evaluate.py

import sys, os
sys.path.insert(0, os.path.dirname(__file__))

import pickle as pkl
import zipfile
from lafan1 import extract, utils, benchmarks


"""
Unzips the data, extracts the LaFAN1 train statistics,
then the LaFAN1 test set and evaluates baselines on the test set.
"""

# output location for unzipped bvhs, stats, results, etc...
out_path = os.path.join(os.path.dirname(__file__), 'output')

# the train/test set actors as in the paper
train_actors = ['subject1', 'subject2', 'subject3', 'subject4']
test_actors = ['subject5']


print('Unzipping the data...\n')
lafan_data = os.path.join(os.path.dirname(__file__), 'lafan1', 'lafan1.zip')
bvh_folder = os.path.join(out_path, 'BVH')
with zipfile.ZipFile(lafan_data, "r") as zip_ref:
    if not os.path.exists(bvh_folder):
        os.makedirs(bvh_folder, exist_ok=True)
    zip_ref.extractall(bvh_folder)


print('Retrieving statistics...')
stats_file = os.path.join(out_path, 'train_stats.pkl')
if not os.path.exists(stats_file):
    x_mean, x_std, offsets = extract.get_train_stats(bvh_folder, train_actors)
    with open(stats_file, 'wb') as f:
        pkl.dump({
            'x_mean': x_mean,
            'x_std': x_std,
            'offsets': offsets,
        }, f, protocol=pkl.HIGHEST_PROTOCOL)
else:
    print('  Reusing stats file: ' + stats_file)
    with open(stats_file, 'rb') as f:
        stats = pkl.load(f)
    x_mean = stats['x_mean']
    x_std = stats['x_std']
    offsets = stats['offsets']


# Get test-set for windows of 65 frames, offset by 40 frames
print('\nBuilding the test set...')
X, Q, parents, contacts_l, contacts_r = extract.get_lafan1_set(bvh_folder, test_actors, window=65, offset=40)
print('  Nb of sequences : {}\n'.format(X.shape[0]))

results = benchmarks.benchmark_interpolation(X, Q, x_mean, x_std, offsets, parents, out_path=out_path)

# save the results for validation if desired
with open(os.path.join(out_path, 'results.pkl'), 'wb') as f:
    pkl.dump(results, f, protocol=pkl.HIGHEST_PROTOCOL)


print('Done.')