#!/usr/bin/env python
# coding: utf-8

# # Visualizing a Gensim model

# To illustrate how to use [`pyLDAvis`](https://github.com/bmabey/pyLDAvis)'s gensim [helper funtions](https://pyldavis.readthedocs.org/en/latest/modules/API.html#module-pyLDAvis.gensim) we will create a model from the [20 Newsgroup corpus](http://qwone.com/~jason/20Newsgroups/). Minimal preprocessing is done and so the model is not the best. However, the goal of this notebook is to demonstrate the helper functions.
# 
# ## Downloading the data

# In[17]:


get_ipython().run_cell_magic('bash', '', '\nmkdir -p data\npushd data\nif [ -d "20news-bydate-train" ]\nthen\n  echo "The data has already been downloaded..."\nelse\n  wget http://qwone.com/%7Ejason/20Newsgroups/20news-bydate.tar.gz\n  tar xfv 20news-bydate.tar.gz\n  rm 20news-bydate.tar.gz\nfi\necho "Lets take a look at the groups..."\nls 20news-bydate-train/\npopd\n')


# ## Exploring the dataset

# Each group dir has a set of files:

# In[2]:


get_ipython().system('ls -lah data/20news-bydate-train/sci.space | tail  -n 5')


# Lets take a peak at one email:

# In[3]:


get_ipython().system('head data/20news-bydate-train/sci.space/61422')


# ## Loading and tokenizing the corpus

# In[4]:


from glob import glob
import re
import string
import funcy as fp
from gensim import models
from gensim.corpora import Dictionary, MmCorpus
import nltk
import pandas as pd

nltk.download('stopwords')


# In[5]:


# quick and dirty....
EMAIL_REGEX = re.compile(r"[a-z0-9\.\+_-]+@[a-z0-9\._-]+\.[a-z]*")
FILTER_REGEX = re.compile(r"[^a-z '#]")
TOKEN_MAPPINGS = [(EMAIL_REGEX, "#email"), (FILTER_REGEX, ' ')]

def tokenize_line(line):
    res = line.lower()
    for regexp, replacement in TOKEN_MAPPINGS:
        res = regexp.sub(replacement, res)
    return res.split()
    
def tokenize(lines, token_size_filter=2):
    tokens = fp.mapcat(tokenize_line, lines)
    return [t for t in tokens if len(t) > token_size_filter]
    

def load_doc(filename):
    group, doc_id = filename.split('/')[-2:]
    with open(filename, errors='ignore') as f:
        doc = f.readlines()
    return {'group': group,
            'doc': doc,
            'tokens': tokenize(doc),
            'id': doc_id}


docs = pd.DataFrame(list(map(load_doc, glob('data/20news-bydate-train/*/*')))).set_index(['group','id'])
docs.head()


# ## Creating the dictionary, and bag of words corpus

# In[6]:


def nltk_stopwords():
    return set(nltk.corpus.stopwords.words('english'))

def prep_corpus(docs, additional_stopwords=set(), no_below=5, no_above=0.5):
  print('Building dictionary...')
  dictionary = Dictionary(docs)
  stopwords = nltk_stopwords().union(additional_stopwords)
  stopword_ids = map(dictionary.token2id.get, stopwords)
  dictionary.filter_tokens(stopword_ids)
  dictionary.compactify()
  dictionary.filter_extremes(no_below=no_below, no_above=no_above, keep_n=None)
  dictionary.compactify()

  print('Building corpus...')
  corpus = [dictionary.doc2bow(doc) for doc in docs]

  return dictionary, corpus



# In[7]:


dictionary, corpus = prep_corpus(docs['tokens'])


# In[8]:


MmCorpus.serialize('newsgroups.mm', corpus)
dictionary.save('newsgroups.dict')


# ## Fitting the LDA model

# In[9]:


get_ipython().run_cell_magic('time', '', "lda = models.ldamodel.LdaModel(corpus=corpus, id2word=dictionary, num_topics=50, passes=10)\n                                      \nlda.save('newsgroups_50_lda.model')\n")


# ## Visualizing the model with pyLDAvis
# 
# Okay, the moment we have all been waiting for is finally here!  You'll notice in the visualization that we have a few junk topics that would probably disappear after better preprocessing of the corpus. This is left as an exercises to the reader. :)

# In[10]:


import pyLDAvis.gensim_models as gensimvis
import pyLDAvis


# In[11]:


vis_data = gensimvis.prepare(lda, corpus, dictionary)
pyLDAvis.display(vis_data)


# ## Fitting the HDP model
# 
# We can both visualize LDA models as well as gensim HDP models with pyLDAvis.
# 
# The difference between HDP and LDA is that HDP is a non-parametric method. Which means that we don't need to specify the number of topics. HDP will fit as many topics as it can and find the optimal number of topics by itself.

# In[12]:


get_ipython().run_cell_magic('time', '', "# The optional parameter T here indicates that HDP should find no more than 50 topics\n# if there exists any.\nhdp = models.hdpmodel.HdpModel(corpus, dictionary, T=50)\n                                      \nhdp.save('newsgroups_hdp.model')\n")


# ## Visualizing the HDP model with pyLDAvis
# 
# As for the LDA model, in order to prepare the visualization you only need to pass it your model, the corpus, and the associated dictionary.

# In[13]:


vis_data = gensimvis.prepare(hdp, corpus, dictionary)
pyLDAvis.display(vis_data)