Upload folder using huggingface_hub

.gitignore

@@ -0,0 +1,3 @@
+.DS_Store
+__pycache__
+hoho.egg-info/

LICENSE.txt

@@ -0,0 +1,13 @@
+Copyright 2024 Jack Langerman & Dmytro Mishkin
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

README.md

@@ -0,0 +1,20 @@
+---
+license: apache-2.0
+---
+# HoHo Tools
+
+Tools and utilities for the [S23DR competition](https://huggingface.co/spaces/usm3d/S23DR) and [HoHo Dataset](https://huggingface.co/datasets/usm3d/usm-training-data)
+
+## Installation 
+```bash
+# pip install over ssh
+pip install git+ssh://git@hf.co/usm3d/tools.git 
+
+# pip install over http
+pip install git+http://hf.co/usm3d/tools.git    
+
+# editable
+git clone http://hf.co/usm3d/tools 
+cd tools
+pip install -e .
+```

hoho/__init__.py

@@ -0,0 +1,27 @@
+from .hoho import *
+from . import vis
+from . import read_write_colmap
+from .wed import compute_WED
+
+import importlib
+import sys
+
+class LazyLoadModule:
+    def __init__(self, module_name):
+        self.module_name = module_name
+        self.module = None
+
+    def __getattribute__(self, attr):
+        if attr == 'module_name' or attr == 'module':
+            return super().__getattribute__(attr)
+
+        if self.module is None:
+            self.module = importlib.import_module(f'hoho.{self.module_name}')
+            sys.modules[self.module_name] = self.module
+
+        return getattr(self.module, attr)
+    
+try:
+    import viz3d
+except ImportError:
+    viz3d = LazyLoadModule('viz3d')

hoho/color_mappings.py

@@ -0,0 +1,206 @@
+import numpy as np 
+
+gestalt_color_mapping = {
+    "unclassified": (215, 62, 138),
+    "apex": (235, 88, 48),
+    "eave_end_point": (248, 130, 228),
+    "flashing_end_point": (71, 11, 161),
+    "ridge": (214, 251, 248),
+    "rake": (13, 94, 47),
+    "eave": (54, 243, 63),
+    "post": (187, 123, 236),
+    "ground_line": (136, 206, 14),
+    "flashing": (162, 162, 32),
+    "step_flashing": (169, 255, 219),
+    "hip": (8, 89, 52),
+    "valley": (85, 27, 65),
+    "roof": (215, 232, 179),
+    "door": (110, 52, 23),
+    "garage": (50, 233, 171),
+    "window": (230, 249, 40),
+    "shutter": (122, 4, 233),
+    "fascia": (95, 230, 240),
+    "soffit": (2, 102, 197),
+    "horizontal_siding": (131, 88, 59),
+    "vertical_siding": (110, 187, 198),
+    "brick": (171, 252, 7),
+    "concrete": (32, 47, 246),
+    "other_wall": (112, 61, 240),
+    "trim": (151, 206, 58),
+    "unknown": (127, 127, 127),
+    "transition_line": (0,0,0),
+}
+
+ade20k_color_mapping = {
+    'wall': (120, 120, 120),
+    'building;edifice': (180, 120, 120),
+    'sky': (6, 230, 230),
+    'floor;flooring': (80, 50, 50),
+    'tree': (4, 200, 3),
+    'ceiling': (120, 120, 80),
+    'road;route': (140, 140, 140),
+    'bed': (204, 5, 255),
+    'windowpane;window': (230, 230, 230),
+    'grass': (4, 250, 7),
+    'cabinet': (224, 5, 255),
+    'sidewalk;pavement': (235, 255, 7),
+    'person;individual;someone;somebody;mortal;soul': (150, 5, 61),
+    'earth;ground': (120, 120, 70),
+    'door;double;door': (8, 255, 51),
+    'table': (255, 6, 82),
+    'mountain;mount': (143, 255, 140),
+    'plant;flora;plant;life': (204, 255, 4),
+    'curtain;drape;drapery;mantle;pall': (255, 51, 7),
+    'chair': (204, 70, 3),
+    'car;auto;automobile;machine;motorcar': (0, 102, 200),
+    'water': (61, 230, 250),
+    'painting;picture': (255, 6, 51),
+    'sofa;couch;lounge': (11, 102, 255),
+    'shelf': (255, 7, 71),
+    'house': (255, 9, 224),
+    'sea': (9, 7, 230),
+    'mirror': (220, 220, 220),
+    'rug;carpet;carpeting': (255, 9, 92),
+    'field': (112, 9, 255),
+    'armchair': (8, 255, 214),
+    'seat': (7, 255, 224),
+    'fence;fencing': (255, 184, 6),
+    'desk': (10, 255, 71),
+    'rock;stone': (255, 41, 10),
+    'wardrobe;closet;press': (7, 255, 255),
+    'lamp': (224, 255, 8),
+    'bathtub;bathing;tub;bath;tub': (102, 8, 255),
+    'railing;rail': (255, 61, 6),
+    'cushion': (255, 194, 7),
+    'base;pedestal;stand': (255, 122, 8),
+    'box': (0, 255, 20),
+    'column;pillar': (255, 8, 41),
+    'signboard;sign': (255, 5, 153),
+    'chest;of;drawers;chest;bureau;dresser': (6, 51, 255),
+    'counter': (235, 12, 255),
+    'sand': (160, 150, 20),
+    'sink': (0, 163, 255),
+    'skyscraper': (140, 140, 140),
+    'fireplace;hearth;open;fireplace': (250, 10, 15),
+    'refrigerator;icebox': (20, 255, 0),
+    'grandstand;covered;stand': (31, 255, 0),
+    'path': (255, 31, 0),
+    'stairs;steps': (255, 224, 0),
+    'runway': (153, 255, 0),
+    'case;display;case;showcase;vitrine': (0, 0, 255),
+    'pool;table;billiard;table;snooker;table': (255, 71, 0),
+    'pillow': (0, 235, 255),
+    'screen;door;screen': (0, 173, 255),
+    'stairway;staircase': (31, 0, 255),
+    'river': (11, 200, 200),
+    'bridge;span': (255 ,82, 0),
+    'bookcase': (0, 255, 245),
+    'blind;screen': (0, 61, 255),
+    'coffee;table;cocktail;table': (0, 255, 112),
+    'toilet;can;commode;crapper;pot;potty;stool;throne': (0, 255, 133),
+    'flower': (255, 0, 0),
+    'book': (255, 163, 0),
+    'hill': (255, 102, 0),
+    'bench': (194, 255, 0),
+    'countertop': (0, 143, 255),
+    'stove;kitchen;stove;range;kitchen;range;cooking;stove': (51, 255, 0),
+    'palm;palm;tree': (0, 82, 255),
+    'kitchen;island': (0, 255, 41),
+    'computer;computing;machine;computing;device;data;processor;electronic;computer;information;processing;system': (0, 255, 173),
+    'swivel;chair': (10, 0, 255),
+    'boat': (173, 255, 0),
+    'bar': (0, 255, 153),
+    'arcade;machine': (255, 92, 0),
+    'hovel;hut;hutch;shack;shanty': (255, 0, 255),
+    'bus;autobus;coach;charabanc;double-decker;jitney;motorbus;motorcoach;omnibus;passenger;vehicle': (255, 0, 245),
+    'towel': (255, 0, 102),
+    'light;light;source': (255, 173, 0),
+    'truck;motortruck': (255, 0, 20),
+    'tower': (255, 184, 184),
+    'chandelier;pendant;pendent': (0, 31, 255),
+    'awning;sunshade;sunblind': (0, 255, 61),
+    'streetlight;street;lamp': (0, 71, 255),
+    'booth;cubicle;stall;kiosk': (255, 0, 204),
+    'television;television;receiver;television;set;tv;tv;set;idiot;box;boob;tube;telly;goggle;box': (0, 255, 194),
+    'airplane;aeroplane;plane': (0, 255, 82),
+    'dirt;track': (0, 10, 255),
+    'apparel;wearing;apparel;dress;clothes': (0, 112, 255),
+    'pole': (51, 0, 255),
+    'land;ground;soil': (0, 194, 255),
+    'bannister;banister;balustrade;balusters;handrail': (0, 122, 255),
+    'escalator;moving;staircase;moving;stairway': (0, 255, 163),
+    'ottoman;pouf;pouffe;puff;hassock': (255, 153, 0),
+    'bottle': (0, 255, 10),
+    'buffet;counter;sideboard': (255, 112, 0),
+    'poster;posting;placard;notice;bill;card': (143, 255, 0),
+    'stage': (82, 0, 255),
+    'van': (163, 255, 0),
+    'ship': (255, 235, 0),
+    'fountain': (8, 184, 170),
+    'conveyer;belt;conveyor;belt;conveyer;conveyor;transporter': (133, 0, 255),
+    'canopy': (0, 255, 92),
+    'washer;automatic;washer;washing;machine': (184, 0, 255),
+    'plaything;toy': (255, 0, 31),
+    'swimming;pool;swimming;bath;natatorium': (0, 184, 255),
+    'stool': (0, 214, 255),
+    'barrel;cask': (255, 0, 112),
+    'basket;handbasket': (92, 255, 0),
+    'waterfall;falls': (0, 224, 255),
+    'tent;collapsible;shelter': (112, 224, 255),
+    'bag': (70, 184, 160),
+    'minibike;motorbike': (163, 0, 255),
+    'cradle': (153, 0, 255),
+    'oven': (71, 255, 0),
+    'ball': (255, 0, 163),
+    'food;solid;food': (255, 204, 0),
+    'step;stair': (255, 0, 143),
+    'tank;storage;tank': (0, 255, 235),
+    'trade;name;brand;name;brand;marque': (133, 255, 0),
+    'microwave;microwave;oven': (255, 0, 235),
+    'pot;flowerpot': (245, 0, 255),
+    'animal;animate;being;beast;brute;creature;fauna': (255, 0, 122),
+    'bicycle;bike;wheel;cycle': (255, 245, 0),
+    'lake': (10, 190, 212),
+    'dishwasher;dish;washer;dishwashing;machine': (214, 255, 0),
+    'screen;silver;screen;projection;screen': (0, 204, 255),
+    'blanket;cover': (20, 0, 255),
+    'sculpture': (255, 255, 0),
+    'hood;exhaust;hood': (0, 153, 255),
+    'sconce': (0, 41, 255),
+    'vase': (0, 255, 204),
+    'traffic;light;traffic;signal;stoplight': (41, 0, 255),
+    'tray': (41, 255, 0),
+    'ashcan;trash;can;garbage;can;wastebin;ash;bin;ash-bin;ashbin;dustbin;trash;barrel;trash;bin': (173, 0, 255),
+    'fan': (0, 245, 255),
+    'pier;wharf;wharfage;dock': (71, 0, 255),
+    'crt;screen': (122, 0, 255),
+    'plate': (0, 255, 184),
+    'monitor;monitoring;device': (0, 92, 255),
+    'bulletin;board;notice;board': (184, 255, 0),
+    'shower': (0, 133, 255),
+    'radiator': (255, 214, 0),
+    'glass;drinking;glass': (25, 194, 194),
+    'clock': (102, 255, 0),
+    'flag': (92, 0, 255),
+}
+
+
+# edge_colors = np.asarray([(214, 251, 248),
+#                           (13, 94, 47),
+#                           (54, 243, 63),
+#                           (187, 123, 236),
+#                           (162, 162, 32),
+#                           (169, 255, 219),
+#                           (8, 89, 52),
+#                           (85, 27, 65),
+#                           (0, 0, 0)]
+
+
+# edge_colors = np.array([[ 54, 243,  63], 
+#                         [214, 251, 248], 
+#                         [169, 255, 219], 
+#                         [ 13,  94,  47], 
+#                         [162, 162,  32], 
+#                         [187, 123, 236], 
+#                         [ 85,  27,  65], 
+#                         [  0,  0,    0]])

hoho/hoho.py

@@ -0,0 +1,338 @@
+import os 
+import json
+import shutil
+from pathlib import Path
+from typing import Dict
+import warnings 
+import contextlib
+import tempfile 
+from PIL import Image
+import io
+import webdataset as wds 
+import numpy as np
+import importlib
+import subprocess
+
+from PIL import ImageFile
+
+from huggingface_hub.utils._headers import build_hf_headers # note: using _headers
+
+
+ImageFile.LOAD_TRUNCATED_IMAGES = True
+
+LOCAL_DATADIR = None
+
+def setup(local_dir='./data/usm-training-data/data'):
+    
+    # If we are in the test environment, we need to link the data directory to the correct location
+    tmp_datadir = Path('/tmp/data/data')
+    local_test_datadir = Path('./data/usm-test-data-x/data')
+    local_val_datadir = Path(local_dir)
+    
+    os.system('pwd')
+    os.system('ls -lahtr .')
+    
+    if tmp_datadir.exists() and not local_test_datadir.exists():
+        global LOCAL_DATADIR
+        LOCAL_DATADIR = local_test_datadir
+        # shutil.move(datadir, './usm-test-data-x/data')
+        print(f"Linking {tmp_datadir} to {LOCAL_DATADIR} (we are in the test environment)")
+        LOCAL_DATADIR.parent.mkdir(parents=True, exist_ok=True)
+        LOCAL_DATADIR.symlink_to(tmp_datadir)
+    else:
+        LOCAL_DATADIR = local_val_datadir
+        print(f"Using {LOCAL_DATADIR} as the data directory (we are running locally)")
+    
+    if not LOCAL_DATADIR.exists():
+        warnings.warn(f"Data directory {LOCAL_DATADIR} does not exist: creating it...")
+        LOCAL_DATADIR.mkdir(parents=True)
+    
+    return LOCAL_DATADIR
+
+
+def download_package(package_name, path_to_save='packages'):
+    """
+    Downloads a package using pip and saves it to a specified directory.
+
+    Parameters:
+    package_name (str): The name of the package to download.
+    path_to_save (str): The path to the directory where the package will be saved.
+    """
+    try:
+        # pip download webdataset -d packages/webdataset --platform manylinux1_x86_64 --python-version 38 --only-binary=:all:
+        subprocess.check_call([subprocess.sys.executable, "-m", "pip", "download", package_name, 
+                               "-d", str(Path(path_to_save)/package_name),  # Download the package to the specified directory
+                               "--platform", "manylinux1_x86_64",  # Specify the platform
+                               "--python-version", "38",  # Specify the Python version
+                               "--only-binary=:all:"])  # Download only binary packages
+        print(f'Package "{package_name}" downloaded successfully')
+    except subprocess.CalledProcessError as e:
+        print(f'Failed to downloaded package "{package_name}". Error: {e}')
+              
+              
+def install_package_from_local_file(package_name, folder='packages'):
+    """
+    Installs a package from a local .whl file or a directory containing .whl files using pip.
+
+    Parameters:
+    path_to_file_or_directory (str): The path to the .whl file or the directory containing .whl files.
+    """
+    try:
+        pth = str(Path(folder) / package_name)
+        subprocess.check_call([subprocess.sys.executable, "-m", "pip", "install", 
+                               "--no-index",  # Do not use package index
+                               "--find-links", pth,  # Look for packages in the specified directory or at the file
+                               package_name])  # Specify the package to install
+        print(f"Package installed successfully from {pth}")
+    except subprocess.CalledProcessError as e:
+        print(f"Failed to install package from {pth}. Error: {e}")
+        
+        
+def importt(module_name, as_name=None):
+    """
+    Imports a module and returns it.
+
+    Parameters:
+    module_name (str): The name of the module to import.
+    as_name (str): The name to use for the imported module. If None, the original module name will be used.
+
+    Returns:
+    The imported module.
+    """
+    for _ in range(2):
+        try:
+            if as_name is None:
+                print(f'imported {module_name}')
+                return importlib.import_module(module_name)
+            else:
+                print(f'imported {module_name} as {as_name}')
+                return importlib.import_module(module_name, as_name)
+        except ModuleNotFoundError as e:
+            install_package_from_local_file(module_name)
+            print(f"Failed to import module {module_name}. Error: {e}")
+            
+    
+def prepare_submission():
+    # Download packages from requirements.txt 
+    if Path('requirements.txt').exists():
+        print('downloading packages from requirements.txt')
+        Path('packages').mkdir(exist_ok=True)
+        with open('requirements.txt') as f:
+            packages = f.readlines()
+            for p in packages:
+                download_package(p.strip())
+                
+    print('all packages downloaded. Don\'t foget to include the packages in the submission by adding them with git lfs.')
+        
+
+def Rt_to_eye_target(im, K, R, t):
+    height = im.height
+    focal_length = K[0,0]
+    fov = 2.0 * np.arctan2((0.5 * height), focal_length) / (np.pi / 180.0)
+
+    x_axis, y_axis, z_axis = R
+
+    eye = -(R.T @ t).squeeze()
+    z_axis = z_axis.squeeze()
+    target = eye + z_axis
+    up = -y_axis
+    
+    return eye, target, up, fov        
+
+
+########## general utilities ##########
+
+
+@contextlib.contextmanager
+def working_directory(path):
+    """Changes working directory and returns to previous on exit."""
+    prev_cwd = Path.cwd()
+    os.chdir(path)
+    try:
+        yield
+    finally:
+        os.chdir(prev_cwd)
+        
+@contextlib.contextmanager
+def temp_working_directory():
+    with tempfile.TemporaryDirectory(dir='.') as D:
+        with working_directory(D):
+            yield
+
+
+############# Dataset ############# 
+def proc(row, split='train'):
+    out = {}
+    out['__key__'] = None
+    out['__imagekey__'] = []
+    for k, v in row.items():
+        key_parts = k.split('.')
+        colname = key_parts[0]
+        if colname == 'ade20k': 
+            out['__imagekey__'].append(key_parts[1])
+        if colname in {'ade20k', 'depthcm', 'gestalt'}:
+            if colname in out:
+                out[colname].append(v)
+            else:
+                out[colname] = [v]
+        elif colname in {'wireframe', 'mesh'}:
+            out.update({a: b for a,b in v.items()})
+        elif colname in 'kr':
+            out[colname.upper()] = v
+        else:
+            out[colname] = v
+    return Sample(out)
+
+
+from . import read_write_colmap
+def decode_colmap(s):
+    with temp_working_directory():
+
+        with open('points3D.bin', 'wb') as stream:
+            stream.write(s['points3d'])
+
+
+        with open('cameras.bin', 'wb') as stream:
+            stream.write(s['cameras'])
+
+
+        with open('images.bin', 'wb') as stream:
+            stream.write(s['images'])
+    
+        
+        cameras, images, points3D = read_write_colmap.read_model(
+            path='.', ext='.bin'
+        )
+    return cameras, images, points3D
+
+
+def decode(row):
+    cameras, images, points3D = decode_colmap(row)
+    
+    out = {}
+    
+    for k, v in row.items():
+        # colname = k.split('.')[0]
+        if k in {'ade20k', 'depthcm', 'gestalt'}:
+            # print(k, len(v), type(v))
+            v = [Image.open(io.BytesIO(im)) for im in v]
+            if k in out:
+                out[k].extend(v)
+            else:
+                out[k] = v
+        elif k in {'wireframe', 'mesh'}:
+            # out.update({a: b.tolist() for a,b in v.items()})
+            v = dict(np.load(io.BytesIO(v)))
+            out.update({a: b for a,b in v.items()})
+        elif k in 'kr':
+            out[k.upper()] = v
+        elif k == 'cameras':
+            out[k] = cameras
+        elif k == 'images':
+            out[k] = images
+        elif k =='points3d':
+            out[k] = points3D
+        else:
+            out[k] = v
+            
+    return Sample(out)
+
+
+class Sample(Dict):
+    def __repr__(self):
+        return str({k: v.shape if hasattr(v, 'shape') else [type(v[0])] if isinstance(v, list) else type(v) for k,v in self.items()})
+
+    
+        
+def get_params():
+    exmaple_param_dict = {
+        "competition_id": "usm3d/S23DR",
+        "competition_type": "script",
+        "metric": "custom",
+        "token": "hf_**********************************",
+        "team_id": "local-test-team_id",
+        "submission_id": "local-test-submission_id",
+        "submission_id_col": "__key__",
+        "submission_cols": [
+            "__key__",
+            "wf_edges",
+            "wf_vertices",
+            "edge_semantics"
+        ],
+        "submission_rows": 180,
+        "output_path": ".",
+        "submission_repo": "<THE HF MODEL ID of THIS REPO",
+        "time_limit": 7200,
+        "dataset": "usm3d/usm-test-data-x",
+        "submission_filenames": [
+            "submission.parquet"
+        ]
+    }
+    
+    param_path = Path('params.json')
+    
+    if not param_path.exists():
+        print('params.json not found (this means we probably aren\'t in the test env). Using example params.')
+        params = exmaple_param_dict
+    else:
+        print('found params.json (this means we are probably in the test env). Using params from file.')
+        with param_path.open() as f:
+            params = json.load(f)
+    print(params)
+    return params
+
+
+
+
+
+SHARD_IDS = {'train': (0, 25), 'val': (25, 26), 'public': (26, 27), 'private': (27, 32)}
+def get_dataset(decode='pil', proc=proc, split='train', dataset_type='webdataset', stream=True):
+    if LOCAL_DATADIR is None:
+        raise ValueError('LOCAL_DATADIR is not set. Please run setup() first.')
+        
+    local_dir = Path(LOCAL_DATADIR)
+    if split != 'all':
+        local_dir = local_dir / split
+    
+    paths = [str(p) for p in local_dir.rglob('*.tar.gz')]
+    msg = f'no tarfiles found in {local_dir}.'
+    if len(paths) == 0:
+        if stream:
+            if split=='all': split = 'train'
+            warnings.warn('streaming isn\'t using with \'all\': changing `split` to \'train\'')
+            warnings.warn(msg)
+            if split == 'val':
+                names = [f'data/val/inputs/hoho_v3_{i:03}-of-032.tar.gz' for i in range(*SHARD_IDS[split])]
+            elif split == 'train':
+                names = [f'data/train/hoho_v3_{i:03}-of-032.tar.gz' for i in range(*SHARD_IDS[split])]
+            
+            auth = build_hf_headers()['authorization']
+            paths = [f"pipe:curl -L -s https://huggingface.co/datasets/usm3d/hoho-train-set/resolve/main/{name} -H 'Authorization: {auth}'" for name in names]
+        else:
+            raise FileNotFoundError(msg)
+    
+    dataset = wds.WebDataset(paths)
+        
+    if decode is not None:
+        dataset = dataset.decode(decode)
+    else:
+        dataset = dataset.decode()
+    
+    dataset = dataset.map(proc)
+    
+    if dataset_type == 'webdataset':
+        return dataset
+    
+    if dataset_type == 'hf':
+        import datasets
+        from datasets import Features, Value, Sequence, Image, Array2D
+       
+        if split == 'train':
+            return datasets.IterableDataset.from_generator(lambda: dataset.iterator())
+        elif split == 'val':
+            return datasets.IterableDataset.from_generator(lambda: dataset.iterator())
+        else:
+            raise NotImplementedError('only train and val are implemented as hf datasets')
+
+    
+    

hoho/read_write_colmap.py

@@ -0,0 +1,489 @@
+# Modified to read from bytes-like object by Dmytro Mishkin.
+# The original license is below:
+# Copyright (c) 2018, ETH Zurich and UNC Chapel Hill.
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+#
+#     * Redistributions of source code must retain the above copyright
+#       notice, this list of conditions and the following disclaimer.
+#
+#     * Redistributions in binary form must reproduce the above copyright
+#       notice, this list of conditions and the following disclaimer in the
+#       documentation and/or other materials provided with the distribution.
+#
+#     * Neither the name of ETH Zurich and UNC Chapel Hill nor the names of
+#       its contributors may be used to endorse or promote products derived
+#       from this software without specific prior written permission.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+# POSSIBILITY OF SUCH DAMAGE.
+#
+# Author: Johannes L. Schoenberger (jsch-at-demuc-dot-de)
+
+import os
+import collections
+import numpy as np
+import struct
+import argparse
+
+
+CameraModel = collections.namedtuple(
+    "CameraModel", ["model_id", "model_name", "num_params"])
+Camera = collections.namedtuple(
+    "Camera", ["id", "model", "width", "height", "params"])
+BaseImage = collections.namedtuple(
+    "Image", ["id", "qvec", "tvec", "camera_id", "name", "xys", "point3D_ids"])
+Point3D = collections.namedtuple(
+    "Point3D", ["id", "xyz", "rgb", "error", "image_ids", "point2D_idxs"])
+
+
+class Image(BaseImage):
+    def qvec2rotmat(self):
+        return qvec2rotmat(self.qvec)
+
+
+CAMERA_MODELS = {
+    CameraModel(model_id=0, model_name="SIMPLE_PINHOLE", num_params=3),
+    CameraModel(model_id=1, model_name="PINHOLE", num_params=4),
+    CameraModel(model_id=2, model_name="SIMPLE_RADIAL", num_params=4),
+    CameraModel(model_id=3, model_name="RADIAL", num_params=5),
+    CameraModel(model_id=4, model_name="OPENCV", num_params=8),
+    CameraModel(model_id=5, model_name="OPENCV_FISHEYE", num_params=8),
+    CameraModel(model_id=6, model_name="FULL_OPENCV", num_params=12),
+    CameraModel(model_id=7, model_name="FOV", num_params=5),
+    CameraModel(model_id=8, model_name="SIMPLE_RADIAL_FISHEYE", num_params=4),
+    CameraModel(model_id=9, model_name="RADIAL_FISHEYE", num_params=5),
+    CameraModel(model_id=10, model_name="THIN_PRISM_FISHEYE", num_params=12)
+}
+CAMERA_MODEL_IDS = dict([(camera_model.model_id, camera_model)
+                         for camera_model in CAMERA_MODELS])
+CAMERA_MODEL_NAMES = dict([(camera_model.model_name, camera_model)
+                           for camera_model in CAMERA_MODELS])
+
+
+def read_next_bytes(fid, num_bytes, format_char_sequence, endian_character="<"):
+    """Read and unpack the next bytes from a binary file.
+    :param fid:
+    :param num_bytes: Sum of combination of {2, 4, 8}, e.g. 2, 6, 16, 30, etc.
+    :param format_char_sequence: List of {c, e, f, d, h, H, i, I, l, L, q, Q}.
+    :param endian_character: Any of {@, =, <, >, !}
+    :return: Tuple of read and unpacked values.
+    """
+    data = fid.read(num_bytes)
+    return struct.unpack(endian_character + format_char_sequence, data)
+
+
+def write_next_bytes(fid, data, format_char_sequence, endian_character="<"):
+    """pack and write to a binary file.
+    :param fid:
+    :param data: data to send, if multiple elements are sent at the same time,
+    they should be encapsuled either in a list or a tuple
+    :param format_char_sequence: List of {c, e, f, d, h, H, i, I, l, L, q, Q}.
+    should be the same length as the data list or tuple
+    :param endian_character: Any of {@, =, <, >, !}
+    """
+    if isinstance(data, (list, tuple)):
+        bytes = struct.pack(endian_character + format_char_sequence, *data)
+    else:
+        bytes = struct.pack(endian_character + format_char_sequence, data)
+    fid.write(bytes)
+
+
+def read_cameras_text(path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::WriteCamerasText(const std::string& path)
+        void Reconstruction::ReadCamerasText(const std::string& path)
+    """
+    cameras = {}
+    with open(path, "r") as fid:
+        while True:
+            line = fid.readline()
+            if not line:
+                break
+            line = line.strip()
+            if len(line) > 0 and line[0] != "#":
+                elems = line.split()
+                camera_id = int(elems[0])
+                model = elems[1]
+                width = int(elems[2])
+                height = int(elems[3])
+                params = np.array(tuple(map(float, elems[4:])))
+                cameras[camera_id] = Camera(id=camera_id, model=model,
+                                            width=width, height=height,
+                                            params=params)
+    return cameras
+
+
+def read_cameras_binary(path_to_model_file=None, fid=None):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::WriteCamerasBinary(const std::string& path)
+        void Reconstruction::ReadCamerasBinary(const std::string& path)
+    """
+    cameras = {}
+    if fid is None:
+        fid = open(path_to_model_file, "rb")
+    num_cameras = read_next_bytes(fid, 8, "Q")[0]
+    for _ in range(num_cameras):
+        camera_properties = read_next_bytes(
+            fid, num_bytes=24, format_char_sequence="iiQQ")
+        camera_id = camera_properties[0]
+        model_id = camera_properties[1]
+        model_name = CAMERA_MODEL_IDS[camera_properties[1]].model_name
+        width = camera_properties[2]
+        height = camera_properties[3]
+        num_params = CAMERA_MODEL_IDS[model_id].num_params
+        params = read_next_bytes(fid, num_bytes=8*num_params,
+                                    format_char_sequence="d"*num_params)
+        cameras[camera_id] = Camera(id=camera_id,
+                                    model=model_name,
+                                    width=width,
+                                    height=height,
+                                    params=np.array(params))
+    assert len(cameras) == num_cameras
+    if path_to_model_file is not None:
+        fid.close()
+    return cameras
+
+
+def write_cameras_text(cameras, path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::WriteCamerasText(const std::string& path)
+        void Reconstruction::ReadCamerasText(const std::string& path)
+    """
+    HEADER = "# Camera list with one line of data per camera:\n" + \
+             "#   CAMERA_ID, MODEL, WIDTH, HEIGHT, PARAMS[]\n" + \
+             "# Number of cameras: {}\n".format(len(cameras))
+    with open(path, "w") as fid:
+        fid.write(HEADER)
+        for _, cam in cameras.items():
+            to_write = [cam.id, cam.model, cam.width, cam.height, *cam.params]
+            line = " ".join([str(elem) for elem in to_write])
+            fid.write(line + "\n")
+
+
+def write_cameras_binary(cameras, path_to_model_file):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::WriteCamerasBinary(const std::string& path)
+        void Reconstruction::ReadCamerasBinary(const std::string& path)
+    """
+    with open(path_to_model_file, "wb") as fid:
+        write_next_bytes(fid, len(cameras), "Q")
+        for _, cam in cameras.items():
+            model_id = CAMERA_MODEL_NAMES[cam.model].model_id
+            camera_properties = [cam.id,
+                                 model_id,
+                                 cam.width,
+                                 cam.height]
+            write_next_bytes(fid, camera_properties, "iiQQ")
+            for p in cam.params:
+                write_next_bytes(fid, float(p), "d")
+    return cameras
+
+
+def read_images_text(path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadImagesText(const std::string& path)
+        void Reconstruction::WriteImagesText(const std::string& path)
+    """
+    images = {}
+    with open(path, "r") as fid:
+        while True:
+            line = fid.readline()
+            if not line:
+                break
+            line = line.strip()
+            if len(line) > 0 and line[0] != "#":
+                elems = line.split()
+                image_id = int(elems[0])
+                qvec = np.array(tuple(map(float, elems[1:5])))
+                tvec = np.array(tuple(map(float, elems[5:8])))
+                camera_id = int(elems[8])
+                image_name = elems[9]
+                elems = fid.readline().split()
+                xys = np.column_stack([tuple(map(float, elems[0::3])),
+                                       tuple(map(float, elems[1::3]))])
+                point3D_ids = np.array(tuple(map(int, elems[2::3])))
+                images[image_id] = Image(
+                    id=image_id, qvec=qvec, tvec=tvec,
+                    camera_id=camera_id, name=image_name,
+                    xys=xys, point3D_ids=point3D_ids)
+    return images
+
+
+def read_images_binary(path_to_model_file=None, fid=None):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadImagesBinary(const std::string& path)
+        void Reconstruction::WriteImagesBinary(const std::string& path)
+    """
+    images = {}
+    if fid is None:
+        fid = open(path_to_model_file, "rb")
+    num_reg_images = read_next_bytes(fid, 8, "Q")[0]
+    for _ in range(num_reg_images):
+        binary_image_properties = read_next_bytes(
+            fid, num_bytes=64, format_char_sequence="idddddddi")
+        image_id = binary_image_properties[0]
+        qvec = np.array(binary_image_properties[1:5])
+        tvec = np.array(binary_image_properties[5:8])
+        camera_id = binary_image_properties[8]
+        image_name = ""
+        current_char = read_next_bytes(fid, 1, "c")[0]
+        while current_char != b"\x00":   # look for the ASCII 0 entry
+            image_name += current_char.decode("utf-8")
+            current_char = read_next_bytes(fid, 1, "c")[0]
+        num_points2D = read_next_bytes(fid, num_bytes=8,
+                                        format_char_sequence="Q")[0]
+        x_y_id_s = read_next_bytes(fid, num_bytes=24*num_points2D,
+                                    format_char_sequence="ddq"*num_points2D)
+        xys = np.column_stack([tuple(map(float, x_y_id_s[0::3])),
+                                tuple(map(float, x_y_id_s[1::3]))])
+        point3D_ids = np.array(tuple(map(int, x_y_id_s[2::3])))
+        images[image_id] = Image(
+            id=image_id, qvec=qvec, tvec=tvec,
+            camera_id=camera_id, name=image_name,
+            xys=xys, point3D_ids=point3D_ids)
+    if path_to_model_file is not None:
+        fid.close()
+    return images
+
+
+def write_images_text(images, path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadImagesText(const std::string& path)
+        void Reconstruction::WriteImagesText(const std::string& path)
+    """
+    if len(images) == 0:
+        mean_observations = 0
+    else:
+        mean_observations = sum((len(img.point3D_ids) for _, img in images.items()))/len(images)
+    HEADER = "# Image list with two lines of data per image:\n" + \
+             "#   IMAGE_ID, QW, QX, QY, QZ, TX, TY, TZ, CAMERA_ID, NAME\n" + \
+             "#   POINTS2D[] as (X, Y, POINT3D_ID)\n" + \
+             "# Number of images: {}, mean observations per image: {}\n".format(len(images), mean_observations)
+
+    with open(path, "w") as fid:
+        fid.write(HEADER)
+        for _, img in images.items():
+            image_header = [img.id, *img.qvec, *img.tvec, img.camera_id, img.name]
+            first_line = " ".join(map(str, image_header))
+            fid.write(first_line + "\n")
+
+            points_strings = []
+            for xy, point3D_id in zip(img.xys, img.point3D_ids):
+                points_strings.append(" ".join(map(str, [*xy, point3D_id])))
+            fid.write(" ".join(points_strings) + "\n")
+
+
+def write_images_binary(images, path_to_model_file):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadImagesBinary(const std::string& path)
+        void Reconstruction::WriteImagesBinary(const std::string& path)
+    """
+    with open(path_to_model_file, "wb") as fid:
+        write_next_bytes(fid, len(images), "Q")
+        for _, img in images.items():
+            write_next_bytes(fid, img.id, "i")
+            write_next_bytes(fid, img.qvec.tolist(), "dddd")
+            write_next_bytes(fid, img.tvec.tolist(), "ddd")
+            write_next_bytes(fid, img.camera_id, "i")
+            for char in img.name:
+                write_next_bytes(fid, char.encode("utf-8"), "c")
+            write_next_bytes(fid, b"\x00", "c")
+            write_next_bytes(fid, len(img.point3D_ids), "Q")
+            for xy, p3d_id in zip(img.xys, img.point3D_ids):
+                write_next_bytes(fid, [*xy, p3d_id], "ddq")
+
+
+def read_points3D_text(path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadPoints3DText(const std::string& path)
+        void Reconstruction::WritePoints3DText(const std::string& path)
+    """
+    points3D = {}
+    with open(path, "r") as fid:
+        while True:
+            line = fid.readline()
+            if not line:
+                break
+            line = line.strip()
+            if len(line) > 0 and line[0] != "#":
+                elems = line.split()
+                point3D_id = int(elems[0])
+                xyz = np.array(tuple(map(float, elems[1:4])))
+                rgb = np.array(tuple(map(int, elems[4:7])))
+                error = float(elems[7])
+                image_ids = np.array(tuple(map(int, elems[8::2])))
+                point2D_idxs = np.array(tuple(map(int, elems[9::2])))
+                points3D[point3D_id] = Point3D(id=point3D_id, xyz=xyz, rgb=rgb,
+                                               error=error, image_ids=image_ids,
+                                               point2D_idxs=point2D_idxs)
+    return points3D
+
+
+def read_points3D_binary(path_to_model_file=None, fid=None):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadPoints3DBinary(const std::string& path)
+        void Reconstruction::WritePoints3DBinary(const std::string& path)
+    """
+    points3D = {}
+    if fid is None:
+        fid = open(path_to_model_file, "rb")
+    num_points = read_next_bytes(fid, 8, "Q")[0]
+    for _ in range(num_points):
+        binary_point_line_properties = read_next_bytes(
+            fid, num_bytes=43, format_char_sequence="QdddBBBd")
+        point3D_id = binary_point_line_properties[0]
+        xyz = np.array(binary_point_line_properties[1:4])
+        rgb = np.array(binary_point_line_properties[4:7])
+        error = np.array(binary_point_line_properties[7])
+        track_length = read_next_bytes(
+            fid, num_bytes=8, format_char_sequence="Q")[0]
+        track_elems = read_next_bytes(
+            fid, num_bytes=8*track_length,
+            format_char_sequence="ii"*track_length)
+        image_ids = np.array(tuple(map(int, track_elems[0::2])))
+        point2D_idxs = np.array(tuple(map(int, track_elems[1::2])))
+        points3D[point3D_id] = Point3D(
+            id=point3D_id, xyz=xyz, rgb=rgb,
+            error=error, image_ids=image_ids,
+            point2D_idxs=point2D_idxs)
+    if path_to_model_file is not None:
+        fid.close()
+    return points3D
+
+
+def write_points3D_text(points3D, path):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadPoints3DText(const std::string& path)
+        void Reconstruction::WritePoints3DText(const std::string& path)
+    """
+    if len(points3D) == 0:
+        mean_track_length = 0
+    else:
+        mean_track_length = sum((len(pt.image_ids) for _, pt in points3D.items()))/len(points3D)
+    HEADER = "# 3D point list with one line of data per point:\n" + \
+             "#   POINT3D_ID, X, Y, Z, R, G, B, ERROR, TRACK[] as (IMAGE_ID, POINT2D_IDX)\n" + \
+             "# Number of points: {}, mean track length: {}\n".format(len(points3D), mean_track_length)
+
+    with open(path, "w") as fid:
+        fid.write(HEADER)
+        for _, pt in points3D.items():
+            point_header = [pt.id, *pt.xyz, *pt.rgb, pt.error]
+            fid.write(" ".join(map(str, point_header)) + " ")
+            track_strings = []
+            for image_id, point2D in zip(pt.image_ids, pt.point2D_idxs):
+                track_strings.append(" ".join(map(str, [image_id, point2D])))
+            fid.write(" ".join(track_strings) + "\n")
+
+
+def write_points3D_binary(points3D, path_to_model_file):
+    """
+    see: src/base/reconstruction.cc
+        void Reconstruction::ReadPoints3DBinary(const std::string& path)
+        void Reconstruction::WritePoints3DBinary(const std::string& path)
+    """
+    with open(path_to_model_file, "wb") as fid:
+        write_next_bytes(fid, len(points3D), "Q")
+        for _, pt in points3D.items():
+            write_next_bytes(fid, pt.id, "Q")
+            write_next_bytes(fid, pt.xyz.tolist(), "ddd")
+            write_next_bytes(fid, pt.rgb.tolist(), "BBB")
+            write_next_bytes(fid, pt.error, "d")
+            track_length = pt.image_ids.shape[0]
+            write_next_bytes(fid, track_length, "Q")
+            for image_id, point2D_id in zip(pt.image_ids, pt.point2D_idxs):
+                write_next_bytes(fid, [image_id, point2D_id], "ii")
+
+
+def detect_model_format(path, ext):
+    if os.path.isfile(os.path.join(path, "cameras"  + ext)) and \
+       os.path.isfile(os.path.join(path, "images"   + ext)) and \
+       os.path.isfile(os.path.join(path, "points3D" + ext)):
+        print("Detected model format: '" + ext + "'")
+        return True
+
+    return False
+
+
+def read_model(path, ext=""):
+    # try to detect the extension automatically
+    if ext == "":
+        if detect_model_format(path, ".bin"):
+            ext = ".bin"
+        elif detect_model_format(path, ".txt"):
+            ext = ".txt"
+        else:
+            print("Provide model format: '.bin' or '.txt'")
+            return
+
+    if ext == ".txt":
+        cameras = read_cameras_text(os.path.join(path, "cameras" + ext))
+        images = read_images_text(os.path.join(path, "images" + ext))
+        points3D = read_points3D_text(os.path.join(path, "points3D") + ext)
+    else:
+        cameras = read_cameras_binary(os.path.join(path, "cameras" + ext))
+        images = read_images_binary(os.path.join(path, "images" + ext))
+        points3D = read_points3D_binary(os.path.join(path, "points3D") + ext)
+    return cameras, images, points3D
+
+
+def write_model(cameras, images, points3D, path, ext=".bin"):
+    if ext == ".txt":
+        write_cameras_text(cameras, os.path.join(path, "cameras" + ext))
+        write_images_text(images, os.path.join(path, "images" + ext))
+        write_points3D_text(points3D, os.path.join(path, "points3D") + ext)
+    else:
+        write_cameras_binary(cameras, os.path.join(path, "cameras" + ext))
+        write_images_binary(images, os.path.join(path, "images" + ext))
+        write_points3D_binary(points3D, os.path.join(path, "points3D") + ext)
+    return cameras, images, points3D
+
+
+def qvec2rotmat(qvec):
+    return np.array([
+        [1 - 2 * qvec[2]**2 - 2 * qvec[3]**2,
+         2 * qvec[1] * qvec[2] - 2 * qvec[0] * qvec[3],
+         2 * qvec[3] * qvec[1] + 2 * qvec[0] * qvec[2]],
+        [2 * qvec[1] * qvec[2] + 2 * qvec[0] * qvec[3],
+         1 - 2 * qvec[1]**2 - 2 * qvec[3]**2,
+         2 * qvec[2] * qvec[3] - 2 * qvec[0] * qvec[1]],
+        [2 * qvec[3] * qvec[1] - 2 * qvec[0] * qvec[2],
+         2 * qvec[2] * qvec[3] + 2 * qvec[0] * qvec[1],
+         1 - 2 * qvec[1]**2 - 2 * qvec[2]**2]])
+
+
+def rotmat2qvec(R):
+    Rxx, Ryx, Rzx, Rxy, Ryy, Rzy, Rxz, Ryz, Rzz = R.flat
+    K = np.array([
+        [Rxx - Ryy - Rzz, 0, 0, 0],
+        [Ryx + Rxy, Ryy - Rxx - Rzz, 0, 0],
+        [Rzx + Rxz, Rzy + Ryz, Rzz - Rxx - Ryy, 0],
+        [Ryz - Rzy, Rzx - Rxz, Rxy - Ryx, Rxx + Ryy + Rzz]]) / 3.0
+    eigvals, eigvecs = np.linalg.eigh(K)
+    qvec = eigvecs[[3, 0, 1, 2], np.argmax(eigvals)]
+    if qvec[0] < 0:
+        qvec *= -1
+    return qvec
+

hoho/vis.py

@@ -0,0 +1,171 @@
+import trimesh
+import numpy as np
+from copy import deepcopy
+from PIL import Image
+
+from . import color_mappings
+
+def line(p1, p2, c=(255,0,0), resolution=10, radius=0.05):
+    '''draws a 3d cylinder along the line (p1, p2)'''
+    # check colors
+    if len(c) == 1:
+        c = [c[0]]*4
+    elif len(c) == 3:
+        c = [*c, 255]
+    elif len(c) != 4:
+        raise ValueError(f'{c} is not a valid color (must have 1,3, or 4 elements).')
+        
+    # compute length and direction of segment
+    p1, p2 = np.asarray(p1), np.asarray(p2)
+    l = np.linalg.norm(p2-p1)
+    
+    direction = (p2 - p1) / l
+    
+    # point z along direction of segment
+    T = np.eye(4)
+    T[:3, 2] = direction
+    T[:3, 3] = (p1+p2)/2
+    
+    #reorthogonalize basis
+    b0, b1 = T[:3, 0], T[:3, 1]
+    if np.abs(np.dot(b0, direction)) < np.abs(np.dot(b1, direction)):
+        T[:3, 1] = -np.cross(b0, direction)
+    else:
+        T[:3, 0] = np.cross(b1, direction)
+    
+    # generate and transform mesh
+    mesh = trimesh.primitives.Cylinder(radius=radius, height=l, transform=T)
+    
+    # apply uniform color
+    mesh.visual.vertex_colors = np.ones_like(mesh.visual.vertex_colors)*c
+         
+    return mesh
+
+def show_wf(row, radius=10, show_vertices=False, vertex_color=(255,0,0, 255)):
+    EDGE_CLASSES = ['eave',
+                    'ridge',
+                    'step_flashing',
+                    'rake',
+                    'flashing',
+                    'post',
+                    'valley',
+                    'hip',
+                    'transition_line']
+    out_meshes = []
+    if show_vertices:
+        out_meshes.extend([trimesh.primitives.Sphere(radius=radius+5, center = center, color=vertex_color) for center in row['wf_vertices']])
+        for m in out_meshes:
+            m.visual.vertex_colors = np.ones_like(m.visual.vertex_colors)*vertex_color
+    if 'edge_semantics' not in row:
+        print ("Warning: edge semantics is not here, skipping")
+        out_meshes.extend([line(a,b, radius=radius, c=(214, 251, 248)) for a,b in np.stack([*row['wf_vertices']])[np.stack(row['wf_edges'])]])
+    elif len(np.stack(row['wf_edges'])) ==  len(row['edge_semantics']):
+        out_meshes.extend([line(a,b, radius=radius, c=color_mappings.gestalt_color_mapping[EDGE_CLASSES[cls_id]]) for (a,b), cls_id in zip(np.stack([*row['wf_vertices']])[np.stack(row['wf_edges'])], row['edge_semantics'])])
+    else:
+        print ("Warning: edge semantics has different length compared to edges, skipping semantics")
+        out_meshes.extend([line(a,b, radius=radius, c=(214, 251, 248)) for a,b in np.stack([*row['wf_vertices']])[np.stack(row['wf_edges'])]])
+    return out_meshes
+    # return [line(a,b, radius=radius, c=color_mappings.edge_colors[cls_id]) for (a,b), cls_id in zip(np.stack([*row['wf_vertices']])[np.stack(row['wf_edges'])], row['edge_semantics'])]
+
+
+def show_grid(edges, meshes=None, row_length=5):
+    '''
+        edges: list of list of meshes
+        meshes: optional corresponding list of meshes
+        row_length: number of meshes per row
+  
+        returns trimesh.Scene()
+    '''
+    
+    T = np.eye(4)
+    out = []
+    edges = [sum(e[1:], e[0]) for e in edges]
+    row_height = 1.1 * max((e.extents for e in edges), key=lambda e: e[1])[1]
+    col_width = 1.1 * max((e.extents for e in edges), key=lambda e: e[0])[0]
+    # print(row_height, col_width)
+    
+    if meshes is None:
+        meshes = [None]*len(edges)
+
+    for i, (gt, mesh) in enumerate(zip(edges, meshes), start=0):
+        mesh = deepcopy(mesh)
+        gt = deepcopy(gt)
+
+        if i%row_length != 0:
+            T[0, 3] += col_width
+
+        else:
+            T[0, 3] = 0
+            T[1, 3] += row_height
+
+        # print(T[0,3]/col_width, T[2,3]/row_height)
+        
+        if mesh is not None:
+            mesh.apply_transform(T)
+            out.append(mesh)
+                            
+        gt.apply_transform(T)
+        out.append(gt)
+        
+                            
+        out.extend([mesh, gt])
+
+            
+    return trimesh.Scene(out)
+
+
+
+
+def visualize_order_images(row_order):
+    return create_image_grid(row_order['ade20k'] + row_order['gestalt'] + [visualize_depth(dm) for dm in row_order['depthcm']], num_per_row=len(row_order['ade20k']))
+
+def create_image_grid(images, target_length=312, num_per_row=2):
+    # Calculate the target size for the first image
+    first_img = images[0]
+    aspect_ratio = first_img.width / first_img.height
+    new_width = int((target_length ** 2 * aspect_ratio) ** 0.5)
+    new_height = int((target_length ** 2 / aspect_ratio) ** 0.5)
+    
+    # Resize the first image
+    resized_images = [img.resize((new_width, new_height), Image.Resampling.LANCZOS) for img in images]
+    
+    # Calculate the grid size
+    num_rows = (len(resized_images) + num_per_row - 1) // num_per_row
+    grid_width = new_width * num_per_row
+    grid_height = new_height * num_rows
+    
+    # Create a new image for the grid
+    grid_img = Image.new('RGB', (grid_width, grid_height))
+    
+    # Paste the images into the grid
+    for i, img in enumerate(resized_images):
+        x_offset = (i % num_per_row) * new_width
+        y_offset = (i // num_per_row) * new_height
+        grid_img.paste(img, (x_offset, y_offset))
+    
+    return grid_img
+
+
+import matplotlib.pyplot as plt
+
+def visualize_depth(depth, min_depth=None, max_depth=None, cmap='rainbow'):
+    depth = np.array(depth)
+    
+    if min_depth is None:
+        min_depth = np.min(depth)
+    if max_depth is None:
+        max_depth = np.max(depth)
+    
+    
+    # Normalize the depth to be between 0 and 1
+    depth = (depth - min_depth) / (max_depth - min_depth)
+    depth = np.clip(depth, 0, 1)
+    
+    # Use the matplotlib colormap to convert the depth to an RGB image
+    cmap = plt.get_cmap(cmap)
+    depth_image = (cmap(depth) * 255).astype(np.uint8)
+    
+    # Convert the depth image to a PIL image
+    depth_image = Image.fromarray(depth_image)
+    
+    return depth_image

hoho/viz3d.py

@@ -0,0 +1,302 @@
+
+"""
+Copyright [2022] [Paul-Edouard Sarlin and Philipp Lindenberger]
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+
+3D visualization based on plotly.
+Works for a small number of points and cameras, might be slow otherwise.
+
+1) Initialize a figure with `init_figure`
+2) Add 3D points, camera frustums, or both as a pycolmap.Reconstruction
+
+Written by Paul-Edouard Sarlin and Philipp Lindenberger.
+"""
+# Slightly modified by Dmytro Mishkin
+
+from typing import Optional
+import numpy as np
+import pycolmap
+import plotly.graph_objects as go
+
+
+### Some helper functions for geometry
+def qvec2rotmat(qvec):
+    return np.array([
+        [1 - 2 * qvec[2]**2 - 2 * qvec[3]**2,
+         2 * qvec[1] * qvec[2] - 2 * qvec[0] * qvec[3],
+         2 * qvec[3] * qvec[1] + 2 * qvec[0] * qvec[2]],
+        [2 * qvec[1] * qvec[2] + 2 * qvec[0] * qvec[3],
+         1 - 2 * qvec[1]**2 - 2 * qvec[3]**2,
+         2 * qvec[2] * qvec[3] - 2 * qvec[0] * qvec[1]],
+        [2 * qvec[3] * qvec[1] - 2 * qvec[0] * qvec[2],
+         2 * qvec[2] * qvec[3] + 2 * qvec[0] * qvec[1],
+         1 - 2 * qvec[1]**2 - 2 * qvec[2]**2]])
+
+
+def to_homogeneous(points):
+    pad = np.ones((points.shape[:-1]+(1,)), dtype=points.dtype)
+    return np.concatenate([points, pad], axis=-1)
+
+def t_to_proj_center(qvec, tvec):
+    Rr = qvec2rotmat(qvec)
+    tt = (-Rr.T) @ tvec
+    return tt
+
+def calib(params):
+    out = np.eye(3)
+    if len(params) == 3:
+        out[0,0] = params[0]
+        out[1,1] = params[0]
+        out[0,2] = params[1]
+        out[1,2] = params[2]
+    else:
+        out[0,0] = params[0]
+        out[1,1] = params[1]
+        out[0,2] = params[2]
+        out[1,2] = params[3]
+    return out
+
+
+### Plotting functions
+
+def init_figure(height: int = 800) -> go.Figure:
+    """Initialize a 3D figure."""
+    fig = go.Figure()
+    axes = dict(
+        visible=False,
+        showbackground=False,
+        showgrid=False,
+        showline=False,
+        showticklabels=True,
+        autorange=True,
+    )
+    fig.update_layout(
+        template="plotly_dark",
+        height=height,
+        scene_camera=dict(
+            eye=dict(x=0., y=-.1, z=-2),
+            up=dict(x=0, y=-1., z=0),
+            projection=dict(type="orthographic")),
+        scene=dict(
+            xaxis=axes,
+            yaxis=axes,
+            zaxis=axes,
+            aspectmode='data',
+            dragmode='orbit',
+        ),
+        margin=dict(l=0, r=0, b=0, t=0, pad=0),
+        legend=dict(
+            orientation="h",
+            yanchor="top",
+            y=0.99,
+            xanchor="left",
+            x=0.1
+        ),
+    )
+    return fig
+
+
+def plot_lines_3d(
+        fig: go.Figure,
+        pts: np.ndarray,
+        color: str = 'rgba(255, 255, 255, 1)',
+        ps: int = 2,
+        colorscale: Optional[str] = None,
+        name: Optional[str] = None):
+    """Plot a set of 3D points."""
+    x = pts[..., 0]
+    y = pts[..., 1]
+    z = pts[..., 2]
+    traces = [go.Scatter3d(x=x1, y=y1, z=z1,
+                        mode='lines',
+                        line=dict(color=color, width=2)) for x1, y1, z1 in zip(x,y,z)]
+    for t in traces:
+        fig.add_trace(t)
+    fig.update_traces(showlegend=False)
+
+
+def plot_points(
+        fig: go.Figure,
+        pts: np.ndarray,
+        color: str = 'rgba(255, 0, 0, 1)',
+        ps: int = 2,
+        colorscale: Optional[str] = None,
+        name: Optional[str] = None):
+    """Plot a set of 3D points."""
+    x, y, z = pts.T
+    tr = go.Scatter3d(
+        x=x, y=y, z=z, mode='markers', name=name, legendgroup=name,
+        marker=dict(
+            size=ps, color=color, line_width=0.0, colorscale=colorscale))
+    fig.add_trace(tr)
+
+def plot_camera(
+        fig: go.Figure,
+        R: np.ndarray,
+        t: np.ndarray,
+        K: np.ndarray,
+        color: str = 'rgb(0, 0, 255)',
+        name: Optional[str] = None,
+        legendgroup: Optional[str] = None,
+        size: float = 1.0):
+    """Plot a camera frustum from pose and intrinsic matrix."""
+    W, H = K[0, 2]*2, K[1, 2]*2
+    corners = np.array([[0, 0], [W, 0], [W, H], [0, H], [0, 0]])
+    if size is not None:
+        image_extent = max(size * W / 1024.0, size * H / 1024.0)
+        world_extent = max(W, H) / (K[0, 0] + K[1, 1]) / 0.5
+        scale = 0.5 * image_extent / world_extent
+    else:
+        scale = 1.0
+    corners = to_homogeneous(corners) @ np.linalg.inv(K).T
+    corners = (corners / 2 * scale) @ R.T + t
+
+    x, y, z = corners.T
+    rect = go.Scatter3d(
+        x=x, y=y, z=z, line=dict(color=color), legendgroup=legendgroup,
+        name=name, marker=dict(size=0.0001), showlegend=False)
+    fig.add_trace(rect)
+
+    x, y, z = np.concatenate(([t], corners)).T
+    i = [0, 0, 0, 0]
+    j = [1, 2, 3, 4]
+    k = [2, 3, 4, 1]
+
+    pyramid = go.Mesh3d(
+        x=x, y=y, z=z, color=color, i=i, j=j, k=k,
+        legendgroup=legendgroup, name=name, showlegend=False)
+    fig.add_trace(pyramid)
+    triangles = np.vstack((i, j, k)).T
+    vertices = np.concatenate(([t], corners))
+    tri_points = np.array([
+        vertices[i] for i in triangles.reshape(-1)
+    ])
+
+    x, y, z = tri_points.T
+    pyramid = go.Scatter3d(
+        x=x, y=y, z=z, mode='lines', legendgroup=legendgroup,
+        name=name, line=dict(color=color, width=1), showlegend=False)
+    fig.add_trace(pyramid)
+
+
+def plot_camera_colmap(
+        fig: go.Figure,
+        image: pycolmap.Image,
+        camera: pycolmap.Camera,
+        name: Optional[str] = None,
+        **kwargs):
+    """Plot a camera frustum from PyCOLMAP objects"""
+    intr = calib(camera.params)
+    if intr[0][0] > 10000:
+        print("Bad camera")
+        return
+    plot_camera(
+        fig,
+        qvec2rotmat(image.qvec).T,
+        t_to_proj_center(image.qvec, image.tvec),
+        intr,#calibration_matrix(),
+        name=name or str(image.id),
+        **kwargs)
+
+
+def plot_cameras(
+        fig: go.Figure,
+        reconstruction,#: pycolmap.Reconstruction,
+        **kwargs):
+    """Plot a camera as a cone with camera frustum."""
+    for image_id, image in reconstruction["images"].items():
+        plot_camera_colmap(
+            fig, image, reconstruction["cameras"][image.camera_id], **kwargs)
+
+
+def plot_reconstruction(
+        fig: go.Figure,
+        rec,
+        color: str = 'rgb(0, 0, 255)',
+        name: Optional[str] = None,
+        points: bool = True,
+        cameras: bool = True,
+        cs: float = 1.0,
+        single_color_points=False,
+        camera_color='rgba(0, 255, 0, 0.5)'):
+    # rec is result of loading reconstruction from "read_write_colmap.py"
+    # Filter outliers
+    xyzs = []
+    rgbs = []
+    for k, p3D in rec['points'].items():
+        xyzs.append(p3D.xyz)
+        rgbs.append(p3D.rgb)
+
+    if points:
+        plot_points(fig, np.array(xyzs), color=color if single_color_points else np.array(rgbs), ps=1, name=name)
+    if cameras:
+        plot_cameras(fig, rec, color=camera_color, legendgroup=name, size=cs)
+
+
+def plot_pointcloud(
+        fig: go.Figure,
+        pts: np.ndarray,
+        colors: np.ndarray,
+        ps: int = 2,
+        name: Optional[str] = None):
+    """Plot a set of 3D points."""
+    plot_points(fig, np.array(pts), color=colors, ps=ps, name=name)
+
+
+def plot_triangle_mesh(
+        fig: go.Figure,
+        vert: np.ndarray,
+        colors: np.ndarray,
+        triangles: np.ndarray,
+        name: Optional[str] = None):
+    """Plot a triangle mesh."""
+    tr = go.Mesh3d(
+        x=vert[:,0],
+        y=vert[:,1],
+        z=vert[:,2],
+        vertexcolor = np.clip(255*colors, 0, 255),
+        # i, j and k give the vertices of triangles
+        # here we represent the 4 triangles of the tetrahedron surface
+        i=triangles[:,0],
+        j=triangles[:,1],
+        k=triangles[:,2],
+        name=name,
+        showscale=False
+    )
+    fig.add_trace(tr)
+
+def plot_estimate_and_gt(pred_vertices, pred_connections, gt_vertices=None, gt_connections=None):
+    fig3d = init_figure()
+    c1 = (30, 20, 255)
+    img_color = [c1 for _ in range(len(pred_vertices))]
+    plot_points(fig3d, pred_vertices, color = img_color, ps = 10)  
+    lines = []
+    for c in pred_connections:
+        v1 = pred_vertices[c[0]]
+        v2 = pred_vertices[c[1]]
+        lines.append(np.stack([v1, v2], axis=0))
+    plot_lines_3d(fig3d, np.array(lines), img_color, ps=4)  
+    if gt_vertices is not None:
+        c2 = (30, 255, 20)
+        img_color2 = [c2 for _ in range(len(gt_vertices))]
+        plot_points(fig3d, gt_vertices, color = img_color2, ps = 10)  
+        if gt_connections is not None:
+            gt_lines = []
+            for c in gt_connections:
+                v1 = gt_vertices[c[0]]
+                v2 = gt_vertices[c[1]]
+                gt_lines.append(np.stack([v1, v2], axis=0))
+        plot_lines_3d(fig3d, np.array(gt_lines), img_color2, ps=4)        
+    fig3d.show()
+    return fig3d

hoho/wed.py

@@ -0,0 +1,107 @@
+from scipy.spatial.distance import cdist
+from scipy.optimize import linear_sum_assignment
+import numpy as np 
+
+
+def preregister_mean_std(verts_to_transform, target_verts, single_scale=True):
+    mu_target = target_verts.mean(axis=0)
+    mu_in = verts_to_transform.mean(axis=0)
+    std_target = np.std(target_verts, axis=0)
+    std_in = np.std(verts_to_transform, axis=0)
+    
+    if np.any(std_in == 0):
+        std_in[std_in == 0] = 1
+    if np.any(std_target == 0):
+        std_target[std_target == 0] = 1
+    if np.any(np.isnan(std_in)):
+        std_in[np.isnan(std_in)] = 1
+    if np.any(np.isnan(std_target)):
+        std_target[np.isnan(std_target)] = 1
+        
+    if single_scale:
+        std_target = np.linalg.norm(std_target)
+        std_in = np.linalg.norm(std_in)
+    
+    transformed_verts = (verts_to_transform - mu_in) / std_in
+    transformed_verts = transformed_verts * std_target + mu_target
+    
+    return transformed_verts
+
+
+def update_cv(cv, gt_vertices):
+    if cv < 0:
+        diameter = cdist(gt_vertices, gt_vertices).max()
+        # Cost of adding or deleting a vertex is set to -cv times the diameter of the ground truth wireframe
+        cv = -cv * diameter
+    return cv
+
+def compute_WED(pd_vertices, pd_edges, gt_vertices, gt_edges, cv_ins=-1/2, cv_del=-1/4, ce=1.0, normalized=True, preregister=True, single_scale=True):
+    '''The function computes the Wireframe Edge Distance (WED) between two graphs.
+    pd_vertices: list of predicted vertices
+    pd_edges: list of predicted edges
+    gt_vertices: list of ground truth vertices
+    gt_edges: list of ground truth edges
+    cv_ins: vertex insertion cost: if positive, the cost in centimeters of inserting vertex, if negative, multiplies diameter to compute cost (default is -1/2)
+    cv_del: vertex deletion cost: if positive, the cost in centimeters of deleting a vertex, if negative, multiplies diameter to compute cost (default is -1/4)
+    ce: edge cost (multiplier of the edge length for edge deletion and insertion, default is 1.0)
+    normalized: if True, the WED is normalized by the total length of the ground truth edges
+    preregister: if True, the predicted vertices have their mean and scale matched to the ground truth vertices
+    '''
+    
+    pd_vertices = np.array(pd_vertices)
+    gt_vertices = np.array(gt_vertices)
+    pd_edges = np.array(pd_edges)
+    gt_edges = np.array(gt_edges)        
+    
+        
+    cv_del = update_cv(cv_del, gt_vertices)
+    cv_ins = update_cv(cv_ins, gt_vertices)
+    
+    # Step 0: Prenormalize / preregister
+    if preregister:
+        pd_vertices = preregister_mean_std(pd_vertices, gt_vertices, single_scale=single_scale)
+    
+
+    # Step 1: Bipartite Matching
+    distances = cdist(pd_vertices, gt_vertices, metric='euclidean')
+    row_ind, col_ind = linear_sum_assignment(distances)
+        
+    
+    # Step 2: Vertex Translation
+    translation_costs = np.sum(distances[row_ind, col_ind])
+    
+    # Step 3: Vertex Deletion
+    unmatched_pd_indices = set(range(len(pd_vertices))) - set(row_ind)
+    deletion_costs = cv_del * len(unmatched_pd_indices)  
+    
+    # Step 4: Vertex Insertion
+    unmatched_gt_indices = set(range(len(gt_vertices))) - set(col_ind)
+    insertion_costs = cv_ins * len(unmatched_gt_indices)  
+    
+    # Step 5: Edge Deletion and Insertion
+    updated_pd_edges = [(col_ind[np.where(row_ind == edge[0])[0][0]], col_ind[np.where(row_ind == edge[1])[0][0]]) for edge in pd_edges if len(edge)==2 and edge[0] in row_ind and edge[1] in row_ind]
+    pd_edges_set = set(map(tuple, [set(edge) for edge in updated_pd_edges]))
+    gt_edges_set = set(map(tuple, [set(edge) for edge in gt_edges]))
+
+    
+    # Delete edges not in ground truth
+    edges_to_delete = pd_edges_set - gt_edges_set
+    
+    vert_tf = [np.where(col_ind == v)[0][0] if v in col_ind else 0 for v in range(len(gt_vertices))]
+    deletion_edge_costs = ce * sum(np.linalg.norm(pd_vertices[vert_tf[edge[0]]] - pd_vertices[vert_tf[edge[1]]]) for edge in edges_to_delete if len(edge) == 2)
+
+    
+    # Insert missing edges from ground truth
+    edges_to_insert = gt_edges_set - pd_edges_set
+    insertion_edge_costs = ce * sum(np.linalg.norm(gt_vertices[edge[0]] - gt_vertices[edge[1]]) for edge in edges_to_insert if len(edge) == 2) 
+    
+    # Step 6: Calculation of WED
+    WED = translation_costs + deletion_costs + insertion_costs + deletion_edge_costs + insertion_edge_costs
+
+    
+    if normalized:
+        total_length_of_gt_edges = np.linalg.norm((gt_vertices[gt_edges[:, 0]] - gt_vertices[gt_edges[:, 1]]), axis=1).sum()
+        WED = WED / total_length_of_gt_edges
+        
+    # print ("Total length", total_length_of_gt_edges)
+    return WED

requirements.txt

@@ -0,0 +1,10 @@
+datasets
+ipywidgets
+matplotlib
+numpy
+pillow
+plotly
+pycolmap
+scipy
+trimesh
+webdataset

setup.py

@@ -0,0 +1,16 @@
+from setuptools import setup, find_packages
+import glob
+
+
+with open('requirements.txt') as f:
+    required = f.read().splitlines()
+
+setup(name='hoho',
+	version='0.0.4',
+	description='Tools and utilites for the HoHo Dataset and S23DR Competition',
+	url='usm3d.github.io',
+	author='Jack Langerman, Dmytro Mishkin, S23DR Orgainizing Team',
+	author_email='hoho@jackml.com',
+	install_requires=required,
+	packages=find_packages(),
+	include_package_data=True)