nlp-lab/Project/naive_approach/naive_approach.py
2018-06-26 15:00:24 +02:00

129 lines
3.7 KiB
Python

# coding: utf-8
# In[1]:
import pandas as pd
from IPython.display import clear_output, Markdown, Math
import ipywidgets as widgets
import os
import unicodedata as uni
import numpy as np
from nltk.stem import PorterStemmer
from nltk.tokenize import sent_tokenize, word_tokenize
from nltk.corpus import wordnet
import math
import pprint
# # Naive Approach
table = pd.read_csv('../Tools/emoji_descriptions.csv')
#######################
# Helper functions
#######################
def stemming(messages):
stemmed_messages = []
ps = PorterStemmer()
for m in messages:
words = word_tokenize(m)
sm = []
for w in words:
sm.append(ps.stem(w))
m = (" ").join(sm)
stemmed_messages.append(m)
return stemmed_messages
# * compare words to emoji descriptions
def evaluate_sentence(sentence, table, description_key = 'description', lang = 'eng'):
tokenized_sentence = word_tokenize(sentence)
n = len(tokenized_sentence)
l = table.shape[0]
matrix_list = []
for index, row in table.iterrows():
emoji_tokens = word_tokenize(row[description_key])
m = len(emoji_tokens)
mat = np.zeros(shape=(m,n))
for i in range(len(emoji_tokens)):
for j in range(len(tokenized_sentence)):
syn1 = wordnet.synsets(emoji_tokens[i],lang=lang)
if len(syn1) == 0:
continue
w1 = syn1[0]
#print(j, tokenized_sentence)
syn2 = wordnet.synsets(tokenized_sentence[j], lang=lang)
if len(syn2) == 0:
continue
w2 = syn2[0]
val = w1.wup_similarity(w2)
if val is None:
continue
mat[i,j] = val
#print(row['character'], mat)
matrix_list.append(mat)
return matrix_list
###########################
#Functions to be called from main script
###########################
# load and preprocess data
# emojis_to_consider can be either a list or "all"
def prepareData(stemming=False, emojis_to_consider="all"):
table.head()
if(stemming):
table['description'] = stemming(table['description'])
#collect the emojis
lookup = {}
emoji_set = []
for index, row in table.iterrows():
if(emojis_to_consider=="all" or (type(emojis_to_consider)==list and row['character'] in emojis_to_consider)):
lookup[index] = row['character']
emoji_set.append(row['character'])
emoji_set = set(emoji_set)
return lookup
# make a prediction for an input sentence
def predict(sentence, lookup, emojis_to_consider="all", criteria="threshold", description_key='description', lang = 'eng', n=10, t=0.9):
result = evaluate_sentence(sentence, table, description_key, lang)
if(criteria=="summed"):
indexes = np.argsort([-np.sum(x) for x in result])[0:n]
elif (criteria=="max_val"):
indexes = np.argsort([-np.max(x) for x in result])[0:n]
elif(criteria=="avg"):
indexes = np.argsort([-np.mean(x) for x in result])[0:n]
else:
indexes= np.argsort([-len(np.where(x>t)[0]) / (x.shape[0] * x.shape[1]) for x in result])[0:n]
if(emojis_to_consider!="all"):
for i in indexes:
if (i not in lookup):
indexes = np.delete(indexes, [i])
# build a result table
table_array = [[lookup[indexes[i]], str(table.iloc[indexes[i]][description_key])] for i in range(n) ]
table_frame = pd.DataFrame(table_array, columns=[criteria, 'description'])
#display(table_frame)
return list(table_frame[criteria])
#predict("I like to travel by train", description_key='description' , lang='eng')