.py

@@ -32,13 +32,8 @@ which is the property that matters and the one `TIMESTAMP` was silently
 breaking.
 """
 
-from __future__ import annotations
-
-from datetime import UTC
-from typing import TYPE_CHECKING, overload
-
-if TYPE_CHECKING:
-    from datetime import datetime
+from datetime import UTC, datetime
+from typing import overload
 
 
 # Overloaded so a value known to be present stays `datetime` on the way out —

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/api/_freshness.py

@@ -30,17 +30,12 @@ more permissive, never shorten what a generous one offered. A caller who wants
 less can pass a pinned URN, which is not a pointer at all.
 """
 
-from __future__ import annotations
-
 import re
+from collections.abc import Mapping
 from datetime import UTC, datetime, timedelta
-from typing import TYPE_CHECKING
 
 from attrs import frozen
 
-if TYPE_CHECKING:
-    from collections.abc import Mapping
-
 __all__ = ["Freshness", "read_freshness"]
 
 _MAX_AGE = re.compile(r"\bmax-age\s*=\s*(\d+)", re.IGNORECASE)

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/api/client.py

@@ -210,7 +210,7 @@ def _check_content_type(response: httpx.Response, fmt: DataFormat, *, url: str)
     msg = (
         f"Asked for {fmt.media_type!r} but the response says {header!r}. The "
         f"provider ignored `Accept`, so the body is not the format requested — "
-        f"pass the format it does serve, or use `observations()`, which parses "
+        f"pass the format it does serve, or use `obs()`, which parses "
         f"whatever arrives. URL: {url}"
     )
     raise ValueError(msg)
@@ -550,51 +550,11 @@ class SdmxClient:
             return f"{self._endpoint}/data/dataflow/{agency}/{id}/{ver}/{key}"
         return f"{self._endpoint}/data/{agency},{id},{version}/{key}"
 
-    def get_data(
-        self,
-        agency: str,
-        id: str,
-        version: str = "latest",
-        *,
-        key: str = "all",
-        start_period: str | None = None,
-        end_period: str | None = None,
-    ) -> bytes:
-        """Fetch observation data as SDMX-CSV bytes.
-
-        Args:
-            agency: Dataflow maintainer agency id.
-            id: Dataflow id.
-            version: Dataflow version. Defaults to ``"latest"``.
-            key: SDMX key filter string, e.g. ``"M.AT+DE.INDPRO"``.
-                 Defaults to ``"all"`` (all observations).
-            start_period: ISO period start, e.g. ``"2020-01"``.
-            end_period: ISO period end.
-
-        Returns:
-            Raw SDMX-CSV bytes.
-        """
-        url = self._data_url(agency, id, version, key)
-        params: dict[str, str] = {}
-        if start_period is not None:
-            params["startPeriod"] = start_period
-        if end_period is not None:
-            params["endPeriod"] = end_period
-        with self._reaching(url):
-            response = self._client.get(
-                url,
-                params=params,
-                headers={"Accept": "text/csv"},
-            )
-        _raise_for_status(response, url=url)
-        return response.content
-
     def get_data_at(self, path: str, params: Mapping[str, str]) -> bytes:
         """Fetch SDMX-CSV bytes for an already-built request path and parameters.
 
         The seam `query._rest.DataRequest` needs: it owns the URL's shape —
-        which differs by API version — and this owns the HTTP. `get_data`
-        builds the path from parts instead, for the older `reg.data()` surface.
+        which differs by API version — and this owns the HTTP.
 
         Args:
             path: Path relative to the service root, as `DataRequest.path`
@@ -634,51 +594,6 @@ class SdmxClient:
                 for chunk in response.iter_bytes():
                     handle.write(chunk)
 
-    def stream_data_to(
-        self,
-        agency: str,
-        id: str,
-        version: str = "latest",
-        *,
-        path: Path,
-        key: str = "all",
-        start_period: str | None = None,
-        end_period: str | None = None,
-    ) -> None:
-        """Stream observation data directly to a file.
-
-        Uses httpx chunked streaming — the response is never fully loaded into
-        memory. Suitable for multi-GB datasets.
-
-        Args:
-            agency: Dataflow maintainer agency id.
-            id: Dataflow id.
-            version: Dataflow version.
-            path: Destination file path.
-            key: SDMX key filter string.
-            start_period: ISO period start.
-            end_period: ISO period end.
-        """
-        url = self._data_url(agency, id, version, key)
-        params: dict[str, str] = {}
-        if start_period is not None:
-            params["startPeriod"] = start_period
-        if end_period is not None:
-            params["endPeriod"] = end_period
-        with (
-            self._reaching(url),
-            self._client.stream(
-                "GET",
-                url,
-                params=params,
-                headers={"Accept": "text/csv"},
-            ) as response,
-        ):
-            _raise_for_status(response, url=url)
-            with path.open("wb") as f:
-                for chunk in response.iter_bytes():
-                    f.write(chunk)
-
     def download_data(
         self,
         agency: str,
@@ -693,9 +608,9 @@ class SdmxClient:
     ) -> bytes | None:
         """Fetch a dataflow's data in `fmt`, unparsed.
 
-        The byte-faithful counterpart to `get_data` / `stream_data_to`, which
+        The byte-faithful counterpart to `get_data_at` / `stream_data_at`, which
         hardcode ``Accept: text/csv`` because their caller is about to hand the
-        body to the SDMX-CSV reader. Here the caller wants what the provider
+        body to DuckDB's CSV reader. Here the caller wants what the provider
         sent, so the format is the caller's to choose and nothing inspects the
         body afterwards.
 

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/api/federated.py

@@ -33,22 +33,17 @@ Example:
 import annotationlib
 import functools
 import inspect
-from typing import TYPE_CHECKING, overload
+from collections.abc import Sequence
+from typing import overload
 
 import httpx
 
 from sdmxlib.api.policy import RegistryUnavailable
-from sdmxlib.model.binding import DriftError, _validate_binding
+from sdmxlib.model.binding import DataflowBinding, DriftError, DriftIssue, _validate_binding
 from sdmxlib.model.ref import Ref, target_urn
+from sdmxlib.model.registry import RegistryReader
 from sdmxlib.model.urn import SdmxUrn
 
-if TYPE_CHECKING:
-    from collections.abc import Sequence
-
-    from sdmxlib.model.binding import DataflowBinding, DriftIssue
-    from sdmxlib.model.registry import RegistryReader
-    from sdmxlib.model.urn import SdmxUrn
-
 
 @functools.cache
 def _get_parameters(source_type: type) -> frozenset[str]:

{sdmxlib-0.68.0 → sdmxlib-0.69.0}/src/sdmxlib/api/policy.py

@@ -27,8 +27,6 @@ so, which cannot mislabel anything:
             ...  # genuinely absent, and only that
 """
 
-from __future__ import annotations
-
 
 class RegistryUnavailable(Exception):  # noqa: N818  # name matches #46 vocabulary
     """A source could not be reached, so no read of it can be believed.