Large data & batch operators

`batch.run`#

An operator run on many rasters in parallel (landschaft's batch, GuidosToolbox's Batch Process): each worker reads one raster, runs the operator on it and writes what it gives to the output folder, named after the input and the operator: maps [].tif, tables .csv, texts .txt, and _.json with the parameters, the files, the numbers and the time (an input whose JSON exists is skipped unless overwrite). MSPA with GuidosToolbox's naming. A row per input: ok, skipped (done before), seconds, files written; the errors in the report. A folder given to any node that takes a map does the same, its maps seen and its settings in its pane (Make it a chain of nodes, in the app).

Output	Type	What it is
`table`	table	a row per input: its number, ok (1/0), skipped (1/0: done before), seconds, files
`report`	text	each input and its outcome

Setting	Type	What it is
`operator`required	text	the operator, e.g. landscape.mspa
`inputs`required	any	a folder, a glob pattern (tiles/*.tif) or a list of rasters
`output_dir`required	text	the folder for the results
`params`	any	the operator's parameters ({"edge_width": 2})
`input`	text	the operator's input taking the raster (default: its first map)
`workers`	integer	rasters at the same time (default: the threads)
`overwrite`	bool	recompute inputs already done
`recursive`	bool	the rasters of the folder's sub-folders too

batch.run#

`batch.run`#