File size: 3,715 Bytes
a85e26f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 | from pathlib import Path
from typing import Iterable, List
import cv2
import json
import numpy as np
import pandas as pd
def norm_by_x(df):
df['x'] = df['x'] - df['x'].min()
df['y'] = df['y'] - df['y'].min()
maxx = df['x'].max()
df['x'] /= maxx
df['y'] /= maxx
return df
def csv2dfs(filenames:Iterable[str], rotate_func:callable=None) -> List[pd.DataFrame]:
''' Extract x,y coordinates from a .csv/.xlsx file. Each file may include multiple outlines.'''
# if multiple x,y columns in the csv/xlsx file
def _sheet2dfs(df, df_name):
dfs_local = []
if 'Unnamed: 0' in df.columns:
df = df.drop(columns=['Unnamed: 0'])
column_pairs = [(df.columns[i], df.columns[i+1]) for i in range(0, len(df.columns)-1, 2)]
for x_col, y_col in column_pairs:
shoe_id = x_col
x, y = df[x_col].iloc[1:].astype(float), df[y_col].iloc[1:].astype(float)
if rotate_func: x, y = rotate_func(x, y)
shoe_df = pd.DataFrame({'x': x, 'y': y}).dropna()
shoe_df.name = shoe_id
dfs_local.append(shoe_df)
else:
df.name = df_name
dfs_local += [df]
return dfs_local
dfs = []
for filename in filenames:
filename = Path(filename)
if filename.suffix.lower() == '.csv':
df = pd.read_csv(filename)
dfs += _sheet2dfs(df, filename.name)
elif filename.suffix.lower() == '.xlsx':
xls = pd.ExcelFile(filename)
for sheet in xls.sheet_names:
df = pd.read_excel(xls, sheet)
dfs += _sheet2dfs(df, sheet)
elif filename.suffix.lower() == '.json':
vgg_json = json.load(filename.open())
for _,v in vgg_json.items():
fn = v['filename']
for region in v['regions']:
if region['shape_attributes']['name'] != 'polygon': continue
xs,ys = region['shape_attributes']['all_points_x'], region['shape_attributes']['all_points_y']
df = pd.DataFrame({'x':xs, 'y':ys})
df.name = fn
dfs.append(df)
break
return dfs
def coordsdf2image(df:pd.DataFrame, target_height:int=256, margin:float=0.1) -> np.ndarray:
''' Render a footprint outline from x,y coordinates in a DataFrame into a binary image (filled shape). '''
x_coords, y_coords = df['x'].values, df['y'].values
# Compute the bounding box of the footprint
x_min, x_max = x_coords.min(), x_coords.max()
y_min, y_max = y_coords.min(), y_coords.max()
width = x_max - x_min
height = y_max - y_min
# Compute canvas width proportional to the footprint's aspect ratio
scale = target_height / (1 + 2 * margin) / height
target_width = int(scale * width + 2 * margin * target_height)
image = np.zeros((target_height, target_width), dtype=np.uint8)
# Scale coordinates to fit within the canvas
x_scaled = ((x_coords - x_min) * scale).astype(np.int32)
y_scaled = ((y_coords - y_min) * scale).astype(np.int32)
# Compute the size of the scaled footprint
scaled_width = x_scaled.max() - x_scaled.min()
scaled_height = y_scaled.max() - y_scaled.min()
# Compute and apply offsets for centering
x_scaled += (target_width - scaled_width) // 2 - x_scaled.min()
y_scaled += (target_height - scaled_height) // 2 - y_scaled.min()
contour = np.array([np.stack((x_scaled, y_scaled), axis=-1)], dtype=np.int32)
cv2.fillPoly(image, contour, color=255) # Fill the shape with white (255)
return image |