Feat/osm dem

CHANGELOG.md

@@ -8,6 +8,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+- **`type = "dem"` source**: elevation, or slope in degrees or percent, from SRTM, Copernicus GLO-30
+  or TOPODATA (or your own `tiles`), read over the grid plus a `margin` and registered as a raster
+  source with checksum and provenance. The slope is computed on a metric (UTM) grid by central
+  differences; asking for a Copernicus slope logs a warning, since that model includes the forest
+  canopy. TOPODATA sheets are downloaded once into a cache. Replaces the DEM script of the Lab15
+  reconstruction.
+- **`type = "osm"` source**: ways from OpenStreetMap through the Overpass API, clipped to the grid
+  plus a `margin`, registered as a vector source with checksum and provenance (request, server,
+  time retrieved, `timestamp_osm_base`, ODbL). The answer is cached by request, so a pipeline reads
+  the same data (and gets the same `spec_hash`) on every run; `date` asks for an older snapshot
+  (`date = "{year}-07-01"` with `years`); several public servers are tried with retries; HTTP 406
+  points to `OSM_CONTACT`. Replaces the OpenStreetMap download script of the Lab15 reconstruction.
+
+## [0.4.0] - 2026-09-30
+
 ### Changed
 - **DisSModel is now optional.** DisSCube no longer depends on it: install
   `disscube[dissmodel]` only to hand a cube to a model. `CubeClient.to_lucc_data()`

README.md

@@ -246,6 +246,8 @@ disscube/
 │   ├── _categorical.py  legends (.qml/.json/.csv) and reclassification
 │   ├── mapbiomas.py  MapBiomas annual land-cover maps (Collection 11, 10 m series)
 │   ├── prodes.py     PRODES deforestation (download + cache, legend from the .qml)
+│   ├── osm.py        OpenStreetMap ways via Overpass (cached, with date snapshots)
+│   ├── dem.py        Elevation and slope from SRTM, Copernicus GLO-30 or TOPODATA
 │   └── classified.py any classified map with its legend, e.g. from SITS
 └── utils.py          Checksums (sha256_file) and BDC tile importer (import_bdc_grids)
 ```

disscube/pipeline/runner.py

@@ -18,11 +18,13 @@
 from disscube.pipeline.schema import (
     BdcSource,
     ClassifiedSource,
+    DemSource,
     DeriveConfig,
     ExportConfig,
     FileSource,
     GridConfig,
     MapbiomasSource,
+    OsmSource,
     PipelineConfig,
     ProdesSource,
     UnionSource,
@@ -56,7 +58,7 @@ def base_dir(self) -> Path:
 @dataclass
 class PlannedSource:
     id: str
-    config: FileSource | BdcSource | MapbiomasSource | ProdesSource | ClassifiedSource | UnionSource
+    config: FileSource | BdcSource | MapbiomasSource | ProdesSource | ClassifiedSource | OsmSource | DemSource | UnionSource
     year: int | None = None
 
 
@@ -412,7 +414,7 @@ def _register_source(
 ):
     c = s.config
     # ── Validação defensiva: apenas fontes dinâmicas em janela na nuvem precisam de bbox_geo ──
-    if isinstance(c, (BdcSource, MapbiomasSource, ProdesSource, ClassifiedSource)):
+    if isinstance(c, (BdcSource, MapbiomasSource, ProdesSource, ClassifiedSource, OsmSource, DemSource)):
         if bbox_geo is None:
             raise PipelineError(
                 f"source {c.id!r} ({c.type}): windowed cloud sources require an"
@@ -448,6 +450,29 @@ def _register_source(
                 files=files,
                 name=c.name,
             )
+        if isinstance(c, OsmSource):
+            from disscube.sources import osm
+
+            try:
+                return osm.register_osm_source(
+                    cube, c.id, c.query, bbox_geo, raw, margin=c.margin, date=c.date,
+                    endpoints=c.endpoints, cache=(base / c.cache) if c.cache else None,
+                    time=c.time, name=c.name,
+                )
+            except (osm.OsmError, ValueError) as exc:
+                raise PipelineError(f"source {c.id!r}: {exc}") from None
+        if isinstance(c, DemSource):
+            from disscube.sources import dem
+
+            try:
+                return dem.register_dem_source(
+                    cube, c.id, bbox_geo, raw, dem=c.dem,
+                    tiles=[_resolve(base, t) for t in c.tiles] if c.tiles else None,
+                    product=c.product, margin=c.margin, resolution=c.resolution,
+                    cache=(base / c.cache) if c.cache else None, name=c.name,
+                )
+            except (dem.DemError, ValueError) as exc:
+                raise PipelineError(f"source {c.id!r}: {exc}") from None
         if isinstance(c, ClassifiedSource):
             from disscube.sources.classified import register_classified_map
 

disscube/pipeline/schema.py

@@ -131,6 +131,49 @@ class ClassifiedSource(_SourceBase):
     producer: str | None = None
 
 
+class OsmSource(_SourceBase):
+    """Ways from OpenStreetMap (Overpass), clipped to the grid plus ``margin``.
+
+    ``query`` holds Overpass statements without the bounding box (``way["highway"]["ref"~"BR-163"]``),
+    separated by ``;`` or given as a list. ``margin`` is in degrees: a ``distance`` needs
+    features beyond the grid, a ``count`` does not. ``date`` asks for the map as it was on
+    that day; with ``years``, ``date = "{year}-07-01"`` gives one snapshot per year. The answer
+    is cached, so a pipeline reads the same data until the cache is cleared. Set ``OSM_CONTACT``.
+    """
+
+    type: Literal["osm"]
+    query: str | list[str]
+    margin: float = Field(default=0.0, ge=0)
+    date: str | None = None
+    endpoints: list[str] | None = None
+    cache: str | None = None
+    time: int | None = None
+
+
+class DemSource(_SourceBase):
+    """Elevation or slope from a DEM (SRTM, Copernicus GLO-30, TOPODATA), over the grid plus ``margin``.
+
+    Give ``dem`` to download one, or ``tiles`` (GeoTIFFs, paths relative to the pipeline file or URLs
+    GDAL can read) to use your own. ``product`` is ``elevation`` (m), ``slope_deg`` or ``slope_pct``;
+    a slope is computed on a metric (UTM) grid of ``resolution`` metres. Prefer SRTM or TOPODATA to
+    Copernicus for slope: Copernicus includes the forest canopy.
+    """
+
+    type: Literal["dem"]
+    dem: Literal["srtm", "copernicus", "topodata"] | None = None
+    tiles: list[str] | None = None
+    product: Literal["elevation", "slope_deg", "slope_pct"] = "elevation"
+    margin: float = Field(default=0.02, ge=0)
+    resolution: float = Field(default=30.0, gt=0)
+    cache: str | None = None
+
+    @model_validator(mode="after")
+    def _one_input(self):
+        if (self.dem is None) == (self.tiles is None):
+            raise ValueError(f"source {self.id!r}: give exactly one of 'dem' or 'tiles'")
+        return self
+
+
 class UnionSource(_SourceBase):
     """The features of several vector sources of this file as one source."""
 
@@ -139,7 +182,7 @@ class UnionSource(_SourceBase):
 
 
 Source = Annotated[
-    FileSource | BdcSource | MapbiomasSource | ProdesSource | ClassifiedSource | UnionSource,
+    FileSource | BdcSource | MapbiomasSource | ProdesSource | ClassifiedSource | OsmSource | DemSource | UnionSource,
     Field(discriminator="type"),
 ]
 
@@ -174,7 +217,7 @@ class PipelineConfig(_Strict):
     grid: GridConfig | None = None
     extent: list[float] | None = Field(default=None, min_length=4, max_length=4)
     """``[min_lon, min_lat, max_lon, max_lat]`` (WGS84) a sources-only file reads
-    windowed sources over (classified maps, BDC, MapBiomas, PRODES)."""
+    windowed sources over (classified maps, BDC, MapBiomas, PRODES, OpenStreetMap, DEM)."""
     sources_from_catalog: bool = False
     """Let [[derive]] blocks use sources this file does not declare, registered
     in the workspace's catalog by another pipeline file. Off by default, so a
@@ -190,14 +233,14 @@ def _grid_or_extent(self):
 
         # Apenas fontes dinâmicas em janela na nuvem precisam de extent.
         # Fontes 'file' e 'union' já têm extensão definida por seus arquivos locais.
-        windowed_types = {"bdc", "mapbiomas", "prodes", "classified"}
+        windowed_types = {"bdc", "mapbiomas", "prodes", "classified", "osm", "dem"}
         needs_extent = any(
             getattr(s, "type", None) in windowed_types for s in self.source
         )
 
         if self.grid is None and self.extent is None and needs_extent:
             raise ValueError(
-                "a sources-only file with windowed sources (BDC, MapBiomas, PRODES)"
+                "a sources-only file with windowed sources (BDC, MapBiomas, PRODES, OpenStreetMap, DEM)"
                 " needs `extent` (or a [grid])"
             )
         return self