Check if string coordinates represent a polygon using Pandas-CodePudding

Using Pandas, how to check if a valid polygon geometry can be created using coordinates from a string?

Input example:

import pandas as pd
coord_series = pd.Series([
    '-33, 50, -30, 38, -40, 27, -33.0, 50.5',
    '-xx, xx, -xx, xx, -xx, xx, -xxxx, xxxx',
    None,
    '-10',
    '-10, 20, -30, 40, -50, 60, -70',
    '-11, 11, -11, 11, -11, 11, -11.1, 11.1'
])

Only the first string forms a valid polygon.

A function is needed which accepts one Series object and outputs one Series object.

CodePudding user response：

The below code looks like a lot, but I couldn't create a shorter solution. shapely.geometry.Polygon.is_valid seems needed to do the assessment, but it's not enough, as sometimes it throws an error.

from shapely.geometry import Polygon

def validate_polygon(s: pd.Series) -> pd.Series:
    def is_digit(lst):
        return [e.strip().lstrip('-').replace('.', '', 1).isdigit() if e else None for e in lst]
    coords = s.str.split(',')
    even_cnt = coords.map(lambda x: x if x and len(x) % 2 == 0 else None)
    point_cnt_ge3 = even_cnt.map(lambda x: x if x and len(x) >= 6 else None)
    all_digits = point_cnt_ge3.map(lambda x: x if x and all(is_digit(x)) else None)
    polygon_shell = all_digits.map(lambda x: list(zip(*[iter(map(float, x))]*2)) if x else None)
    is_valid = polygon_shell.map(lambda x: Polygon(x).is_valid if x else False)
    return is_valid

print(validate_polygon(coord_series))
# 0     True
# 1    False
# 2    False
# 3    False
# 4    False
# 5    False
# dtype: bool

CodePudding user response：

We can use shapely.geometry.Polygon.is_valid together with try...except:

from shapely.geometry import Polygon

def validate_polygon(ser: pd.Series) -> pd.Series:
    def is_valid(s):
        try:
            p_shell = list(zip(*[iter(map(float, s.split(',')))]*2))
            return Polygon(p_shell).is_valid if p_shell != [] else False
        except (ValueError, AttributeError):
            return False
    return ser.map(is_valid)

print(validate_polygon(coord_series))
# 0     True
# 1    False
# 2    False
# 3    False
# 4    False
# 5    False
# dtype: bool