diff options
| author | Étienne Simon <esimon@esimon.eu> | 2015-04-29 19:42:47 -0400 |
|---|---|---|
| committer | Étienne Simon <esimon@esimon.eu> | 2015-04-29 19:43:12 -0400 |
| commit | 12fbc9b96ea1ef2727c87c02fe8d2305235b4d15 (patch) | |
| tree | 2a95141115eb1a0f507a10eb2d40b2e9fe181f99 /convert_data.py | |
| parent | 43e106e6630030dd34813295fe1d07bb86025402 (diff) | |
| download | taxi-12fbc9b96ea1ef2727c87c02fe8d2305235b4d15.tar.gz taxi-12fbc9b96ea1ef2727c87c02fe8d2305235b4d15.zip | |
Change origin_call ids
Diffstat (limited to 'convert_data.py')
| -rwxr-xr-x | convert_data.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/convert_data.py b/convert_data.py index f81b40f..9684fa9 100755 --- a/convert_data.py +++ b/convert_data.py @@ -104,8 +104,9 @@ def convert(input_directory, save_path): h5file = h5py.File(save_path, 'w') split = {} split.update(read_stands(input_directory, h5file)) - split.update(read_taxis(input_directory, h5file, 'test', 'test_')) split.update(read_taxis(input_directory, h5file, 'train', '')) + print 'First origin_call not present in training set: ', len(origin_call_dict) + split.update(read_taxis(input_directory, h5file, 'test', 'test_')) split.update(unique(h5file)) h5file.attrs['split'] = H5PYDataset.create_split_array(split) h5file.flush() |