asp_plot.selections =================== .. py:module:: asp_plot.selections .. autoapi-nested-parse:: Reproducible "figure selections" for asp_plot reports. When re-processing the same scene with different ASP parameters, the diagnostic figures silently change *what they show* between runs: a fresh ICESat-2 request returns a slightly different point set, the "best" profile track flips, the best/worst agreement segments move, and the detailed-hillshade clip boxes are re-selected from the (re-processed) intersection-error raster. That makes before/after comparison impossible. This module persists every non-deterministic selection a run made to a small YAML sidecar next to the report, and reads it back so a later run can replay the same choices. It deliberately imports nothing from ``report.py`` / ``fpdf`` so it stays safe to use from notebooks. See issue: https://github.com/uw-cryo/asp_plot/issues/121 Attributes ---------- .. autoapisummary:: asp_plot.selections.HILLSHADE_CLIP_LABELS asp_plot.selections.SELECTIONS_SCHEMA_VERSION asp_plot.selections.logger Classes ------- .. autoapisummary:: asp_plot.selections.FigureSelections Functions --------- .. autoapisummary:: asp_plot.selections.bbox_to_pixel_offset asp_plot.selections.pixel_window_to_bbox asp_plot.selections.read_selections_yaml asp_plot.selections.reproject_bbox asp_plot.selections.write_selections_yaml Module Contents --------------- .. py:class:: FigureSelections Container for the reproducible selections made while building a report. All nested values are plain JSON/YAML-serializable types (dicts, lists, numbers, strings) so the object round-trips cleanly through YAML. .. attribute:: asp_plot_version asp_plot version that wrote the file (informational). :type: str or None .. attribute:: dem_filename DEM the selections were computed against (informational). :type: str or None .. attribute:: map_crs Map CRS string used by the report (informational). :type: str or None .. attribute:: detailed_hillshade ``{"subset_km": float, "intersection_error_percentiles": [..], "dem_crs": "EPSG:XXXX", "clips": [{"label": str, "bbox": [xmin, ymin, xmax, ymax], "pixel_offset": [row, col]}, ...]}``. ``bbox`` is in ``dem_crs`` map coordinates (robust to a re-gridded DEM). :type: dict or None .. attribute:: icesat2 ``{"request": {..}, "parquet_cache": {key: path}, "profile_track": {"rgt": int, "cycle": int, "spot": int}, "segments": {"best": {..}, "worst": {..}}}``. Omitted (None) for planetary DEMs. :type: dict or None .. py:method:: from_dict(data) :classmethod: Build a FigureSelections from a parsed YAML/JSON dict. .. py:method:: to_dict() Return a plain dict suitable for YAML serialization. .. py:attribute:: asp_plot_version :type: Optional[str] :value: None .. py:attribute:: dem_filename :type: Optional[str] :value: None .. py:attribute:: detailed_hillshade :type: Optional[dict] :value: None .. py:attribute:: icesat2 :type: Optional[dict] :value: None .. py:attribute:: map_crs :type: Optional[str] :value: None .. py:attribute:: schema_version :type: int :value: 1 .. py:function:: bbox_to_pixel_offset(transform, bbox) Convert a map-coordinate bbox to a top-left pixel offset (row, col). The window size is *not* returned: on reuse the subset size is recomputed from ``subset_km`` and the current DEM's GSD so the ground footprint stays constant even if the DEM resolution changed. Only the top-left anchor is needed. :param transform: Geotransform of the DEM being clipped on reuse. :type transform: affine.Affine :param bbox: ``[xmin, ymin, xmax, ymax]`` in the DEM's CRS. :type bbox: sequence of float :returns: ``(row, col)`` top-left pixel offset (clamped to be non-negative). :rtype: tuple of int .. py:function:: pixel_window_to_bbox(transform, row, col, n_rows, n_cols) Convert a pixel window (top-left row/col + size) to a map-coordinate bbox. :param transform: Raster geotransform (``raster.ds.transform``). :type transform: affine.Affine :param row: Top-left pixel of the window. :type row: int :param col: Top-left pixel of the window. :type col: int :param n_rows: Window height/width in pixels. :type n_rows: int :param n_cols: Window height/width in pixels. :type n_cols: int :returns: ``[xmin, ymin, xmax, ymax]`` in the raster's CRS. :rtype: list of float .. py:function:: read_selections_yaml(path) Read a FigureSelections from a YAML file. :param path: Path to a previously written selections file. :type path: str :rtype: FigureSelections .. py:function:: reproject_bbox(bbox, src_crs, dst_crs) Reproject a map-coordinate bbox from one CRS to another. Used when replaying detailed-hillshade clips against a DEM in a different CRS than the run that wrote them (e.g. a mapprojected vs. non-mapprojected stereo variant of the same scene, which can land in different projections). Returns the input unchanged when either CRS is missing or they are equal. :param bbox: ``[xmin, ymin, xmax, ymax]`` in ``src_crs``. :type bbox: sequence of float :param src_crs: CRS the bbox is currently expressed in. :type src_crs: str or rasterio.crs.CRS or None :param dst_crs: Target CRS (the DEM being clipped on reuse). :type dst_crs: str or rasterio.crs.CRS or None :returns: ``[xmin, ymin, xmax, ymax]`` in ``dst_crs``. :rtype: list of float .. py:function:: write_selections_yaml(path, selections) Write a FigureSelections to a YAML file. :param path: Destination path (e.g. ``_figure_selections.yml``). :type path: str :param selections: Selections to serialize. :type selections: FigureSelections .. py:data:: HILLSHADE_CLIP_LABELS :value: ['low', 'medium', 'high'] .. py:data:: SELECTIONS_SCHEMA_VERSION :value: 1 .. py:data:: logger