def get_predictions(data_fp: Path, predictor: Predictor, incl_labels: bool,
vocab: Vocabulary) -> List[Dict[str, Any]]:
'''
:param data_fp: File Path to the dataset file that you wish to predict on
:param predictor: A predictor that can be used to generate predictions
:param incl_labels: Wether or not to include the original gold labels in
the return
:param vocab: Required to get the original gold labels
:returns: A List of dictionaries that store the data read from the dataset
file and with included predictions and optional gold labels
'''
data_samples = read_dataset(data_fp, incl_labels=incl_labels, vocab=vocab)
data_samples = iter(data_samples)
batch_size = 64
data_exists = True
new_data_samples = []
while data_exists:
data_batch = []
for _ in range(batch_size):
try:
data_batch.append(next(data_samples))
except StopIteration:
data_exists = False
if data_batch:
predictions = predictor.predict_batch_json(data_batch)
for prediction, data_sample in zip(predictions, data_batch):
data_sample['prediction'] = prediction['labels']
new_data_samples.append(data_sample)
return new_data_samples
def batched_predict_json(predictor: Predictor,
examples: List[Dict[str, Any]],
batch_size: int = 16) -> List[Dict[str, Any]]:
results = [] # type: List[Dict[str, Any]]
for i in range(0, len(examples), batch_size):
batch_examples = examples[i:i + batch_size]
batch_results = predictor.predict_batch_json(batch_examples)
results.extend(batch_results)
return results
