from revscoring.features import wikitext
from revscoring.features.modifiers import max, sub
from revscoring.languages import english
from revscoring import Feature
from revscoring.datasources import Datasource, revision_oriented
from revscoring.dependencies import solve
from gensim.models.keyedvectors import KeyedVectors
import numpy as np
import pdb

word2vec = None
VECTORS_DIM = 300

def load_word2vec(filepath):
    global word2vec
    if word2vec is not None:
        return word2vec
    #word2vec = KeyedVectors.load_word2vec_format(filepath,
    #                                               binary=True, limit=20000)
    print("called here")
    word2vec = np.random.uniform(0, 1, size=300)
    return word2vec


def get_word_vectors(non_stop_tokens):
    word2vec = load_word2vec('')
    vector = np.zeros((1, VECTORS_DIM))
    words_added = 0
    for tok in non_stop_tokens:
        try:
            vec = word2vec[tok]
            vector += vec
            words_added += 1
        except:
            continue
    return vector/words_added

word_vectors = Datasource("word_vectors",
                         get_word_vectors,
                         depends_on=[english.stopwords.revision.datasources.non_stopwords])

def get_feature_at(idx):
    return Feature(
        "word_vector_{}".format(idx),
        lambda vec: vec[idx],
        depends_on=[word_vectors],
        returns=float
    )

features = []
for i in range(0,300):
    features.append(get_feature_at(i))

vectors = solve(features, cache={revision_oriented.revision.text: 'Hey ther, how are you?'})