Source code for pygmt.src.blockmedian
"""
blockmedian - Block average (x,y,z) data tables by median estimation.
"""
import pandas as pd
from pygmt.clib import Session
from pygmt.exceptions import GMTInvalidInput
from pygmt.helpers import (
GMTTempFile,
build_arg_string,
data_kind,
dummy_context,
fmt_docstring,
kwargs_to_strings,
use_alias,
)
[docs]@fmt_docstring
@use_alias(I="spacing", R="region", V="verbose")
@kwargs_to_strings(R="sequence")
def blockmedian(table, outfile=None, **kwargs):
r"""
Block average (x,y,z) data tables by median estimation.
Reads arbitrarily located (x,y,z) triples [or optionally weighted
quadruples (x,y,z,w)] from a table and writes to the output a median
position and value for every non-empty block in a grid region defined by
the region and spacing arguments.
Full option list at :gmt-docs:`blockmedian.html`
{aliases}
Parameters
----------
table : pandas.DataFrame or str
Either a pandas dataframe with (x, y, z) or (longitude, latitude,
elevation) values in the first three columns, or a file name to an
ASCII data table.
spacing : str
*xinc*\[\ *unit*\][**+e**\|\ **n**]
[/*yinc*\ [*unit*][**+e**\|\ **n**]].
*xinc* [and optionally *yinc*] is the grid spacing.
region : str or list
*xmin/xmax/ymin/ymax*\[\ **+r**\][**+u**\ *unit*].
Specify the region of interest.
outfile : str
Required if ``table`` is a file. The file name for the output ASCII
file.
{V}
Returns
-------
output : pandas.DataFrame or None
Return type depends on whether the outfile parameter is set:
- pandas.DataFrame table with (x, y, z) columns if outfile is not set
- None if outfile is set (filtered output will be stored in outfile)
"""
kind = data_kind(table)
with GMTTempFile(suffix=".csv") as tmpfile:
with Session() as lib:
if kind == "matrix":
if not hasattr(table, "values"):
raise GMTInvalidInput(f"Unrecognized data type: {type(table)}")
file_context = lib.virtualfile_from_matrix(table.values)
elif kind == "file":
if outfile is None:
raise GMTInvalidInput("Please pass in a str to 'outfile'")
file_context = dummy_context(table)
else:
raise GMTInvalidInput(f"Unrecognized data type: {type(table)}")
with file_context as infile:
if outfile is None:
outfile = tmpfile.name
arg_str = " ".join([infile, build_arg_string(kwargs), "->" + outfile])
lib.call_module(module="blockmedian", args=arg_str)
# Read temporary csv output to a pandas table
if outfile == tmpfile.name: # if user did not set outfile, return pd.DataFrame
result = pd.read_csv(tmpfile.name, sep="\t", names=table.columns)
elif outfile != tmpfile.name: # return None if outfile set, output in outfile
result = None
return result