changes to preprocessor, trainer, inference & seq modeling

- [trainer]rename "batchify" to "make_batch" in trainer
- [trainer]pack (batch_x_pad, seq_len) into batch_x in make_batch for seq labeling, because seq length before pad is needed to make masks
- [trainer]unpack it in data_forward
- [model]shorten model definition
- [inference]build inference class. test_POS_pipeline.py is OK to infer
- [preprocessor]handle pickles in a nicer manner
- [FastNLP] add fastNLP.py as high-level API, not finished yet

fastNLP/action/inference.py

@@ -1,26 +1,116 @@
+import torch
+
+from fastNLP.action.action import Batchifier, SequentialSampler
+from fastNLP.loader.preprocess import load_pickle, DEFAULT_UNKNOWN_LABEL
+
+
 class Inference(object):
     """
     This is an interface focusing on predicting output based on trained models.
-    It does not care about evaluations of the model.
+    It does not care about evaluations of the model, which is different from Tester.
+    This is a high-level model wrapper to be called by FastNLP.
 
     """
 
-    def __init__(self):
-        pass
+    def __init__(self, pickle_path):
+        self.batch_size = 1
+        self.batch_output = []
+        self.iterator = None
+        self.pickle_path = pickle_path
+        self.index2label = load_pickle(self.pickle_path, "id2class.pkl")
+        self.word2index = load_pickle(self.pickle_path, "word2id.pkl")
 
-    def predict(self, model, data):
+    def predict(self, network, data):
         """
-        this is actually a forward pass. shall be shared by Trainer/Tester
-        :param model:
-        :param data:
-        :return result: the output results
+        Perform inference.
+        :param network:
+        :param data: multi-level lists of strings
+        :return result: the model outputs
         """
-        raise NotImplementedError
+        # transform strings into indices
+        data = self.prepare_input(data)
 
-    def prepare_input(self, data_path):
+        # turn on the testing mode; clean up the history
+        self.mode(network, test=True)
+
+        self.iterator = iter(Batchifier(SequentialSampler(data), self.batch_size, drop_last=False))
+
+        num_iter = len(data) // self.batch_size
+
+        for step in range(num_iter):
+            batch_x = self.batchify(data)
+
+            prediction = self.data_forward(network, batch_x)
+
+            self.batch_output.append(prediction)
+
+        return self.prepare_output(self.batch_output)
+
+    def mode(self, network, test=True):
+        if test:
+            network.eval()
+        else:
+            network.train()
+        self.batch_output.clear()
+
+    def data_forward(self, network, x):
         """
-        This can also be shared.
-        :param data_path:
+        This is only for sequence labeling with CRF decoder. To do: more general ?
+        :param network:
+        :param x:
         :return:
         """
-        raise NotImplementedError
+        seq_len = [len(seq) for seq in x]
+        x = torch.Tensor(x).long()
+        y = network(x)
+        prediction = network.prediction(y, seq_len)
+        # To do: hide framework
+        results = torch.Tensor(prediction).view(-1, )
+        return list(results.data)
+
+    def batchify(self, data):
+        indices = next(self.iterator)
+        batch_x = [data[idx] for idx in indices]
+        batch_x = self.pad(batch_x)
+        return batch_x
+
+    @staticmethod
+    def pad(batch, fill=0):
+        """
+        Pad a batch of samples to maximum length.
+        :param batch: list of list
+        :param fill: word index to pad, default 0.
+        :return: a padded batch
+        """
+        max_length = max([len(x) for x in batch])
+        for idx, sample in enumerate(batch):
+            if len(sample) < max_length:
+                batch[idx] = sample + [fill * (max_length - len(sample))]
+        return batch
+
+    def prepare_input(self, data):
+        """
+        Transform three-level list of strings into that of index.
+        :param data:
+        [
+            [word_11, word_12, ...],
+            [word_21, word_22, ...],
+            ...
+        ]
+        """
+        data_index = []
+        default_unknown_index = self.word2index[DEFAULT_UNKNOWN_LABEL]
+        for example in data:
+            data_index.append([self.word2index.get(w, default_unknown_index) for w in example])
+        return data_index
+
+    def prepare_output(self, batch_outputs):
+        """
+        Transform list of batch outputs into strings.
+        :param batch_outputs: list of list [num_batch, tag_seq_length]
+        :return:
+        """
+        results = []
+        for batch in batch_outputs:
+            results.append([self.index2label[int(x.data)] for x in batch])
+        return results

fastNLP/action/trainer.py

@@ -86,7 +86,7 @@ class BaseTrainer(Action):
 
             # training iterations in one epoch
             for step in range(iterations):
-                batch_x, batch_y = self.batchify(data_train)  # pad ?
+                batch_x, batch_y = self.make_batch(data_train)
 
                 prediction = self.data_forward(network, batch_x)
 
@@ -180,7 +180,7 @@ class BaseTrainer(Action):
         """
         raise NotImplementedError
 
-    def batchify(self, data, output_length=True):
+    def make_batch(self, data, output_length=True):
         """
         1. Perform batching from data and produce a batch of training data.
         2. Add padding.
@@ -191,9 +191,12 @@ class BaseTrainer(Action):
                     [[word_21, word_22, word_23], [label_21. label_22]],  # sample 2
                     ...
                 ]
-        :return batch_x: list. Each entry is a list of features of a sample. [batch_size, max_len]
+        :return (batch_x, seq_len): tuple of two elements, if output_length is true.
+                     batch_x: list. Each entry is a list of features of a sample. [batch_size, max_len]
+                     seq_len: list. The length of the pre-padded sequence, if output_length is True.
                  batch_y: list. Each entry is a list of labels of a sample.  [batch_size, num_labels]
-                 seq_len: list. The length of the pre-padded sequence, if output_length is True.
+
+                 return batch_x and batch_y, if output_length is False
         """
         indices = next(self.iterator)
         batch = [data[idx] for idx in indices]
@@ -202,7 +205,7 @@ class BaseTrainer(Action):
         batch_x_pad = self.pad(batch_x)
         if output_length:
             seq_len = [len(x) for x in batch_x]
-            return batch_x_pad, batch_y, seq_len
+            return (batch_x_pad, seq_len), batch_y
         else:
             return batch_x_pad, batch_y
 
@@ -292,17 +295,23 @@ class POSTrainer(BaseTrainer):
         data_dev = _pickle.load(open(data_path + "/data_train.pkl", "rb"))
         return data_train, data_dev, 0, 1
 
-    def data_forward(self, network, x):
+    def data_forward(self, network, inputs):
         """
         :param network: the PyTorch model
-        :param x: list of list, [batch_size, max_len]
+        :param inputs: list of list, [batch_size, max_len],
+                        or tuple of (batch_x, seq_len), batch_x == [batch_size, max_len]
         :return y: [batch_size, max_len, tag_size]
         """
-        self.seq_len = [len(seq) for seq in x]
+        # unpack the returned value from make_batch
+        if isinstance(inputs, tuple):
+            x = inputs[0]
+            self.seq_len = inputs[1]
+        else:
+            x = inputs
         x = torch.Tensor(x).long()
         self.batch_size = x.size(0)
         self.max_len = x.size(1)
-        # self.mask = seq_mask(seq_len, self.max_len)
+
         y = network(x)
         return y
 
@@ -325,11 +334,12 @@ class POSTrainer(BaseTrainer):
     def get_loss(self, predict, truth):
         """
         Compute loss given prediction and ground truth.
-        :param predict: prediction label vector, [batch_size, tag_size, tag_size]
+        :param predict: prediction label vector, [batch_size, max_len, tag_size]
         :param truth: ground truth label vector, [batch_size, max_len]
         :return: a scalar
         """
         truth = torch.Tensor(truth)
+        assert truth.shape == (self.batch_size, self.max_len)
         if self.loss_func is None:
             if hasattr(self.model, "loss"):
                 self.loss_func = self.model.loss
@@ -347,6 +357,35 @@ class POSTrainer(BaseTrainer):
         else:
             return False
 
+    def make_batch(self, data, output_length=True):
+        """
+        1. Perform batching from data and produce a batch of training data.
+        2. Add padding.
+        :param data: list. Each entry is a sample, which is also a list of features and label(s).
+            E.g.
+                [
+                    [[word_11, word_12, word_13], [label_11. label_12]],  # sample 1
+                    [[word_21, word_22, word_23], [label_21. label_22]],  # sample 2
+                    ...
+                ]
+        :return (batch_x, seq_len): tuple of two elements, if output_length is true.
+                     batch_x: list. Each entry is a list of features of a sample. [batch_size, max_len]
+                     seq_len: list. The length of the pre-padded sequence, if output_length is True.
+                 batch_y: list. Each entry is a list of labels of a sample.  [batch_size, num_labels]
+
+                 return batch_x and batch_y, if output_length is False
+        """
+        indices = next(self.iterator)
+        batch = [data[idx] for idx in indices]
+        batch_x = [sample[0] for sample in batch]
+        batch_y = [sample[1] for sample in batch]
+        batch_x_pad = self.pad(batch_x)
+        if output_length:
+            seq_len = [len(x) for x in batch_x]
+            return (batch_x_pad, seq_len), batch_y
+        else:
+            return batch_x_pad, batch_y
+
 
 class LanguageModelTrainer(BaseTrainer):
     """
@@ -438,7 +477,7 @@ class ClassTrainer(BaseTrainer):
 
             # training iterations in one epoch
             step = 0
-            for batch_x, batch_y in self.batchify(data_train):
+            for batch_x, batch_y in self.make_batch(data_train):
                 prediction = self.data_forward(network, batch_x)
 
                 loss = self.get_loss(prediction, batch_y)
@@ -533,7 +572,7 @@ class ClassTrainer(BaseTrainer):
         """Apply gradient."""
         self.optimizer.step()
 
-    def batchify(self, data):
+    def make_batch(self, data):
         """Batch and pad data."""
         for indices in self.iterator:
             batch = [data[idx] for idx in indices]
@@ -559,4 +598,4 @@ if __name__ == "__name__":
     train_args = {"epochs": 1, "validate": False, "batch_size": 3, "pickle_path": "./"}
     trainer = BaseTrainer(train_args)
     data_train = [[[1, 2, 3, 4], [0]] * 10] + [[[1, 3, 5, 2], [1]] * 10]
-    trainer.batchify(data=data_train)
+    trainer.make_batch(data=data_train)

fastNLP/fastNLP.py

@@ -0,0 +1,104 @@
+from fastNLP.action.inference import Inference
+from fastNLP.loader.config_loader import ConfigLoader, ConfigSection
+from fastNLP.loader.model_loader import ModelLoader
+
+"""
+mapping from model name to [URL, file_name.class_name]
+Notice that the class of the model should be in "models" directory.
+
+Example:
+    "zh_pos_tag_model": ["www.fudan.edu.cn", "sequence_modeling.SeqLabeling"]
+"""
+FastNLP_MODEL_COLLECTION = {
+    "zh_pos_tag_model": ["www.fudan.edu.cn", "sequence_modeling.SeqLabeling"]
+}
+
+
+class FastNLP(object):
+    """
+    High-level interface for direct model inference.
+    Usage:
+        fastnlp = FastNLP()
+        fastnlp.load("zh_pos_tag_model")
+        text = "这是最好的基于深度学习的中文分词系统。"
+        result = fastnlp.run(text)
+        print(result)  # ["这", "是", "最好", "的", "基于", "深度学习", "的", "中文", "分词", "系统", "。"]
+    """
+
+    def __init__(self, model_dir="./"):
+        self.model_dir = model_dir
+        self.model = None
+
+    def load(self, model_name):
+        """
+        Load a pre-trained FastNLP model together with additional data.
+        :param model_name: str, the name of a FastNLP model.
+        """
+        assert type(model_name) is str
+        if model_name not in FastNLP_MODEL_COLLECTION:
+            raise ValueError("No FastNLP model named {}.".format(model_name))
+
+        if not self.model_exist(model_dir=self.model_dir):
+            self._download(model_name, FastNLP_MODEL_COLLECTION[model_name][0])
+
+        model_class = self._get_model_class(FastNLP_MODEL_COLLECTION[model_name][1])
+
+        model_args = ConfigSection()
+        # To do: customized config file for model init parameters
+        ConfigLoader.load_config(self.model_dir + "default.cfg", model_args)
+
+        model = model_class(model_args)
+
+        # To do: framework independent
+        ModelLoader.load_pytorch(model, self.model_dir + model_name)
+
+        self.model = model
+
+        print("Model loaded. ")
+
+    def run(self, infer_input):
+        """
+        Perform inference over given input using the loaded model.
+        :param infer_input: str, raw text
+        :return results:
+        """
+        infer = Inference()
+        data = infer.prepare_input(infer_input)
+        results = infer.predict(self.model, data)
+        return results
+
+    @staticmethod
+    def _get_model_class(file_class_name):
+        """
+        Feature the class specified by <file_class_name>
+        :param file_class_name: str, contains the name of the Python module followed by the name of the class.
+                Example: "sequence_modeling.SeqLabeling"
+        :return module: the model class
+        """
+        import_prefix = "fastNLP.models."
+        parts = (import_prefix + file_class_name).split(".")
+        from_module = ".".join(parts[:-1])
+        module = __import__(from_module)
+        for sub in parts[1:]:
+            module = getattr(module, sub)
+        return module
+
+    def _load(self, model_dir, model_name):
+        # To do
+        return 0
+
+    def _download(self, model_name, url):
+        """
+        Download the model weights from <url> and save in <self.model_dir>.
+        :param model_name:
+        :param url:
+        """
+        print("Downloading {} from {}".format(model_name, url))
+        # To do
+
+    def model_exist(self, model_dir):
+        """
+        Check whether the desired model is already in the directory.
+        :param model_dir:
+        """
+        pass

fastNLP/loader/base_loader.py

@@ -17,7 +17,7 @@ class BaseLoader(object):
     def load_lines(self):
         with open(self.data_path, "r", encoding="utf=8") as f:
             text = f.readlines()
-        return text
+        return [line.strip() for line in text]
 
 
 class ToyLoader0(BaseLoader):

fastNLP/loader/model_loader.py

@@ -11,9 +11,11 @@ class ModelLoader(BaseLoader):
     def __init__(self, data_name, data_path):
         super(ModelLoader, self).__init__(data_name, data_path)
 
-    def load_pytorch(self, empty_model):
+    @staticmethod
+    def load_pytorch(empty_model, model_path):
         """
         Load model parameters from .pkl files into the empty PyTorch model.
         :param empty_model: a PyTorch model with initialized parameters.
+        :param model_path: str, the path to the saved model.
         """
-        empty_model.load_state_dict(torch.load(self.data_path))
+        empty_model.load_state_dict(torch.load(model_path))

fastNLP/loader/preprocess.py

@@ -1,346 +1,361 @@
-import _pickle
-import os
-
-DEFAULT_PADDING_LABEL = '<pad>'  # dict index = 0
-DEFAULT_UNKNOWN_LABEL = '<unk>'  # dict index = 1
-DEFAULT_RESERVED_LABEL = ['<reserved-2>',
-                          '<reserved-3>',
-                          '<reserved-4>']  # dict index = 2~4
-
-DEFAULT_WORD_TO_INDEX = {DEFAULT_PADDING_LABEL: 0, DEFAULT_UNKNOWN_LABEL: 1,
-                         DEFAULT_RESERVED_LABEL[0]: 2, DEFAULT_RESERVED_LABEL[1]: 3,
-                         DEFAULT_RESERVED_LABEL[2]: 4}
-
-
-# the first vocab in dict with the index = 5
-
-
-class BasePreprocess(object):
-
-    def __init__(self, data, pickle_path):
-        super(BasePreprocess, self).__init__()
-        self.data = data
-        self.pickle_path = pickle_path
-        if not self.pickle_path.endswith('/'):
-            self.pickle_path = self.pickle_path + '/'
-
-
-class POSPreprocess(BasePreprocess):
-    """
-        This class are used to preprocess the pos datasets.
-
-    """
-
-    def __init__(self, data, pickle_path="./", train_dev_split=0):
-        """
-        Preprocess pipeline, including building mapping from words to index, from index to words,
-        from labels/classes to index, from index to labels/classes.
-        :param data: three-level list
-            [
-                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
-                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
-                ...
-            ]
-        :param pickle_path: str, the directory to the pickle files. Default: "./"
-        :param train_dev_split: float in [0, 1]. The ratio of dev data split from training data. Default: 0.
-
-        To do:
-        1. simplify __init__
-        """
-        super(POSPreprocess, self).__init__(data, pickle_path)
-
-        self.pickle_path = pickle_path
-
-        if self.pickle_exist("word2id.pkl"):
-            # load word2index because the construction of the following objects needs it
-            with open(os.path.join(self.pickle_path, "word2id.pkl"), "rb") as f:
-                self.word2index = _pickle.load(f)
-        else:
-            self.word2index, self.label2index = self.build_dict(data)
-            with open(os.path.join(self.pickle_path, "word2id.pkl"), "wb") as f:
-                _pickle.dump(self.word2index, f)
-
-        if self.pickle_exist("class2id.pkl"):
-            with open(os.path.join(self.pickle_path, "class2id.pkl"), "rb") as f:
-                self.label2index = _pickle.load(f)
-        else:
-            with open(os.path.join(self.pickle_path, "class2id.pkl"), "wb") as f:
-                _pickle.dump(self.label2index, f)
-            #something will be wrong if word2id.pkl is found but class2id.pkl is not found
-
-        if not self.pickle_exist("id2word.pkl"):
-            index2word = self.build_reverse_dict(self.word2index)
-            with open(os.path.join(self.pickle_path, "id2word.pkl"), "wb") as f:
-                _pickle.dump(index2word, f)
-
-        if not self.pickle_exist("id2class.pkl"):
-            index2label = self.build_reverse_dict(self.label2index)
-            with open(os.path.join(self.pickle_path, "word2id.pkl"), "wb") as f:
-                _pickle.dump(index2label, f)
-
-        if not self.pickle_exist("data_train.pkl"):
-            data_train = self.to_index(data)
-            if train_dev_split > 0 and not self.pickle_exist("data_dev.pkl"):
-                data_dev = data_train[: int(len(data_train) * train_dev_split)]
-                with open(os.path.join(self.pickle_path, "data_dev.pkl"), "wb") as f:
-                    _pickle.dump(data_dev, f)
-            with open(os.path.join(self.pickle_path, "data_train.pkl"), "wb") as f:
-                _pickle.dump(data_train, f)
-
-    def build_dict(self, data):
-        """
-        Add new words with indices into self.word_dict, new labels with indices into self.label_dict.
-        :param data: three-level list
-            [
-                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
-                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
-                ...
-            ]
-        :return word2index: dict of {str, int}
-                label2index: dict of {str, int}
-        """
-        label2index = {}
-        word2index = DEFAULT_WORD_TO_INDEX
-        for example in data:
-            for word, label in zip(example[0], example[1]):
-                if word not in word2index:
-                    word2index[word] = len(word2index)
-                if label not in label2index:
-                    label2index[label] = len(label2index)
-        return word2index, label2index
-
-    def pickle_exist(self, pickle_name):
-        """
-        :param pickle_name: the filename of target pickle file
-        :return: True if file exists else False
-        """
-        if not os.path.exists(self.pickle_path):
-            os.makedirs(self.pickle_path)
-        file_name = os.path.join(self.pickle_path, pickle_name)
-        if os.path.exists(file_name):
-            return True
-        else:
-            return False
-
-    def build_reverse_dict(self, word_dict):
-        id2word = {word_dict[w]: w for w in word_dict}
-        return id2word
-
-    def to_index(self, data):
-        """
-        Convert word strings and label strings into indices.
-        :param data: three-level list
-            [
-                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
-                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
-                ...
-            ]
-        :return data_index: the shape of data, but each string is replaced by its corresponding index
-        """
-        data_index = []
-        for example in data:
-            word_list = []
-            label_list = []
-            for word, label in zip(example[0], example[1]):
-                word_list.append(self.word2index[word])
-                label_list.append(self.label2index[label])
-            data_index.append([word_list, label_list])
-        return data_index
-
-    @property
-    def vocab_size(self):
-        return len(self.word2index)
-
-    @property
-    def num_classes(self):
-        return len(self.label2index)
-
-
-class ClassPreprocess(BasePreprocess):
-    """
-    Pre-process the classification datasets.
-
-    Params:
-        pickle_path - directory to save result of pre-processing
-    Saves:
-        word2id.pkl
-        id2word.pkl
-        class2id.pkl
-        id2class.pkl
-        embedding.pkl
-        data_train.pkl
-        data_dev.pkl
-        data_test.pkl
-    """
-
-    def __init__(self, pickle_path):
-        # super(ClassPreprocess, self).__init__(data, pickle_path)
-        self.word_dict = None
-        self.label_dict = None
-        self.pickle_path = pickle_path  # save directory
-
-    def process(self, data, save_name):
-        """
-        Process data.
-
-        Params:
-            data - nested list, data = [sample1, sample2, ...],
-                sample = [sentence, label], sentence = [word1, word2, ...]
-            save_name - name of processed data, such as data_train.pkl
-        Returns:
-            vocab_size - vocabulary size
-            n_classes - number of classes
-        """
-        self.build_dict(data)
-        self.word2id()
-        vocab_size = self.id2word()
-        self.class2id()
-        num_classes = self.id2class()
-        self.embedding()
-        self.data_generate(data, save_name)
-
-        return vocab_size, num_classes
-
-    def build_dict(self, data):
-        """Build vocabulary."""
-
-        # just read if word2id.pkl and class2id.pkl exists
-        if self.pickle_exist("word2id.pkl") and \
-                self.pickle_exist("class2id.pkl"):
-            file_name = os.path.join(self.pickle_path, "word2id.pkl")
-            with open(file_name, 'rb') as f:
-                self.word_dict = _pickle.load(f)
-            file_name = os.path.join(self.pickle_path, "class2id.pkl")
-            with open(file_name, 'rb') as f:
-                self.label_dict = _pickle.load(f)
-            return
-
-        # build vocabulary from scratch if nothing exists
-        self.word_dict = {
-            DEFAULT_PADDING_LABEL: 0,
-            DEFAULT_UNKNOWN_LABEL: 1,
-            DEFAULT_RESERVED_LABEL[0]: 2,
-            DEFAULT_RESERVED_LABEL[1]: 3,
-            DEFAULT_RESERVED_LABEL[2]: 4}
-        self.label_dict = {}
-
-        # collect every word and label
-        for sent, label in data:
-            if len(sent) <= 1:
-                continue
-
-            if label not in self.label_dict:
-                index = len(self.label_dict)
-                self.label_dict[label] = index
-
-            for word in sent:
-                if word not in self.word_dict:
-                    index = len(self.word_dict)
-                    self.word_dict[word[0]] = index
-
-    def pickle_exist(self, pickle_name):
-        """
-        Check whether a pickle file exists.
-
-        Params
-            pickle_name: the filename of target pickle file
-        Return
-            True if file exists else False
-        """
-        if not os.path.exists(self.pickle_path):
-            os.makedirs(self.pickle_path)
-        file_name = os.path.join(self.pickle_path, pickle_name)
-        if os.path.exists(file_name):
-            return True
-        else:
-            return False
-
-    def word2id(self):
-        """Save vocabulary of {word:id} mapping format."""
-        # nothing will be done if word2id.pkl exists
-        if self.pickle_exist("word2id.pkl"):
-            return
-
-        file_name = os.path.join(self.pickle_path, "word2id.pkl")
-        with open(file_name, "wb") as f:
-            _pickle.dump(self.word_dict, f)
-
-    def id2word(self):
-        """Save vocabulary of {id:word} mapping format."""
-        # nothing will be done if id2word.pkl exists
-        if self.pickle_exist("id2word.pkl"):
-            file_name = os.path.join(self.pickle_path, "id2word.pkl")
-            with open(file_name, 'rb') as f:
-                id2word_dict = _pickle.load(f)
-            return len(id2word_dict)
-
-        id2word_dict = {self.word_dict[w]: w for w in self.word_dict}
-        file_name = os.path.join(self.pickle_path, "id2word.pkl")
-        with open(file_name, "wb") as f:
-            _pickle.dump(id2word_dict, f)
-        return len(id2word_dict)
-
-    def class2id(self):
-        """Save mapping of {class:id}."""
-        # nothing will be done if class2id.pkl exists
-        if self.pickle_exist("class2id.pkl"):
-            return
-
-        file_name = os.path.join(self.pickle_path, "class2id.pkl")
-        with open(file_name, "wb") as f:
-            _pickle.dump(self.label_dict, f)
-
-    def id2class(self):
-        """Save mapping of {id:class}."""
-        # nothing will be done if id2class.pkl exists
-        if self.pickle_exist("id2class.pkl"):
-            file_name = os.path.join(self.pickle_path, "id2class.pkl")
-            with open(file_name, "rb") as f:
-                id2class_dict = _pickle.load(f)
-            return len(id2class_dict)
-
-        id2class_dict = {self.label_dict[c]: c for c in self.label_dict}
-        file_name = os.path.join(self.pickle_path, "id2class.pkl")
-        with open(file_name, "wb") as f:
-            _pickle.dump(id2class_dict, f)
-        return len(id2class_dict)
-
-    def embedding(self):
-        """Save embedding lookup table corresponding to vocabulary."""
-        # nothing will be done if embedding.pkl exists
-        if self.pickle_exist("embedding.pkl"):
-            return
-
-        # retrieve vocabulary from pre-trained embedding (not implemented)
-
-    def data_generate(self, data_src, save_name):
-        """Convert dataset from text to digit."""
-
-        # nothing will be done if file exists
-        save_path = os.path.join(self.pickle_path, save_name)
-        if os.path.exists(save_path):
-            return
-
-        data = []
-        # for every sample
-        for sent, label in data_src:
-            if len(sent) <= 1:
-                continue
-
-            label_id = self.label_dict[label]  # label id
-            sent_id = []  # sentence ids
-            for word in sent:
-                if word in self.word_dict:
-                    sent_id.append(self.word_dict[word])
-                else:
-                    sent_id.append(self.word_dict[DEFAULT_UNKNOWN_LABEL])
-            data.append([sent_id, label_id])
-
-        # save data
-        with open(save_path, "wb") as f:
-            _pickle.dump(data, f)
-
-
-class LMPreprocess(BasePreprocess):
-    def __init__(self, data, pickle_path):
-        super(LMPreprocess, self).__init__(data, pickle_path)
+import _pickle
+import os
+
+DEFAULT_PADDING_LABEL = '<pad>'  # dict index = 0
+DEFAULT_UNKNOWN_LABEL = '<unk>'  # dict index = 1
+DEFAULT_RESERVED_LABEL = ['<reserved-2>',
+                          '<reserved-3>',
+                          '<reserved-4>']  # dict index = 2~4
+
+DEFAULT_WORD_TO_INDEX = {DEFAULT_PADDING_LABEL: 0, DEFAULT_UNKNOWN_LABEL: 1,
+                         DEFAULT_RESERVED_LABEL[0]: 2, DEFAULT_RESERVED_LABEL[1]: 3,
+                         DEFAULT_RESERVED_LABEL[2]: 4}
+
+
+# the first vocab in dict with the index = 5
+
+def save_pickle(obj, pickle_path, file_name):
+    with open(os.path.join(pickle_path, file_name), "wb") as f:
+        _pickle.dump(obj, f)
+    print("{} saved. ".format(file_name))
+
+
+def load_pickle(pickle_path, file_name):
+    with open(os.path.join(pickle_path, file_name), "rb") as f:
+        obj = _pickle.load(f)
+    return obj
+
+
+def pickle_exist(pickle_path, pickle_name):
+    """
+    :param pickle_path: the directory of target pickle file
+    :param pickle_name: the filename of target pickle file
+    :return: True if file exists else False
+    """
+    if not os.path.exists(pickle_path):
+        os.makedirs(pickle_path)
+    file_name = os.path.join(pickle_path, pickle_name)
+    if os.path.exists(file_name):
+        return True
+    else:
+        return False
+
+
+class BasePreprocess(object):
+
+    def __init__(self, data, pickle_path):
+        super(BasePreprocess, self).__init__()
+        # self.data = data
+        self.pickle_path = pickle_path
+        if not self.pickle_path.endswith('/'):
+            self.pickle_path = self.pickle_path + '/'
+
+
+class POSPreprocess(BasePreprocess):
+    """
+        This class are used to preprocess the POS Tag datasets.
+
+    """
+
+    def __init__(self, data, pickle_path="./", train_dev_split=0):
+        """
+        Preprocess pipeline, including building mapping from words to index, from index to words,
+        from labels/classes to index, from index to labels/classes.
+        :param data: three-level list
+            [
+                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
+                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
+                ...
+            ]
+        :param pickle_path: str, the directory to the pickle files. Default: "./"
+        :param train_dev_split: float in [0, 1]. The ratio of dev data split from training data. Default: 0.
+
+        """
+        super(POSPreprocess, self).__init__(data, pickle_path)
+
+        self.pickle_path = pickle_path
+
+        if pickle_exist(pickle_path, "word2id.pkl") and pickle_exist(pickle_path, "class2id.pkl"):
+            self.word2index = load_pickle(self.pickle_path, "word2id.pkl")
+            self.label2index = load_pickle(self.pickle_path, "class2id.pkl")
+        else:
+            self.word2index, self.label2index = self.build_dict(data)
+            save_pickle(self.word2index, self.pickle_path, "word2id.pkl")
+            save_pickle(self.label2index, self.pickle_path, "class2id.pkl")
+
+        if not pickle_exist(pickle_path, "id2word.pkl"):
+            index2word = self.build_reverse_dict(self.word2index)
+            save_pickle(index2word, self.pickle_path, "id2word.pkl")
+
+        if not pickle_exist(pickle_path, "id2class.pkl"):
+            index2label = self.build_reverse_dict(self.label2index)
+            save_pickle(index2label, self.pickle_path, "id2class.pkl")
+
+        if not pickle_exist(pickle_path, "data_train.pkl"):
+            data_train = self.to_index(data)
+            if train_dev_split > 0 and not pickle_exist(pickle_path, "data_dev.pkl"):
+                data_dev = data_train[: int(len(data_train) * train_dev_split)]
+                save_pickle(data_dev, self.pickle_path, "data_dev.pkl")
+            save_pickle(data_train, self.pickle_path, "data_train.pkl")
+
+    def build_dict(self, data):
+        """
+        Add new words with indices into self.word_dict, new labels with indices into self.label_dict.
+        :param data: three-level list
+            [
+                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
+                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
+                ...
+            ]
+        :return word2index: dict of {str, int}
+                label2index: dict of {str, int}
+        """
+        label2index = {}
+        word2index = DEFAULT_WORD_TO_INDEX
+        for example in data:
+            for word, label in zip(example[0], example[1]):
+                if word not in word2index:
+                    word2index[word] = len(word2index)
+                if label not in label2index:
+                    label2index[label] = len(label2index)
+        return word2index, label2index
+
+    def build_reverse_dict(self, word_dict):
+        id2word = {word_dict[w]: w for w in word_dict}
+        return id2word
+
+    def to_index(self, data):
+        """
+        Convert word strings and label strings into indices.
+        :param data: three-level list
+            [
+                [ [word_11, word_12, ...], [label_1, label_1, ...] ],
+                [ [word_21, word_22, ...], [label_2, label_1, ...] ],
+                ...
+            ]
+        :return data_index: the shape of data, but each string is replaced by its corresponding index
+        """
+        data_index = []
+        for example in data:
+            word_list = []
+            label_list = []
+            for word, label in zip(example[0], example[1]):
+                word_list.append(self.word2index[word])
+                label_list.append(self.label2index[label])
+            data_index.append([word_list, label_list])
+        return data_index
+
+    @property
+    def vocab_size(self):
+        return len(self.word2index)
+
+    @property
+    def num_classes(self):
+        return len(self.label2index)
+
+
+class ClassPreprocess(BasePreprocess):
+    """
+    Pre-process the classification datasets.
+
+    Params:
+        pickle_path - directory to save result of pre-processing
+    Saves:
+        word2id.pkl
+        id2word.pkl
+        class2id.pkl
+        id2class.pkl
+        embedding.pkl
+        data_train.pkl
+        data_dev.pkl
+        data_test.pkl
+    """
+
+    def __init__(self, pickle_path):
+        # super(ClassPreprocess, self).__init__(data, pickle_path)
+        self.word_dict = None
+        self.label_dict = None
+        self.pickle_path = pickle_path  # save directory
+
+    def process(self, data, save_name):
+        """
+        Process data.
+
+        Params:
+            data - nested list, data = [sample1, sample2, ...],
+                sample = [sentence, label], sentence = [word1, word2, ...]
+            save_name - name of processed data, such as data_train.pkl
+        Returns:
+            vocab_size - vocabulary size
+            n_classes - number of classes
+        """
+        self.build_dict(data)
+        self.word2id()
+        vocab_size = self.id2word()
+        self.class2id()
+        num_classes = self.id2class()
+        self.embedding()
+        self.data_generate(data, save_name)
+
+        return vocab_size, num_classes
+
+    def build_dict(self, data):
+        """Build vocabulary."""
+
+        # just read if word2id.pkl and class2id.pkl exists
+        if self.pickle_exist("word2id.pkl") and \
+                self.pickle_exist("class2id.pkl"):
+            file_name = os.path.join(self.pickle_path, "word2id.pkl")
+            with open(file_name, 'rb') as f:
+                self.word_dict = _pickle.load(f)
+            file_name = os.path.join(self.pickle_path, "class2id.pkl")
+            with open(file_name, 'rb') as f:
+                self.label_dict = _pickle.load(f)
+            return
+
+        # build vocabulary from scratch if nothing exists
+        self.word_dict = {
+            DEFAULT_PADDING_LABEL: 0,
+            DEFAULT_UNKNOWN_LABEL: 1,
+            DEFAULT_RESERVED_LABEL[0]: 2,
+            DEFAULT_RESERVED_LABEL[1]: 3,
+            DEFAULT_RESERVED_LABEL[2]: 4}
+        self.label_dict = {}
+
+        # collect every word and label
+        for sent, label in data:
+            if len(sent) <= 1:
+                continue
+
+            if label not in self.label_dict:
+                index = len(self.label_dict)
+                self.label_dict[label] = index
+
+            for word in sent:
+                if word not in self.word_dict:
+                    index = len(self.word_dict)
+                    self.word_dict[word[0]] = index
+
+    def pickle_exist(self, pickle_name):
+        """
+        Check whether a pickle file exists.
+
+        Params
+            pickle_name: the filename of target pickle file
+        Return
+            True if file exists else False
+        """
+        if not os.path.exists(self.pickle_path):
+            os.makedirs(self.pickle_path)
+        file_name = os.path.join(self.pickle_path, pickle_name)
+        if os.path.exists(file_name):
+            return True
+        else:
+            return False
+
+    def word2id(self):
+        """Save vocabulary of {word:id} mapping format."""
+        # nothing will be done if word2id.pkl exists
+        if self.pickle_exist("word2id.pkl"):
+            return
+
+        file_name = os.path.join(self.pickle_path, "word2id.pkl")
+        with open(file_name, "wb") as f:
+            _pickle.dump(self.word_dict, f)
+
+    def id2word(self):
+        """Save vocabulary of {id:word} mapping format."""
+        # nothing will be done if id2word.pkl exists
+        if self.pickle_exist("id2word.pkl"):
+            file_name = os.path.join(self.pickle_path, "id2word.pkl")
+            with open(file_name, 'rb') as f:
+                id2word_dict = _pickle.load(f)
+            return len(id2word_dict)
+
+        id2word_dict = {self.word_dict[w]: w for w in self.word_dict}
+        file_name = os.path.join(self.pickle_path, "id2word.pkl")
+        with open(file_name, "wb") as f:
+            _pickle.dump(id2word_dict, f)
+        return len(id2word_dict)
+
+    def class2id(self):
+        """Save mapping of {class:id}."""
+        # nothing will be done if class2id.pkl exists
+        if self.pickle_exist("class2id.pkl"):
+            return
+
+        file_name = os.path.join(self.pickle_path, "class2id.pkl")
+        with open(file_name, "wb") as f:
+            _pickle.dump(self.label_dict, f)
+
+    def id2class(self):
+        """Save mapping of {id:class}."""
+        # nothing will be done if id2class.pkl exists
+        if self.pickle_exist("id2class.pkl"):
+            file_name = os.path.join(self.pickle_path, "id2class.pkl")
+            with open(file_name, "rb") as f:
+                id2class_dict = _pickle.load(f)
+            return len(id2class_dict)
+
+        id2class_dict = {self.label_dict[c]: c for c in self.label_dict}
+        file_name = os.path.join(self.pickle_path, "id2class.pkl")
+        with open(file_name, "wb") as f:
+            _pickle.dump(id2class_dict, f)
+        return len(id2class_dict)
+
+    def embedding(self):
+        """Save embedding lookup table corresponding to vocabulary."""
+        # nothing will be done if embedding.pkl exists
+        if self.pickle_exist("embedding.pkl"):
+            return
+
+        # retrieve vocabulary from pre-trained embedding (not implemented)
+
+    def data_generate(self, data_src, save_name):
+        """Convert dataset from text to digit."""
+
+        # nothing will be done if file exists
+        save_path = os.path.join(self.pickle_path, save_name)
+        if os.path.exists(save_path):
+            return
+
+        data = []
+        # for every sample
+        for sent, label in data_src:
+            if len(sent) <= 1:
+                continue
+
+            label_id = self.label_dict[label]  # label id
+            sent_id = []  # sentence ids
+            for word in sent:
+                if word in self.word_dict:
+                    sent_id.append(self.word_dict[word])
+                else:
+                    sent_id.append(self.word_dict[DEFAULT_UNKNOWN_LABEL])
+            data.append([sent_id, label_id])
+
+        # save data
+        with open(save_path, "wb") as f:
+            _pickle.dump(data, f)
+
+
+class LMPreprocess(BasePreprocess):
+    def __init__(self, data, pickle_path):
+        super(LMPreprocess, self).__init__(data, pickle_path)
+
+
+def infer_preprocess(pickle_path, data):
+    """
+        Preprocess over inference data.
+        Transform three-level list of strings into that of index.
+        [
+            [word_11, word_12, ...],
+            [word_21, word_22, ...],
+            ...
+        ]
+    """
+    word2index = load_pickle(pickle_path, "word2id.pkl")
+    data_index = []
+    for example in data:
+        data_index.append([word2index.get(w, DEFAULT_UNKNOWN_LABEL) for w in example])
+    return data_index

fastNLP/models/sequence_modeling.py

@@ -9,17 +9,12 @@ class SeqLabeling(BaseModel):
     PyTorch Network for sequence labeling
     """
 
-    def __init__(self, hidden_dim,
-                 rnn_num_layer,
-                 num_classes,
-                 vocab_size,
-                 word_emb_dim=100,
-                 init_emb=None,
-                 rnn_mode="gru",
-                 bi_direction=False,
-                 dropout=0.5,
-                 use_crf=True):
+    def __init__(self, args):
         super(SeqLabeling, self).__init__()
+        vocab_size = args["vocab_size"]
+        word_emb_dim = args["word_emb_dim"]
+        hidden_dim = args["rnn_hidden_units"]
+        num_classes = args["num_classes"]
 
         self.Embedding = encoder.embedding.Embedding(vocab_size, word_emb_dim)
         self.Rnn = encoder.lstm.Lstm(word_emb_dim, hidden_dim)
@@ -29,7 +24,7 @@ class SeqLabeling(BaseModel):
     def forward(self, x):
         """
         :param x: LongTensor, [batch_size, mex_len]
-        :return y: [batch_size, tag_size, tag_size]
+        :return y: [batch_size, mex_len, tag_size]
         """
         x = self.Embedding(x)
         # [batch_size, max_len, word_emb_dim]
@@ -64,7 +59,7 @@ class SeqLabeling(BaseModel):
 
     def prediction(self, x, seq_length):
         """
-        :param x: FloatTensor, [batch_size, tag_size, tag_size]
+        :param x: FloatTensor, [batch_size, max_len, tag_size]
         :param seq_length: int
         :return prediction: list of tuple of (decode path(list), best score)
         """

fastNLP/modules/encoder/lstm.py

@@ -13,7 +13,7 @@ class Lstm(nn.Module):
     bidirectional : If True, becomes a bidirectional RNN. Default: False.
     """
 
-    def __init__(self, input_size, hidden_size=100, num_layers=1, dropout=0.5, bidirectional=False):
+    def __init__(self, input_size, hidden_size=100, num_layers=1, dropout=0, bidirectional=False):
         super(Lstm, self).__init__()
         self.lstm = nn.LSTM(input_size, hidden_size, num_layers, bias=True, batch_first=True,
                             dropout=dropout, bidirectional=bidirectional)

test/data_for_tests/config

@@ -74,3 +74,9 @@ save_dev_input = false
 save_loss = true
 batch_size = 1
 pickle_path = "./data_for_tests/"
+rnn_hidden_units = 100
+rnn_layers = 1
+rnn_bi_direction = true
+word_emb_dim = 100
+dropout = 0.5
+use_crf = true

test/data_for_tests/people_infer.txt

@@ -0,0 +1,2 @@
+迈向充满希望的新世纪——一九九八年新年讲话
+（附图片1张）

test/test_POS_pipeline.py

@@ -4,8 +4,8 @@ sys.path.append("..")
 
 from fastNLP.loader.config_loader import ConfigLoader, ConfigSection
 from fastNLP.action.trainer import POSTrainer
-from fastNLP.loader.dataset_loader import POSDatasetLoader
-from fastNLP.loader.preprocess import POSPreprocess
+from fastNLP.loader.dataset_loader import POSDatasetLoader, BaseLoader
+from fastNLP.loader.preprocess import POSPreprocess, load_pickle
 from fastNLP.saver.model_saver import ModelSaver
 from fastNLP.loader.model_loader import ModelLoader
 from fastNLP.action.tester import POSTester
@@ -15,32 +15,49 @@ from fastNLP.action.inference import Inference
 data_name = "people.txt"
 data_path = "data_for_tests/people.txt"
 pickle_path = "data_for_tests"
+data_infer_path = "data_for_tests/people_infer.txt"
 
 
-def test_infer():
+def infer():
+    # Load infer configuration, the same as test
+    test_args = ConfigSection()
+    ConfigLoader("config.cfg", "").load_config("./data_for_tests/config", {"POS_test": test_args})
+
+    # fetch dictinary size and number of labels from pickle files
+    word2index = load_pickle(pickle_path, "word2id.pkl")
+    test_args["vocab_size"] = len(word2index)
+    index2label = load_pickle(pickle_path, "id2class.pkl")
+    test_args["num_classes"] = len(index2label)
+
     # Define the same model
-    model = SeqLabeling(hidden_dim=train_args["rnn_hidden_units"], rnn_num_layer=train_args["rnn_layers"],
-                        num_classes=train_args["num_classes"], vocab_size=train_args["vocab_size"],
-                        word_emb_dim=train_args["word_emb_dim"], bi_direction=train_args["rnn_bi_direction"],
-                        rnn_mode="gru", dropout=train_args["dropout"], use_crf=train_args["use_crf"])
+    model = SeqLabeling(test_args)
 
     # Dump trained parameters into the model
-    ModelLoader("arbitrary_name", "./saved_model.pkl").load_pytorch(model)
+    ModelLoader.load_pytorch(model, "./saved_model.pkl")
     print("model loaded!")
 
     # Data Loader
-    pos_loader = POSDatasetLoader(data_name, data_path)
-    infer_data = pos_loader.load_lines()
-
-    # Preprocessor
-    POSPreprocess(infer_data, pickle_path)
+    raw_data_loader = BaseLoader(data_name, data_infer_path)
+    infer_data = raw_data_loader.load_lines()
+    """
+        Transform strings into list of list of strings. 
+        [
+            [word_11, word_12, ...],
+            [word_21, word_22, ...],
+            ...
+        ]
+        In this case, each line in "people_infer.txt" is already a sentence. So load_lines() just splits them.
+    """
 
     # Inference interface
-    infer = Inference()
+    infer = Inference(pickle_path)
     results = infer.predict(model, infer_data)
 
+    print(results)
+    print("Inference finished!")
 
-if __name__ == "__main__":
+
+def train_test():
     # Config Loader
     train_args = ConfigSection()
     ConfigLoader("config.cfg", "").load_config("./data_for_tests/config", {"POS": train_args})
@@ -58,10 +75,7 @@ if __name__ == "__main__":
     trainer = POSTrainer(train_args)
 
     # Model
-    model = SeqLabeling(hidden_dim=train_args["rnn_hidden_units"], rnn_num_layer=train_args["rnn_layers"],
-                        num_classes=train_args["num_classes"], vocab_size=train_args["vocab_size"],
-                        word_emb_dim=train_args["word_emb_dim"], bi_direction=train_args["rnn_bi_direction"],
-                        rnn_mode="gru", dropout=train_args["dropout"], use_crf=train_args["use_crf"])
+    model = SeqLabeling(train_args)
 
     # Start training
     trainer.train(model)
@@ -75,13 +89,10 @@ if __name__ == "__main__":
     del model, trainer, pos_loader
 
     # Define the same model
-    model = SeqLabeling(hidden_dim=train_args["rnn_hidden_units"], rnn_num_layer=train_args["rnn_layers"],
-                        num_classes=train_args["num_classes"], vocab_size=train_args["vocab_size"],
-                        word_emb_dim=train_args["word_emb_dim"], bi_direction=train_args["rnn_bi_direction"],
-                        rnn_mode="gru", dropout=train_args["dropout"], use_crf=train_args["use_crf"])
+    model = SeqLabeling(train_args)
 
     # Dump trained parameters into the model
-    ModelLoader("arbitrary_name", "./saved_model.pkl").load_pytorch(model)
+    ModelLoader.load_pytorch(model, "./saved_model.pkl")
     print("model loaded!")
 
     # Load test configuration
@@ -97,3 +108,7 @@ if __name__ == "__main__":
     # print test results
     print(tester.show_matrices())
     print("model tested!")
+
+
+if __name__ == "__main__":
+    infer()