feat: Sentinel-1 GRD pipeline support

.gitignore

@@ -0,0 +1,63 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+*.egg-info/
+dist/
+build/
+*.egg
+
+# Testing
+.pytest_cache/
+.coverage
+htmlcov/
+.tox/
+
+# Type checking
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Linting
+.ruff_cache/
+
+# Environment
+.env
+.env.local
+.env.*.local
+.venv
+env/
+venv/
+ENV/
+*.local.yaml
+
+# IDE
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# Jupyter
+.ipynb_checkpoints/
+
+# Kubernetes
+.work/
+kubeconfig*
+.kube/
+
+# Temporary files
+*.tmp
+*.log
+runs/
+generated/
+.archive/
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Project-specific
+*.zarr
+out/

README.md

@@ -23,18 +23,24 @@ kubectl get wf -n devseed -w
 [![Python](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)
 [![Tests](https://github.com/EOPF-Explorer/data-pipeline/workflows/Tests/badge.svg)](https://github.com/EOPF-Explorer/data-pipeline/actions)
 
+- **Multi-sensor support**: Sentinel-1 GRD and Sentinel-2 L2A
 - STAC item registration with retry logic
-- GeoZarr format conversion
-- Cloud-native workflows
+- GeoZarr format conversion with cloud-optimized overviews
+- Cloud-native workflows with Argo
+- Interactive visualization with TiTiler
 
 ## What It Does
 
-Transforms Sentinel-2 satellite data into web-ready visualizations:
+Transforms Sentinel satellite data into web-ready visualizations:
 
 **Input:** STAC item URL → **Output:** Interactive web map (~5-10 min)
 
 **Pipeline:** Convert (5 min) → Register (30 sec) → Augment (10 sec)
 
+**Supported sensors:**
+- **Sentinel-1** L1 GRD: SAR backscatter (VH/VV polarizations)
+- **Sentinel-2** L2A: Multispectral reflectance (10m/20m/60m)
+
 ## Quick Start
 
 📖 **New to the project?** See [GETTING_STARTED.md](GETTING_STARTED.md) for complete setup (15 min).
@@ -262,10 +268,12 @@ pytest -v -k e2e   # End-to-end tests only
 1. **Edit workflow:** `workflows/template.yaml`
 2. **Update scripts:** `scripts/*.py`
 3. **Test locally:** `pytest tests/ -v`
-4. **Build image:** `docker build -t ghcr.io/eopf-explorer/data-pipeline:dev -f docker/Dockerfile .`
+4. **Build image:** `docker buildx build --platform linux/amd64 -t ghcr.io/eopf-explorer/data-pipeline:dev -f docker/Dockerfile . --push`
 5. **Deploy:** `kubectl apply -f workflows/template.yaml -n devseed`
 6. **Monitor:** `kubectl get wf -n devseed -w`
 
+⚠️ **Important:** Always use `--platform linux/amd64` when building images for Kubernetes clusters.
+
 See [CONTRIBUTING.md](CONTRIBUTING.md) for coding standards and development workflow.
 
 ## License

docker/Dockerfile

@@ -1,3 +1,4 @@
+# Build for linux/amd64: docker buildx build --platform linux/amd64 -t <tag> . --push
 FROM python:3.11-slim
 
 # System dependencies (including GDAL for rasterio)
@@ -18,13 +19,13 @@ RUN pip install -U pip uv
 # Cachebust for data-model installation (change timestamp to force fresh install)
 ARG CACHEBUST=2025-10-06T11:00:00Z
 
-# Install eopf-geozarr from minimal fix branch
-# Includes critical set_spatial_dims() fix before write_crs() calls
+# Install eopf-geozarr from fix/s1-encoding-conflict branch (temporary until merged)
 RUN uv pip install --system --no-cache \
-    git+https://github.com/EOPF-Explorer/data-model.git@fix/spatial-dims-minimal \
+    git+https://github.com/EOPF-Explorer/data-model.git@fix/s1-encoding-conflict \
     pystac>=1.10.0 \
     httpx>=0.27.0 \
-    boto3>=1.34.0
+    boto3>=1.34.0 \
+    tenacity>=8.0.0
 
 # Force fresh copy of scripts (invalidate cache)
 ARG SCRIPTS_VERSION=2025-10-06T02:05:00Z

scripts/augment_stac_item.py

@@ -47,6 +47,69 @@ def _encode_quicklook_query() -> str:
 
 DEFAULT_QUICKLOOK_QUERY = _encode_quicklook_query()
 
+
+def _get_s1_polarization(item: Item) -> str:
+    """Extract first available polarization from S1 item assets.
+
+    Args:
+        item: PySTAC Item with S1 assets
+
+    Returns:
+        Uppercase polarization code (VH, VV, HH, or HV). Defaults to VH.
+    """
+    for pol in _S1_POLARIZATIONS:
+        if pol in item.assets:
+            return pol.upper()
+    return "VH"
+
+
+def _encode_s1_preview_query(item: Item) -> str:
+    """Generate S1 GRD preview query for TiTiler.
+
+    S1 GRD structure in converted GeoZarr:
+    /S01SIWGRD_{timestamp}_{id}_VH/measurements with grd variable
+
+    TiTiler needs the full path to the measurements group with the grd variable.
+
+    Args:
+        item: PySTAC Item with S1 GRD data
+
+    Returns:
+        Query string for TiTiler (variables, bidx, rescale)
+    """
+    pol = _get_s1_polarization(item)
+    asset = item.assets.get(pol.lower())
+
+    if not asset or not asset.href:
+        # Fallback to simple path
+        pairs = [
+            ("variables", "/measurements:grd"),
+            ("bidx", "1"),
+            ("rescale", "0,219"),
+        ]
+        return "&".join(f"{key}={urllib.parse.quote_plus(value)}" for key, value in pairs)
+
+    # Extract group path from asset href
+    # Example: s3://.../S01SIWGRD_..._VH/measurements -> /S01SIWGRD_..._VH/measurements:grd
+    href = asset.href
+    if ".zarr/" in href:
+        # Extract path after .zarr/
+        zarr_path = href.split(".zarr/")[1]
+        # zarr_path is like: S01SIWGRD_..._VH/measurements
+        # Build variable reference: /S01SIWGRD_..._VH/measurements:grd
+        variable_path = f"/{zarr_path}:grd"
+    else:
+        # Fallback
+        variable_path = "/measurements:grd"
+
+    pairs = [
+        ("variables", variable_path),
+        ("bidx", "1"),
+        ("rescale", "0,219"),  # Typical S1 GRD range
+    ]
+    return "&".join(f"{key}={urllib.parse.quote_plus(value)}" for key, value in pairs)
+
+
 _ALLOWED_SCHEMES = {"http", "https"}
 _USER_AGENT = "augment-stac-item/1.0"
 _DEFAULT_TIMEOUT = float(os.getenv("HTTP_TIMEOUT", "30"))
@@ -65,6 +128,14 @@ def _encode_quicklook_query() -> str:
 _S2_DATASET_KEYS = ("SR_10m", "SR_20m", "SR_60m")
 _S2_QUICKLOOK_KEYS = ("TCI_10m", "TCI", "TCI_20m")
 
+_S1_COLLECTION_ID = "sentinel-1-l1-grd"
+_S1_POLARIZATIONS = ("vh", "vv", "hh", "hv")
+
+
+def _is_s1_collection(collection_id: str) -> bool:
+    """Check if collection is Sentinel-1 GRD."""
+    return collection_id.startswith("sentinel-1-l1-grd")
+
 
 def _coerce_epsg(value: Any) -> int | None:
     if isinstance(value, bool):
@@ -462,17 +533,27 @@ def add_visualization_links(
     item.links = [link for link in item.links if link.rel not in filtered_rels]
     item_id = item.id
     viewer_href = f"{base_raster_url}/collections/{coll}/items/{item_id}/viewer"
-    asset_key = _select_preview_asset(item)
-    preview_asset = item.assets.get(asset_key) if asset_key else None
-    is_quicklook = _is_quicklook_asset(preview_asset)
-    default_query = DEFAULT_QUICKLOOK_QUERY if is_quicklook else DEFAULT_TRUE_COLOR_QUERY
+
+    # Determine preview query based on collection type
+    asset_key: str | None
+    if _is_s1_collection(coll):
+        # Sentinel-1: Use GRD polarization preview
+        default_query = _encode_s1_preview_query(item)
+        xyz_title = os.getenv("PREVIEW_XYZ_TITLE", f"GRD {_get_s1_polarization(item)}")
+        asset_key = _get_s1_polarization(item).lower()  # vh or vv
+    else:
+        # Sentinel-2: Use quicklook or true color
+        asset_key = _select_preview_asset(item)
+        preview_asset = item.assets.get(asset_key) if asset_key else None
+        is_quicklook = _is_quicklook_asset(preview_asset)
+        default_query = DEFAULT_QUICKLOOK_QUERY if is_quicklook else DEFAULT_TRUE_COLOR_QUERY
+        xyz_title = os.getenv("PREVIEW_XYZ_TITLE", "True Color Image (10m)")
+
     xyz_query = _resolve_preview_query(
         os.getenv("PREVIEW_XYZ_QUERY"),
         default_query=default_query,
     )
 
-    xyz_title = os.getenv("PREVIEW_XYZ_TITLE", "True Color Image (10m)")
-
     def _add_link(rel: str, target: str, media_type: str, title: str | None = None) -> None:
         item.add_link(
             Link(

scripts/get_zarr_url.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python3
+import json
+import sys
+from urllib.request import urlopen
+
+
+def get_zarr_url(stac_item_url: str) -> str:
+    with urlopen(stac_item_url) as response:
+        item = json.loads(response.read())
+
+    assets = item.get("assets", {})
+
+    # Priority: product, zarr, then any .zarr asset
+    for key in ["product", "zarr"]:
+        if key in assets:
+            href = assets[key].get("href")
+            if href:
+                return str(href)
+
+    # Fallback
+    for asset in assets.values():
+        href = asset.get("href", "")
+        if ".zarr" in href:
+            return str(href)
+
+    raise RuntimeError("No Zarr asset found")
+
+
+if __name__ == "__main__":
+    print(get_zarr_url(sys.argv[1]))