dev dans master (#171)

* feat: offline mode for color

* update las comparison (#168)

* update las comparison

* remove print

* Update ChangeLog (#169)

* bump version (#170)

---------

Co-authored-by: Nathan Lenglet <nathan.lenglet@ign.fr>
Co-authored-by: nlenglet-ign <58478624+nlenglet-ign@users.noreply.github.com>

Author: alavenant

.pre-commit-config.yaml

@@ -3,7 +3,7 @@ repos:
     rev: 23.12.0
     hooks:
       - id: black
-        language_version: python3.11
+        language_version: python3.12
   - repo: https://github.com/pycqa/flake8
     rev: 6.1.0
     hooks:

CHANGELOG.md

@@ -1,3 +1,7 @@
+# 1.15.7
+- [update] Update las_comparison with tolerance 
+- [update] Colorisation of Las with stream or files
+
 # 1.15.6
 - [fix] fix use of tempory file in windows
 

pdaltools/_version.py

@@ -1,4 +1,4 @@
-__version__ = "1.15.6"
+__version__ = "1.15.7"
 
 
 if __name__ == "__main__":

pdaltools/color.py

@@ -31,7 +31,7 @@ def match_min_max_with_pixel_size(min_d: float, max_d: float, pixel_per_meter: f
     return min_d, max_d
 
 
-def color(
+def color_from_stream(
     input_file: str,
     output_file: str,
     proj="",
@@ -156,28 +156,58 @@ def color(
     return tmp_ortho, tmp_ortho_irc
 
 
-def parse_args():
-    parser = argparse.ArgumentParser("Colorize tool", formatter_class=argparse.RawTextHelpFormatter)
-    parser.add_argument("--input", "-i", type=str, required=True, help="Input file")
-    parser.add_argument("--output", "-o", type=str, default="", help="Output file")
-    parser.add_argument(
-        "--proj", "-p", type=str, default="", help="Projection, default will use projection from metadata input"
+def color_from_files(
+    input_file: str,
+    output_file: str,
+    rgb_image: str,
+    irc_image: str,
+    color_rvb_enabled=True,
+    color_ir_enabled=True,
+    veget_index_file="",
+    vegetation_dim="Deviation",
+):
+    pipeline = pdal.Reader.las(filename=input_file)
+
+    writer_extra_dims = "all"
+
+    if veget_index_file and veget_index_file != "":
+        print(f"Remplissage du champ {vegetation_dim} à partir du fichier {veget_index_file}")
+        pipeline |= pdal.Filter.colorization(raster=veget_index_file, dimensions=f"{vegetation_dim}:1:256.0")
+        writer_extra_dims = [f"{vegetation_dim}=ushort"]
+
+    # Warning: the initial color is multiplied by 256 despite its initial 8-bits encoding
+    # which turns it to a 0 to 255*256 range.
+    # It is kept this way because of other dependencies that have been tuned to fit this range
+    if color_rvb_enabled:
+        pipeline |= pdal.Filter.colorization(raster=rgb_image, dimensions="Red:1:256.0, Green:2:256.0, Blue:3:256.0")
+    if color_ir_enabled:
+        pipeline |= pdal.Filter.colorization(raster=irc_image, dimensions="Infrared:1:256.0")
+
+    pipeline |= pdal.Writer.las(
+        filename=output_file, extra_dims=writer_extra_dims, minor_version="4", dataformat_id="8", forward="all"
     )
-    parser.add_argument("--resolution", "-r", type=float, default=5, help="Resolution, in pixel per meter")
-    parser.add_argument("--timeout", "-t", type=int, default=300, help="Timeout, in seconds")
-    parser.add_argument("--rvb", action="store_true", help="Colorize RVB")
-    parser.add_argument("--ir", action="store_true", help="Colorize IR")
-    parser.add_argument(
-        "--vegetation",
-        type=str,
-        default="",
-        help="Vegetation file (raster), value will be stored in 'vegetation_dim' field",
+
+    print("Traitement du nuage de point")
+    pipeline.execute()
+
+
+def argument_parser():
+    parser = argparse.ArgumentParser("Colorize tool")
+    subparsers = parser.add_subparsers(required=True)
+
+    # first command is 'from_stream'
+    from_stream = subparsers.add_parser("from_stream", help="Images are downloaded from streams")
+    from_stream.add_argument(
+        "--proj", "-p", type=str, default="", help="Projection, default will use projection from metadata input"
     )
-    parser.add_argument(
-        "--vegetation_dim", type=str, default="Deviation", help="name of the extra_dim uses for the vegetation value"
+    from_stream.add_argument("--timeout", "-t", type=int, default=300, help="Timeout, in seconds")
+    from_stream.add_argument("--rvb", action="store_true", help="Colorize RVB")
+    from_stream.add_argument("--ir", action="store_true", help="Colorize IR")
+    from_stream.add_argument("--resolution", "-r", type=float, default=5, help="Resolution, in pixel per meter")
+    from_stream.add_argument(
+        "--check-images", "-c", action="store_true", help="Check that downloaded image is not white"
     )
-    parser.add_argument("--check-images", "-c", action="store_true", help="Check that downloaded image is not white")
-    parser.add_argument(
+    from_stream.add_argument(
         "--stream-RGB",
         type=str,
         default="ORTHOIMAGERY.ORTHOPHOTOS",
@@ -186,27 +216,49 @@ def parse_args():
 for 20cm resolution rasters, use HR.ORTHOIMAGERY.ORTHOPHOTOS
 for 50 cm resolution rasters, use ORTHOIMAGERY.ORTHOPHOTOS.BDORTHO""",
     )
-    parser.add_argument(
+    from_stream.add_argument(
         "--stream-IRC",
         type=str,
         default="ORTHOIMAGERY.ORTHOPHOTOS.IRC",
         help="""WMS raster stream for IRC colorization. Default to ORTHOIMAGERY.ORTHOPHOTOS.IRC
 Documentation about possible stream : https://geoservices.ign.fr/services-web-experts-ortho""",
     )
-    parser.add_argument(
+    from_stream.add_argument(
         "--size-max-GPF",
         type=int,
         default=5000,
         help="Maximum edge size (in pixels) of downloaded images."
         " If input file needs more, several images are downloaded and merged.",
     )
+    add_common_options(from_stream)
+    from_stream.set_defaults(func=from_stream_func)
 
-    return parser.parse_args()
+    # second command is 'from_files'
+    from_files = subparsers.add_parser("from_files", help="Images are in directories from RGB/IRC")
+    from_files.add_argument("--image_RGB", type=str, required=True, help="RGB image filepath")
+    from_files.add_argument("--image_IRC", type=str, required=True, help="IRC image filepath")
+    add_common_options(from_files)
+    from_files.set_defaults(func=from_files_func)
 
+    return parser
 
-if __name__ == "__main__":
-    args = parse_args()
-    color(
+
+def add_common_options(parser):
+    parser.add_argument("--input", "-i", type=str, required=True, help="Input file")
+    parser.add_argument("--output", "-o", type=str, default="", help="Output file")
+    parser.add_argument(
+        "--vegetation",
+        type=str,
+        default="",
+        help="Vegetation file (raster), value will be stored in 'vegetation_dim' field",
+    )
+    parser.add_argument(
+        "--vegetation_dim", type=str, default="Deviation", help="name of the extra_dim uses for the vegetation value"
+    )
+
+
+def from_stream_func(args):
+    color_from_stream(
         input_file=args.input,
         output_file=args.output,
         proj=args.proj,
@@ -221,3 +273,33 @@ def parse_args():
         stream_IRC=args.stream_IRC,
         size_max_gpf=args.size_max_GPF,
     )
+
+
+def from_files_func(args):
+    if args.image_RGB and args.image_RGB != "":
+        color_rvb_enabled = True
+    else:
+        color_rvb_enabled = False
+    if args.image_IRC and args.image_IRC != "":
+        color_irc_enabled = True
+    else:
+        color_irc_enabled = False
+
+    if not color_rvb_enabled and not color_irc_enabled:
+        raise ValueError("At least one of --rvb or --ir must be provided")
+
+    color_from_files(
+        input_file=args.input,
+        output_file=args.output,
+        rgb_image=args.image_RGB,
+        irc_image=args.image_IRC,
+        color_rvb_enabled=color_rvb_enabled,
+        color_ir_enabled=color_irc_enabled,
+        veget_index_file=args.vegetation,
+        vegetation_dim=args.vegetation_dim,
+    )
+
+
+if __name__ == "__main__":
+    args = argument_parser.parse_args()
+    args.func(args)

pdaltools/las_comparison.py

@@ -1,12 +1,13 @@
 import argparse
-from pathlib import Path
-from typing import Tuple
+from typing import Tuple, Dict, Optional
 
 import laspy
 import numpy as np
 
+from pathlib import Path
+
 
-def compare_las_dimensions(file1: Path, file2: Path, dimensions: list = None) -> Tuple[bool, int, float]:
+def compare_las_dimensions(file1: Path, file2: Path, dimensions: list = None, precision: Optional[Dict[str, float]] = None) -> Tuple[bool, int, float]:
     """
     Compare specified dimensions between two LAS files.
     If no dimensions are specified, compares all available dimensions.
@@ -16,6 +17,8 @@ def compare_las_dimensions(file1: Path, file2: Path, dimensions: list = None) ->
         file1: Path to the first LAS file
         file2: Path to the second LAS file
         dimensions: List of dimension names to compare (optional)
+        precision: Dictionary mapping dimension names to tolerance values for float comparison.
+                   If None or dimension not in dict, uses exact comparison (default: None)
 
     Returns:
         bool: True if all specified dimensions are identical, False otherwise
@@ -59,20 +62,42 @@ def compare_las_dimensions(file1: Path, file2: Path, dimensions: list = None) ->
         # Compare each dimension
         for dim in dimensions:
             try:
+                
                 # Get sorted dimension arrays
                 dim1 = np.array(las1[dim])[sort_idx1]
                 dim2 = np.array(las2[dim])[sort_idx2]
 
+                # Get precision for this dimension (if specified)
+                dim_precision = None
+                if precision is not None and dim in precision:
+                    dim_precision = precision[dim]
+
                 # Compare dimensions
-                if not np.array_equal(dim1, dim2):
-                    # Find differences
-                    diff_indices = np.where(dim1 != dim2)[0]
-                    print(f"Found {len(diff_indices)} points with different {dim}:")
-                    for idx in diff_indices[:10]:  # Show first 10 differences
-                        print(f"Point {idx}: file1={dim1[idx]}, file2={dim2[idx]}")
-                    if len(diff_indices) > 10:
-                        print(f"... and {len(diff_indices) - 10} more differences")
-                    return False, len(diff_indices), 100 * len(diff_indices) / len(las1)
+                if dim_precision is not None:
+                    # Use tolerance-based comparison for floats
+                    are_equal = np.allclose(dim1, dim2, rtol=0, atol=dim_precision)
+                    if not are_equal:
+                        # Find differences
+                        diff_mask = ~np.isclose(dim1, dim2, rtol=0, atol=dim_precision)
+                        diff_indices = np.where(diff_mask)[0]
+                        print(f"Found {len(diff_indices)} points with different {dim} (tolerance={dim_precision}):")
+                        for idx in diff_indices[:10]:  # Show first 10 differences
+                            diff_value = abs(dim1[idx] - dim2[idx])
+                            print(f"Point {idx}: file1={dim1[idx]}, file2={dim2[idx]}, diff={diff_value}")
+                        if len(diff_indices) > 10:
+                            print(f"... and {len(diff_indices) - 10} more differences")
+                        return False, len(diff_indices), 100 * len(diff_indices) / len(las1)
+                else:
+                    # Exact comparison
+                    if not np.array_equal(dim1, dim2):
+                        # Find differences
+                        diff_indices = np.where(dim1 != dim2)[0]
+                        print(f"Found {len(diff_indices)} points with different {dim}:")
+                        for idx in diff_indices[:10]:  # Show first 10 differences
+                            print(f"Point {idx}: file1={dim1[idx]}, file2={dim2[idx]}")
+                        if len(diff_indices) > 10:
+                            print(f"... and {len(diff_indices) - 10} more differences")
+                        return False, len(diff_indices), 100 * len(diff_indices) / len(las1)
 
             except KeyError:
                 print(f"Dimension '{dim}' not found in one or both files")
@@ -93,12 +118,32 @@ def compare_las_dimensions(file1: Path, file2: Path, dimensions: list = None) ->
 
 # Update main function to use the new compare function
 def main():
-    parser = argparse.ArgumentParser(description="Compare dimensions between two LAS files")
+    parser = argparse.ArgumentParser(
+        description="Compare dimensions between two LAS files",
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        epilog="""
+Examples:
+  # Compare all dimensions with exact match
+  python las_comparison.py file1.las file2.las
+
+  # Compare specific dimensions with precision per dimension
+  python las_comparison.py file1.las file2.las --dimensions X Y Z --precision X=0.001 Y=0.001 Z=0.0001
+
+  # Compare all dimensions with precision for specific ones
+  python las_comparison.py file1.las file2.las --precision X=0.001 Y=0.001
+        """
+    )
     parser.add_argument("file1", type=str, help="Path to first LAS file")
     parser.add_argument("file2", type=str, help="Path to second LAS file")
     parser.add_argument(
         "--dimensions", nargs="*", help="List of dimensions to compare. If not specified, compares all dimensions."
     )
+    parser.add_argument(
+        "--precision", nargs="*", metavar="DIM=VAL",
+        help="Tolerance for float comparison per dimension (format: DIMENSION=PRECISION). "
+             "Example: --precision X=0.001 Y=0.001 Z=0.0001. "
+             "Dimensions not specified will use exact comparison."
+    )
 
     args = parser.parse_args()
 
@@ -109,7 +154,18 @@ def main():
         print("Error: One or both files do not exist")
         exit(1)
 
-    result = compare_las_dimensions(file1, file2, args.dimensions)
+    # Parse precision dictionary from command line arguments
+    precision_dict = None
+    if args.precision:
+        precision_dict = {}
+        for prec_spec in args.precision:
+            try:
+                dim_name, prec_value = prec_spec.split('=', 1)
+                precision_dict[dim_name] = float(prec_value)
+            except ValueError:
+                parser.error(f"Invalid precision format: '{prec_spec}'. Expected format: DIMENSION=PRECISION (e.g., X=0.001)")
+
+    result = compare_las_dimensions(file1, file2, args.dimensions, precision_dict)
     print(f"Dimensions comparison result: {'identical' if result[0] else 'different'}")
     return result