Python 里的 `moraine.cli`¶

import moraine.cli as mc 给出基于文件的函数：与命令相同的操作，在 Python 里以相同的参数调用（mc.amp_disp('raw/rslc.zarr', 'ps/ras_adi.zarr', cuda=True)），外加查看器和几个不是命令的辅助函数。

import moraine.cli as mc
logger = mc.get_logger()                       # 命令行打印的那些日志
mc.amp_disp('raw/rslc.zarr', 'ps/ras_adi.zarr', cuda=True)
mc.view('ps/ras_adi_pyramid', label='amplitude dispersion index', clim=(0, 1))

view¶

view ¶

view(data, x=None, y=None, show=None, dates=None, series=None, polygons: str = None, cmap=None, clim: tuple = None, opacity: float = 1.0, label: str = None, size: tuple = None, terrain=None, bounds: tuple = None, resolution: float = None, image_pairs: str = None, sliders: dict = None, crs: str = None)

View a raster or point cloud: an interactive map in a notebook (the last expression of a cell) or a PNG (.png(path)); a * b overlays views on one map, a + b shows maps side by side with linked zoom and pan and shared sliders.

On the map: zoom and pan read only what is on screen; sliders choose the image of a stack; the cursor shows the values of all layers; click a pixel or point to plot its time series, double click one to make it the reference of the time series; draw polygons with the polygon button. In Python, .selected, .reference, .polygons and .index (slider values) follow the map. Point clouds are drawn as individual points when zoomed in; web mercator coordinates are drawn north up over a base map, or over the terrain in 3D with terrain.

Parameters:

Name	Type	Description	Default
`data`	`str or ndarray`	a pyramid directory made by `ras_pyramid` or `pc_pyramid`; a raster, (nlines, width[, n[, m]]), as array or zarr path (read into memory, at most 512 MiB); or point data, (n_points[, n[, m]]), with `x` and `y`	required
`x`	`str or ndarray`	x coordinate (range or web mercator x) of point data that is not a pyramid, (n_points,)	`None`
`y`	`str or ndarray`	y coordinate (azimuth or web mercator y) of point data that is not a pyramid, (n_points,)	`None`
`show`	`str or callable`	what to show of a stack: 'phase', 'intf_0' (interferograms with the first image), 'intf_seq' (sequential interferograms), 'intf_all' (any pair, sliders ref and sec), 'coh' / 'coh_abs' (phase / magnitude of compressed coherence (..., n_pairs), sliders ref and sec), or a function `f(v, sliders)` of the visible stack `v` (index it like a numpy array, e.g. `v[..., ref]`) whose other arguments are sliders named like them, e.g. `lambda v, ref, sec: np.angle(v[..., ref] np.conj(v[..., sec]))`; the phase for complex data, the values otherwise by default	`None`
`dates`	`list or str`	date of each image, or a toml file with `dates` (e.g. of `load_gamma_metadata`), shown with the sliders and on the time axis	`None`
`series`	`str or ndarray`	time series plotted for a clicked pixel or point: (nlines, width, n) for rasters, (n_points, n) in the order of the points; the stack itself by default	`None`
`polygons`	`str`	GeoJSON file of the polygons drawn on the map: read if it exists, written on every change; use it with `polygon_mask`	`None`
`cmap`	`str or list`	matplotlib colour map name or list of colours; cyclic for phases, viridis otherwise by default	`None`
`clim`	`tuple`	(min, max) of the colours; (-pi, pi) for phases, the 1 % - 99 % range otherwise by default	`None`
`opacity`	`float`	opacity of the layer, 0 - 1	`1.0`
`label`	`str`	name of the layer; the pyramid directory name by default	`None`
`size`	`tuple`	(width, height) of the map in screen pixels; by default the map takes the width of the notebook (at most 700 pixels high) with the aspect of the scene (at most 1:4). Drag the lower right corner of a map to resize it	`None`
`terrain`	`bool or str`	3D view over the terrain in a notebook, web mercator layers only: True for the public elevation tiles (AWS Terrain Tiles: SRTM, EU-DEM, 3DEP, ... heights of about 30 m, zoom levels up to 15, fetched by the browser like the base map), or the URL template `'https://.../{z}/{x}/{y}.png'` of another service of Terrarium encoded tiles. The layers and the satellite base map are draped over the terrain; the right mouse button (or ctrl + drag) tilts and rotates the view, which zooms two levels further than the 2D map. Polygons are shown but drawn in 2D views; `.png` stays the 2D image	`None`
`bounds`	`tuple`	raster only: (x0, y0, xm, ym), coordinates of the centres of the first and the last pixel; pixel (i, j) at range j, azimuth i by default	`None`
`resolution`	`float`	point data only: cell size of the rasterized points; 1 for integer coordinates	`None`
`image_pairs`	`str`	image pair file (`image_pairs`) of compressed coherence for 'coh' / 'coh_abs'; all pairs by default	`None`
`sliders`	`dict`	number of values of each slider of a `show` function, e.g. {'ref': 17, 'sec': 17}; from the stack shape or `dates` by default	`None`
`crs`	`str`	'grid' (radar grid, azimuth down) or 'web_mercator' (metres, north up); from the coordinates by default	`None`

Returns:

Type	Description
`layer`	displayed in a notebook; combine with `*` and `+`; `.png(path)` saves an image

view 背后的地图部件是 moraine.cli.TileView（anywidget 加 Leaflet；瓦片由内核渲染，通过 notebook 通道发送，所以不需要服务器或端口转发，决策 0018）。moraine view PYRAMID ... -o view.ipynb 生成一个由这种视图组成的 notebook。

get_logger¶

get_logger ¶

get_logger(logfile: str = None)

get logger for decorrelation cli application