La interfaz está localizada; las descripciones y documentos de las Skills se traducen automáticamente con IA a su idioma. El código, los comandos, rutas y términos técnicos se mantienen en el original, y prevalece el documento del autor.
Explorar Skills
Datos e investigación · SKILL.md

gemgis

gemgis

Spatial data processing for geological modelling with GemPy. Use when the agent needs to: (1) Prepare spatial data for GemPy models, (2) Extract interface points from geological maps, (3) Process orientations/dip measurements, (4) Sample DEMs along profiles or cross-sections, (5) Convert between GIS formats and GemPy inputs, (6) Clip/transform vector/raster…

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentación de la Skill

GitHub

name: gemgis description: | Spatial data processing for geological modelling with GemPy. Use when the agent needs to: (1) Prepare spatial data for GemPy models, (2) Extract interface points from geological maps, (3) Process orientations/dip measurements, (4) Sample DEMs along profiles or cross-sections, (5) Convert between GIS formats and GemPy inputs, (6) Clip/transform vector/raster data for modeling, (7) Create model extents from geospatial bounds. license: MIT metadata: version: "1.0.2" author: Geoscience Skills tags: '["GIS", "Geospatial", "Data Preparation", "DEM", "Geological Modelling"]' dependencies: '["gemgis>=1.1.9", "geopandas", "rasterio"]' complements: '["gempy", "loopstructural", "pyvista"]' workflow_role: processing skill_type: domain

GIS contacts and DEM elevations

Use GemGIS to extract geological observations from spatial data. Establish the horizontal CRS, vertical units/datum, geometry type and missing-data policy before sampling. Model coordinates must share a projected metre CRS; converting horizontal coordinates does not transform elevations or their datum.

Extract contact vertices

The following function accepts a single-band DEM already in a metre projected CRS, with elevations in metres. dem is an open Rasterio dataset, not a path. GemGIS repeats attributes when a line becomes multiple points: do not overwrite those attributes using the original feature index.

import gemgis as gg
from rasterio.transform import rowcol
import numpy as np
from pyproj import CRS

def contacts_at_dem(contacts, dem):
    crs = CRS.from_user_input(dem.crs)
    if (contacts.crs is None or not crs.is_projected or dem.count != 1
            or any(not np.isclose(a.unit_conversion_factor, 1.) for a in crs.axis_info[:2])):
        raise ValueError('Require a known vector CRS and a single-band metric DEM')
    if contacts.empty or not contacts.is_valid.all() or contacts.geometry.has_z.any():
        raise ValueError('Require valid nonempty 2D contact geometry')
    if 'formation' not in contacts or contacts['formation'].isna().any():
        raise ValueError('Formation labels are required')
    points = gg.vector.extract_xy(contacts.to_crs(dem.crs))
    xy = points[['X', 'Y']].to_numpy()
    rows, cols = np.asarray(rowcol(dem.transform, xy[:, 0], xy[:, 1]))
    if np.any((rows < 0) | (rows >= dem.height) | (cols < 0) | (cols >= dem.width)):
        raise ValueError('Vertices outside DEM')
    samples = np.ma.vstack(list(dem.sample(xy, indexes=1, masked=True)))[:, 0]
    if np.ma.getmaskarray(samples).any() or not np.isfinite(samples.data).all():
        raise ValueError('Missing DEM elevations')
    return gg.vector.extract_xyz(gdf=points, dem=dem)[['X', 'Y', 'Z', 'formation']]

Open with with rasterio.open(dem_path) as dem: and call this function while the dataset is open. Pixel sampling uses the containing cell; it is not bilinear interpolation. Do not fill NoData or outside-extent samples with zero.

Orientations and helper

Dip is in degrees, 0–90. Output azimuth is dip direction clockwise from DEM grid north, wrapped to [0, 360). Convert strike with (strike + 90) % 360 only when the source explicitly follows the right-hand rule. Preserve supplied polarity (+1/−1); the helper uses +1 only when the column is absent. Never clamp invalid dip. Declare --azimuth-reference dem-grid for bearings already relative to the DEM grid, or true-north for geographic bearings. The latter applies local meridian convergence only on locally conformal projections; other projections are rejected. Magnetic bearings need a documented declination correction first.

The preparation helper validates these rules, reprojects vectors to the DEM CRS, converts declared ft elevations to metres, and writes CSVs plus spatial_metadata.json into a fresh output directory. Its --target-crs is an assertion that must match the DEM; resample/reproject a raster separately when needed.

python scripts/prepare_gempy_data.py --contacts contacts.gpkg \
  --orientations orientations.gpkg --azimuth-reference true-north --dem dem.tif --dem-z-unit m \
  --vertical-datum "documented source datum" --output-dir prepared

Resolve this command relative to the installed skill directory. The output schema uses X/Y/Z and formation, plus dip/azimuth/polarity for orientations; pass these to the current modelling library's input API rather than assuming all GemPy versions accept identical keyword arguments.

Conditional operations

Read profiles and extraction for sampled profiles and orientation conventions. Read CRS, clipping and raster handling for bounds order, masks, reprojection and provenance. Existing 3D geometry needs an explicit choice between measured Z and DEM Z before using this helper.

Verification scope

GemGIS 1.1.9, GeoPandas 1.1.4 and Rasterio 1.4.4 were checked with a real GeoTIFF/GeoPackage fixture: unequal vertex counts and nonconsecutive indices, CRS conversion, feet elevations, NoData, outside points, orientations and CLI CSV readback. Synthetic elevations test data handling, not field DEM accuracy or vertical-datum transformation.

Official GemGIS source, Rasterio sampling API, checked 2026-09-14.

Skills guardadas