final commit

2025_03_21/full.py

@@ -0,0 +1,93 @@
+import pandas as pd
+import geopandas as gpd
+from shapely.geometry import Point
+import numpy as np
+import pyproj
+import math
+import sys
+
+# Function to convert WGS84 to Lambert 93 for the entire DataFrame
+def convert_to_lambert93(row, transf):
+    if(math.isnan(row[x_col]) or math.isnan(row[y_col])):
+        x, y = transf.transform(row[x_inter_col], row[y_inter_col])
+        row[x_inter_col] = x
+        row[y_inter_col] = y
+    return row
+
+def point_position_on_line(pt):
+    return line.project(pt)
+
+# Load CSV
+df = pd.read_csv(sys.argv[1], encoding='UTF-8', on_bad_lines='skip', delimiter=';')
+
+x_col = sys.argv[4]
+y_col = sys.argv[5]
+
+x_inter_col = sys.argv[6]
+y_inter_col = sys.argv[7]
+
+df[x_col] = pd.to_numeric(df[x_col], errors='coerce')
+df[y_col] = pd.to_numeric(df[y_col], errors='coerce')
+
+# Load shapefile (must contain LineString)
+line_gdf = gpd.read_file(sys.argv[2])
+line = line_gdf.union_all()  # merge if multiple lines
+
+# Create geometry
+df['geometry'] = df.apply(
+    lambda row: Point(row[x_col], row[y_col]) if not pd.isna(row[x_col]) and not pd.isna(row[y_col]) else None,
+    axis=1
+)
+
+# Get position
+df['distance_on_line'] = df['geometry'].apply(lambda g: line.project(g) if g else np.nan)
+
+# Interpolate
+df['distance_on_line'] = df['distance_on_line'].interpolate(method='linear', limit_direction='both')
+
+# Create interpolated points
+df['geometry'] = df['distance_on_line'].apply(lambda d: line.interpolate(d))
+
+# Start with original values
+df[x_inter_col] = df[x_col]
+df[y_inter_col] = df[y_col]
+
+# Update only missing ones
+df.loc[df[x_col].isna(), x_inter_col] = df.loc[df[x_col].isna(), 'geometry'].apply(lambda g: g.x)
+df.loc[df[y_col].isna(), y_inter_col] = df.loc[df[y_col].isna(), 'geometry'].apply(lambda g: g.y)
+
+
+# Load CSV and shape
+df = pd.read_csv("in/raw.csv", delimiter=';', encoding='utf-8')
+df[x_col] = pd.to_numeric(df[x_col], errors='coerce')
+df[y_col] = pd.to_numeric(df[y_col], errors='coerce')
+
+# Load the shapefile (must be a LineString)
+line = gpd.read_file("in/célé.shp").unary_union  # Merge multiple lines if needed
+
+# Assign geometry to known points
+df['geometry'] = df.apply(
+    lambda r: Point(r[x_col], r[y_col]) if pd.notna(r[x_col]) else None,
+    axis=1
+)
+
+# Compute the position of known points along the line
+df['distance_on_line'] = df['geometry'].apply(lambda g: line.project(g) if g else np.nan)
+
+# Interpolate missing distances (ensures every point has a valid position)
+df['distance_on_line'] = df['distance_on_line'].interpolate(method='linear', limit_direction='both')
+
+# Interpolate new coordinates from the reference shape
+df['geometry'] = df['distance_on_line'].apply(lambda d: line.interpolate(d) if pd.notna(d) else None)
+df[x_inter_col] = df['geometry'].apply(lambda g: g.x if g else np.nan)
+df[y_inter_col] = df['geometry'].apply(lambda g: g.y if g else np.nan)
+
+# Apply the conversion to the entire DataFrame row-wise
+wgs84 = pyproj.CRS("EPSG:4326")
+lambert93 = pyproj.CRS("EPSG:2154")
+transformer = pyproj.Transformer.from_crs(wgs84, lambert93, always_xy=True)
+
+df = df.apply(lambda row : convert_to_lambert93(row, transformer), axis=1)
+
+# Save fixed CSV
+df.to_csv(sys.argv[3], sep=';', index=False)