Cleanup events and training task (#500)

* interface clean up

* clean up events

* revert rename

* fix unit tests

test/test_beam_search.py

@@ -3,7 +3,7 @@
 import dynet as dy
 
 from xnmt.modelparts.attenders import MlpAttender
-from xnmt import batchers, events
+from xnmt import batchers, event_trigger, events
 from xnmt.modelparts.bridges import CopyBridge
 from xnmt.modelparts.decoders import AutoRegressiveDecoder
 from xnmt.modelparts.embedders import SimpleWordEmbedder
@@ -44,7 +44,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))
@@ -81,7 +81,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))
@@ -117,7 +117,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))
@@ -156,7 +156,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))

test/test_decoders.py

@@ -3,7 +3,7 @@
 import dynet as dy
 
 from xnmt.modelparts.attenders import MlpAttender
-from xnmt import batchers, events
+from xnmt import batchers, event_trigger, events
 from xnmt.modelparts.bridges import CopyBridge
 from xnmt.modelparts.decoders import AutoRegressiveDecoder
 from xnmt.modelparts.embedders import SimpleWordEmbedder
@@ -44,7 +44,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))
@@ -83,7 +83,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))
@@ -120,7 +120,7 @@ def setUp(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(False)
+    event_trigger.set_train(False)
 
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.model.trg_reader.read_sents("examples/data/head.en"))

test/test_encoders.py

@@ -17,7 +17,7 @@
 from xnmt.modelparts.transforms import NonLinear
 from xnmt.models.translators import DefaultTranslator
 from xnmt.vocabs import Vocab
-from xnmt import batchers, events
+from xnmt import batchers, event_trigger, events
 
 class TestEncoder(unittest.TestCase):
 
@@ -32,18 +32,11 @@ def setUp(self):
     self.src_data = list(self.src_reader.read_sents("examples/data/head.ja"))
     self.trg_data = list(self.trg_reader.read_sents("examples/data/head.en"))
 
-  @events.register_xnmt_event
-  def set_train(self, val):
-    pass
-  @events.register_xnmt_event
-  def start_sent(self, src):
-    pass
-
   def assert_in_out_len_equal(self, model):
     dy.renew_cg()
-    self.set_train(True)
+    event_trigger.set_train(True)
     src = self.src_data[0]
-    self.start_sent(src)
+    event_trigger.start_sent(src)
     embeddings = model.src_embedder.embed_sent(src)
     encodings = model.encoder.transduce(embeddings)
     self.assertEqual(len(embeddings), len(encodings))
@@ -119,11 +112,11 @@ def test_py_lstm_encoder_len(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.set_train(True)
+    event_trigger.set_train(True)
     for sent_i in range(10):
       dy.renew_cg()
       src = self.src_data[sent_i].create_padded_sent(4 - (self.src_data[sent_i].sent_len() % 4))
-      self.start_sent(src)
+      event_trigger.start_sent(src)
       embeddings = model.src_embedder.embed_sent(src)
       encodings = model.encoder.transduce(embeddings)
       self.assertEqual(int(math.ceil(len(embeddings) / float(4))), len(encodings))
@@ -149,11 +142,11 @@ def test_py_lstm_mask(self):
     train_src, _ = \
       batcher.pack(self.src_data, self.trg_data)
 
-    self.set_train(True)
+    event_trigger.set_train(True)
     for sent_i in range(3):
       dy.renew_cg()
       src = train_src[sent_i]
-      self.start_sent(src)
+      event_trigger.start_sent(src)
       embeddings = model.src_embedder.embed_sent(src)
       encodings = model.encoder.transduce(embeddings)
       if train_src[sent_i].mask is None:

test/test_segmenting.py

@@ -11,7 +11,7 @@
 from xnmt.modelparts.decoders import AutoRegressiveDecoder
 from xnmt.modelparts.embedders import SimpleWordEmbedder
 import xnmt.events
-from xnmt import batchers
+from xnmt import batchers, event_trigger
 from xnmt.input_readers import PlainTextReader
 from xnmt.input_readers import CharFromWordTextReader
 from xnmt.transducers.recurrent import UniLSTMSeqTransducer
@@ -83,7 +83,7 @@ def setUp(self):
                                     trg_embed_dim=layer_dim,
                                     bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(True)
+    event_trigger.set_train(True)
 
     self.layer_dim = layer_dim
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
@@ -96,7 +96,7 @@ def test_reinforce_loss(self):
     fertility_loss = GlobalFertilityLoss()
     mle_loss = MLELoss()
     loss = CompositeLoss(losses=[mle_loss, fertility_loss]).calc_loss(self.model, self.src[0], self.trg[0])
-    reinforce_loss = self.model.calc_additional_loss(self.trg[0], self.model, loss)
+    reinforce_loss = event_trigger.calc_additional_loss(self.trg[0], self.model, loss)
     pl = self.model.encoder.policy_learning
     # Ensure correct length
     src = self.src[0]
@@ -117,7 +117,7 @@ def test_reinforce_loss(self):
 
   def calc_loss_single_batch(self):
     loss = MLELoss().calc_loss(self.model, self.src[0], self.trg[0])
-    reinforce_loss = self.model.calc_additional_loss(self.trg[0], self.model, loss)
+    reinforce_loss = event_trigger.calc_additional_loss(self.trg[0], self.model, loss)
     return loss, reinforce_loss
 
   def test_gold_input(self):
@@ -134,24 +134,24 @@ def test_sample_input(self):
     self.assertEqual(self.model.encoder.policy_learning.sampling_action, PolicyGradient.SamplingAction.PREDEFINED)
 
   def test_policy_train_test(self):
-    self.model.set_train(True)
+    event_trigger.set_train(True)
     self.calc_loss_single_batch()
     self.assertEqual(self.model.encoder.policy_learning.sampling_action, PolicyGradient.SamplingAction.POLICY_CLP)
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     self.calc_loss_single_batch()
     self.assertEqual(self.model.encoder.policy_learning.sampling_action, PolicyGradient.SamplingAction.POLICY_AMAX)
 
   def test_no_policy_train_test(self):
     self.model.encoder.policy_learning = None
-    self.model.set_train(True)
+    event_trigger.set_train(True)
     self.calc_loss_single_batch()
     self.assertEqual(self.model.encoder.segmenting_action, SegmentingSeqTransducer.SegmentingAction.PURE_SAMPLE)
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     self.calc_loss_single_batch()
     self.assertEqual(self.model.encoder.segmenting_action, SegmentingSeqTransducer.SegmentingAction.PURE_SAMPLE)
 
   def test_sample_during_search(self):
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     self.model.encoder.sample_during_search = True
     self.calc_loss_single_batch()
     self.assertEqual(self.model.encoder.segmenting_action, SegmentingSeqTransducer.SegmentingAction.POLICY)
@@ -199,7 +199,7 @@ def setUp(self):
                                     trg_embed_dim=layer_dim,
                                     bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    self.model.set_train(True)
+    event_trigger.set_train(True)
 
     self.layer_dim = layer_dim
     self.src_data = list(self.model.src_reader.read_sents("examples/data/head.ja"))
@@ -209,7 +209,7 @@ def setUp(self):
     dy.renew_cg(immediate_compute=True, check_validity=True)
 
   def inp_emb(self, idx=0):
-    self.model.start_sent(self.src[idx])
+    event_trigger.start_sent(self.src[idx])
     embed = self.model.src_embedder.embed_sent(self.src[idx])
     return embed
 
@@ -268,19 +268,19 @@ def test_convolution_composer(self):
     enc.segment_composer = ConvolutionComposer(ngram_size=1,
                                                embed_dim=self.layer_dim,
                                                hidden_dim=self.layer_dim)
-    self.model.set_train(True)
+    event_trigger.set_train(True)
     enc.transduce(self.inp_emb(0))
     enc.segment_composer = ConvolutionComposer(ngram_size=3,
                                                embed_dim=self.layer_dim,
                                                hidden_dim=self.layer_dim)
-    self.model.set_train(True)
+    event_trigger.set_train(True)
     enc.transduce(self.inp_emb(0))
 
   def test_transducer_composer(self):
     enc = self.segmenting_encoder
     enc.segment_composer = SeqTransducerComposer(seq_transducer=BiLSTMSeqTransducer(input_dim=self.layer_dim,
                                                                                     hidden_dim=self.layer_dim))
-    self.model.set_train(True)
+    event_trigger.set_train(True)
     enc.transduce(self.inp_emb(0))
 
 if __name__ == "__main__":

test/test_training.py

@@ -21,7 +21,7 @@
 from xnmt.models.translators import DefaultTranslator
 from xnmt.modelparts.scorers import Softmax
 from xnmt.vocabs import Vocab
-from xnmt import sent
+from xnmt import event_trigger, sent
 
 class TestTruncatedBatchTraining(unittest.TestCase):
 
@@ -91,7 +91,7 @@ def test_loss_model1(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model)
 
   def test_loss_model2(self):
@@ -113,7 +113,7 @@ def test_loss_model2(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model, pad_src_to_multiple=4)
 
   def test_loss_model3(self):
@@ -135,7 +135,7 @@ def test_loss_model3(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model)
 
   def test_loss_model4(self):
@@ -157,7 +157,7 @@ def test_loss_model4(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model)
 
 class TestBatchTraining(unittest.TestCase):
@@ -232,7 +232,7 @@ def test_loss_model1(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model)
 
   def test_loss_model2(self):
@@ -254,7 +254,7 @@ def test_loss_model2(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model, pad_src_to_multiple=4)
 
   def test_loss_model3(self):
@@ -276,7 +276,7 @@ def test_loss_model3(self):
                                 scorer=Softmax(input_dim=layer_dim, vocab_size=100),
                                 bridge=CopyBridge(dec_dim=layer_dim, dec_layers=1)),
     )
-    model.set_train(False)
+    event_trigger.set_train(False)
     self.assert_single_loss_equals_batch_loss(model)
 
 

xnmt/eval/tasks.py

@@ -12,7 +12,7 @@
 from xnmt.eval.metrics import LossScore
 from xnmt.losses import FactoredLossExpr, FactoredLossVal
 import xnmt.xnmt_evaluate
-from xnmt import events, reports, utils
+from xnmt import event_trigger, events, reports, utils
 
 class EvalTask(object):
   """
@@ -62,7 +62,7 @@ def eval(self) -> 'EvalScore':
     Returns:
       Evaluated score
     """
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     if self.src_data is None:
       self.src_data, self.ref_data, self.src_batches, self.ref_batches = \
         input_readers.read_parallel_corpus(src_reader=self.model.src_reader,
@@ -126,7 +126,7 @@ def __init__(self, src_file: Union[str,Sequence[str]], ref_file: Union[str,Seque
     self.desc=desc
 
   def eval(self):
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     self.report_corpus_info({"ref_file":self.ref_file})
     self.inference.perform_inference(generator=self.model,
                                      src_file=self.src_file,
@@ -160,7 +160,7 @@ def __init__(self, src_file: Union[str,Sequence[str]], hyp_file: str, model: 'mo
     self.inference = inference or self.model.inference
 
   def eval(self):
-    self.model.set_train(False)
+    event_trigger.set_train(False)
     self.inference.perform_inference(generator=self.model,
                                      src_file=self.src_file,
                                      trg_file=self.hyp_file)