from typing import List, Tuple
import pickle
import tensorflow as tf
import numpy as np
from ..stemmer_morph_analyzer import StemmerAnalyzer
from ..tokenizer import TreebankWordTokenize
from ..utils import check_and_download, load_keras_tokenizer
from ._treestack_utils import (
create_pos_tagger_model,
process_single_word_input,
)
# Resolving parent dependencies
from inspect import getsourcefile
import os
import sys
current_path = os.path.abspath(getsourcefile(lambda: 0))
current_dir = os.path.dirname(current_path)
parent_dir = current_dir[: current_dir.rfind(os.path.sep)]
sys.path.insert(0, parent_dir)
RESOURCES_PATH = os.path.join(os.path.dirname(__file__), "resources/")
PROD_WEIGHTS_LOC = RESOURCES_PATH + "PoS_TreeStack_prod.weights"
EVAL_WEIGHTS_LOC = RESOURCES_PATH + "PoS_TreeStack_eval.weights"
WORD_EMBEDDING_MATRIX_LOC = os.path.abspath(
os.path.join(
os.path.dirname(__file__),
"..",
"resources/TBWTokenized_word_embedding.matrix",
)
)
PROD_WEIGHTS_LINK = "https://vnlp-model-weights.s3.eu-west-1.amazonaws.com/PoS_TreeStack_prod.weights"
EVAL_WEIGHTS_LINK = "https://vnlp-model-weights.s3.eu-west-1.amazonaws.com/PoS_TreeStack_eval.weights"
WORD_EMBEDDING_MATRIX_LINK = "https://vnlp-model-weights.s3.eu-west-1.amazonaws.com/TBWTokenized_word_embedding.matrix"
TOKENIZER_WORD_LOC = os.path.abspath(
os.path.join(
os.path.dirname(__file__), "..", "resources/TB_word_tokenizer.json"
)
)
TOKENIZER_MORPH_TAG_LOC_LOC = os.path.abspath(
os.path.join(
os.path.dirname(__file__),
"..",
"stemmer_morph_analyzer/resources/Stemmer_morph_tag_tokenizer.json",
)
) # using the tokenizer of stemmer_morph_analyzer
TOKENIZER_POS_LABEL_LOC = RESOURCES_PATH + "PoS_label_tokenizer.json"
# Data Preprocessing Config
SENTENCE_MAX_LEN = 40
TAG_MAX_LEN = 15
WORD_OOV_TOKEN = "<OOV>"
# Loading Tokenizers
# Have to load tokenizers here because model config depends on them
tokenizer_word = load_keras_tokenizer(TOKENIZER_WORD_LOC)
# This is transferred from StemmerAnalyzer
tokenizer_morph_tag = load_keras_tokenizer(TOKENIZER_MORPH_TAG_LOC_LOC)
tokenizer_pos_label = load_keras_tokenizer(TOKENIZER_POS_LABEL_LOC)
POS_VOCAB_SIZE = len(tokenizer_pos_label.word_index)
# Model Config
WORD_EMBEDDING_VECTOR_SIZE = 128 # Word2Vec_medium.model
WORD_EMBEDDING_VOCAB_SIZE = 63_992 # Word2Vec_medium.model
# WORD_EMBEDDING_MATRIX and TAG_EMBEDDING MATRIX are initialized as Zeros, will be overwritten when model is loaded.
WORD_EMBEDDING_MATRIX = np.zeros(
(WORD_EMBEDDING_VOCAB_SIZE, WORD_EMBEDDING_VECTOR_SIZE)
)
TAG_EMBEDDING_MATRIX = np.zeros((127, 32))
NUM_RNN_STACKS = 2
RNN_UNITS_MULTIPLIER = 2
TAG_NUM_RNN_UNITS = WORD_EMBEDDING_VECTOR_SIZE
LC_NUM_RNN_UNITS = TAG_NUM_RNN_UNITS * RNN_UNITS_MULTIPLIER
RC_NUM_RNN_UNITS = TAG_NUM_RNN_UNITS * RNN_UNITS_MULTIPLIER
FC_UNITS_MULTIPLIERS = (2, 1)
WORD_FORM = "whole"
DROPOUT = 0.2
[docs]class TreeStackPoS:
"""
Tree-stack Part of Speech Tagger class.
- This Part of Speech Tagger is *inspired* by `Tree-stack LSTM in Transition Based Dependency Parsing <https://aclanthology.org/K18-2012/>`_.
- "Inspire" is emphasized because this implementation uses the approach of using Morphological Tags, Pre-trained word embeddings and POS tags as input for the model, rather than implementing the exact network proposed in the paper.
- It achieves 0.89 Accuracy and 0.71 F1_macro_score on test sets of Universal Dependencies 2.9.
- Input data is processed by NLTK.tokenize.TreebankWordTokenizer.
- For more details about the training procedure, dataset and evaluation metrics, see `ReadMe <https://github.com/vngrs-ai/VNLP/blob/main/vnlp/part_of_speech_tagger/ReadMe.md>`_.
"""
def __init__(self, evaluate, stemmer_analyzer=None):
self.model = create_pos_tagger_model(
WORD_EMBEDDING_VOCAB_SIZE,
WORD_EMBEDDING_VECTOR_SIZE,
WORD_EMBEDDING_MATRIX,
POS_VOCAB_SIZE,
SENTENCE_MAX_LEN,
TAG_MAX_LEN,
NUM_RNN_STACKS,
TAG_NUM_RNN_UNITS,
LC_NUM_RNN_UNITS,
RC_NUM_RNN_UNITS,
DROPOUT,
TAG_EMBEDDING_MATRIX,
FC_UNITS_MULTIPLIERS,
)
# Check and download word embedding matrix and model weights
check_and_download(
WORD_EMBEDDING_MATRIX_LOC, WORD_EMBEDDING_MATRIX_LINK
)
if evaluate:
MODEL_WEIGHTS_LOC = EVAL_WEIGHTS_LOC
MODEL_WEIGHTS_LINK = EVAL_WEIGHTS_LINK
else:
MODEL_WEIGHTS_LOC = PROD_WEIGHTS_LOC
MODEL_WEIGHTS_LINK = PROD_WEIGHTS_LINK
check_and_download(MODEL_WEIGHTS_LOC, MODEL_WEIGHTS_LINK)
# Load Word embedding matrix
word_embedding_matrix = np.load(WORD_EMBEDDING_MATRIX_LOC)
# Load Model weights
with open(MODEL_WEIGHTS_LOC, "rb") as fp:
model_weights = pickle.load(fp)
# Insert word embedding weights to correct position (1 for Part of Speech Tagger model)
model_weights.insert(1, word_embedding_matrix)
# Set model weights
self.model.set_weights(model_weights)
self.tokenizer_word = tokenizer_word
self.tokenizer_morph_tag = tokenizer_morph_tag
self.tokenizer_pos_label = tokenizer_pos_label
# I don't want StemmerAnalyzer to occupy any memory in GPU!
if stemmer_analyzer is None:
with tf.device("/cpu:0"):
stemmer_analyzer = StemmerAnalyzer()
self.stemmer_analyzer = stemmer_analyzer
[docs] def predict(self, sentence: str) -> List[Tuple[str, str]]:
"""
Args:
sentence:
Input text(sentence).
Returns:
List of (token, pos_label).
"""
whole_tokens_in_sentence = TreebankWordTokenize(sentence)
sentence_analysis_result = self.stemmer_analyzer.predict(sentence)
sentence_analysis_result = [
sentence_analysis.replace("^", "+")
for sentence_analysis in sentence_analysis_result
]
num_tokens_in_sentence = len(whole_tokens_in_sentence)
# This is for debugging purposes in case an inconsistency occurs during tokenization.
if not len(sentence_analysis_result) == num_tokens_in_sentence:
raise Exception(
sentence,
"Length of sentence and sentence_analysis_result don't match",
)
pos_int_labels = []
for t in range(num_tokens_in_sentence):
# t is the index of token/word
X = process_single_word_input(
t,
whole_tokens_in_sentence,
sentence_analysis_result,
SENTENCE_MAX_LEN,
TAG_MAX_LEN,
self.tokenizer_word,
self.tokenizer_morph_tag,
self.tokenizer_pos_label,
pos_int_labels,
WORD_FORM,
)
# Predicting
raw_pred = self.model(X).numpy()[0]
pos_int_label = np.argmax(raw_pred, axis=-1)
pos_int_labels.append(pos_int_label)
# Converting integer labels to text form
pos_labels = [
self.tokenizer_pos_label.sequences_to_texts([[pos_int_label]])[0]
for pos_int_label in pos_int_labels
]
result = [
(token, pos_label)
for token, pos_label in zip(whole_tokens_in_sentence, pos_labels)
]
return result