resolvekit 0.1.0__tar.gz → 0.1.2__tar.gz

{resolvekit-0.1.0 → resolvekit-0.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: resolvekit
-Version: 0.1.0
+Version: 0.1.2
 Summary: Entity and place resolution system that maps messy place/entity strings and codes to canonical entities
 Keywords: entity-resolution,geocoding,place-names,data-commons,iso-codes,offline,disambiguation,normalization
 Author: Jorge Rivera

{resolvekit-0.1.0 → resolvekit-0.1.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "resolvekit"
-version = "0.1.0"
+version = "0.1.2"
 description = "Entity and place resolution system that maps messy place/entity strings and codes to canonical entities"
 requires-python = ">=3.12"
 

{resolvekit-0.1.0 → resolvekit-0.1.2}/src/resolvekit/NOTICE.md

@@ -43,12 +43,25 @@ license requires attribution, that attribution is given below.
 
 ## OECD DAC (Development Assistance Committee)
 
-- **Contributes:** DAC country codes and recipient classifications used in the
-  geo entity store.
-- **Upstream:** https://www.oecd.org/dac/financing-sustainable-development/development-finance-standards/dacandcrscodelists.htm
-- **License:** OECD terms of use — data is freely available for non-commercial
-  and research use with attribution.
-  https://www.oecd.org/termsandconditions/
+- **Contributes:** DAC recipient, provider, channel, and agency codes with
+  their English/French names and an ISO3 crosswalk. These populate DAC codes
+  on countries and regions in the geo entity store and provider and
+  government-agency entities in the org entity store.
+- **Upstream:** https://development-finance-codelists.oecd.org/ (DAC and CRS
+  code lists)
+- **License:** OECD Terms and Conditions. Since 1 July 2024 most OECD data and
+  content is published under Creative Commons Attribution 4.0 International
+  (CC BY 4.0), which permits reuse — including commercial use — with
+  attribution; earlier content is available on terms similar to CC BY 4.0.
+  https://www.oecd.org/en/about/terms-conditions.html
+  https://creativecommons.org/licenses/by/4.0/
+- **Modifications:** only the codelists (codes, names, and the ISO3 crosswalk)
+  were extracted and repackaged; no OECD statistical, financial, or aid-flow
+  data is redistributed.
+- **Attribution required:** OECD DAC codelists are used under the OECD Terms
+  and Conditions (CC BY 4.0). Source: OECD Development Assistance Committee
+  (DAC). The OECD logo and branding are not covered by this license and are
+  not used.
 
 ## HDX Python Country / UN M49
 

{resolvekit-0.1.0 → resolvekit-0.1.2}/src/resolvekit/core/api/loading/module_catalog.py

@@ -134,12 +134,41 @@ def _load_and_separate_datapacks(
     return base_packs, overlay_packs
 
 
+def _remote_dependency_uncached(
+    module_id: str,
+    available: dict[str, Path],
+    manifest_overrides: dict[str, dict[str, object]],
+) -> bool:
+    """Return True if ``module_id`` is a remote-distribution module whose data
+    is not in the local cache.
+
+    Such a dependency may be absent from a load set without error: remote
+    packs the user hasn't downloaded are not hard errors (partial caches are
+    first-class, and ``Resolver.auto()`` never triggers a network fetch).
+    Bundled dependencies and unknown module ids stay hard errors.
+    """
+    from resolvekit.core.module_registry import load_module_metadata
+    from resolvekit.core.remote import is_cached
+
+    if module_id not in available:
+        return False
+    path = available[module_id]
+    if not (path / "metadata.json").exists():
+        return False
+    metadata = load_module_metadata(module_id, path, overrides=manifest_overrides)
+    return metadata.distribution == "remote" and not is_cached(metadata)
+
+
 def _validate_module_dependencies(
     base_packs: dict[str, LoadedDataPack],
     overlay_packs: dict[str, LoadedDataPack],
     pack_filter: set[str],
 ) -> None:
+    from resolvekit.core.module_registry import get_manifest_overrides
+
     available_module_ids = set(base_packs) | set(overlay_packs)
+    registry: dict[str, Path] | None = None
+    manifest_overrides: dict[str, dict[str, object]] | None = None
     for loaded in [*base_packs.values(), *overlay_packs.values()]:
         if pack_filter and loaded.pack_id not in pack_filter:
             continue
@@ -149,6 +178,22 @@ def _validate_module_dependencies(
             for module_id in loaded.metadata.module_dependencies
             if module_id not in available_module_ids
         ]
+        if missing:
+            # An absent dependency is only a hard error when its data could
+            # have been loaded — a declared dep on a remote pack the user
+            # hasn't downloaded is skipped, mirroring the auto-mode intent in
+            # _resolve_requested_module_paths.
+            if registry is None:
+                registry = list_available_modules()
+                manifest_overrides = get_manifest_overrides()
+            assert manifest_overrides is not None
+            missing = [
+                module_id
+                for module_id in missing
+                if not _remote_dependency_uncached(
+                    module_id, registry, manifest_overrides
+                )
+            ]
         if missing:
             raise MissingModuleDependencyError(loaded.module_id, missing)
 
@@ -250,8 +295,10 @@ def _resolve_requested_module_paths(
                 continue
             # In auto mode, silently skip dependencies whose data isn't
             # locally available (remote packs the user hasn't downloaded
-            # are not hard errors — see v1-scope §225). In explicit mode
-            # we always queue them so the queue-loop's own
+            # are not hard errors — see v1-scope §225). In explicit mode,
+            # likewise skip remote-uncached dependencies so requesting one
+            # module never transitively forces sibling downloads; unknown
+            # module ids are still queued so the queue-loop's own
             # ``module_id not in available`` check raises
             # ``DataModuleNotFoundError``.
             if auto_mode and (
@@ -261,6 +308,10 @@ def _resolve_requested_module_paths(
                 )
             ):
                 continue
+            if not auto_mode and _remote_dependency_uncached(
+                dependency, available, manifest_overrides
+            ):
+                continue
             queue.append(dependency)
         if metadata.is_overlay:
             for dependency in metadata.base_module_ids or []:

{resolvekit-0.1.0 → resolvekit-0.1.2}/src/resolvekit/core/remote.py

@@ -221,6 +221,14 @@ def download_module_data(
     return cache_dir
 
 
+def _progressbar_available() -> bool:
+    """Pooch's progress bar requires tqdm, which is not a resolvekit
+    dependency; downloads degrade to silent when it isn't installed."""
+    import importlib.util
+
+    return importlib.util.find_spec("tqdm") is not None
+
+
 def _download_one_artifact(
     metadata: DataPackMetadata,
     artifact_type: str,
@@ -234,7 +242,9 @@ def _download_one_artifact(
     processor = pooch.Decompress() if fname.endswith(".gz") else None
 
     try:
-        downloaded_path = fetcher.fetch(fname, progressbar=True, processor=processor)
+        downloaded_path = fetcher.fetch(
+            fname, progressbar=_progressbar_available(), processor=processor
+        )
     except Exception as exc:
         logger.error(
             "Download failed for %s/%s: %s", metadata.module_id, artifact_type, exc