major bugs fixed in data initialization, normalization, and random nu…

…mber generation

README.md

@@ -53,19 +53,21 @@ pip3 install bitermplus
 ```python
 import bitermplus as btm
 import numpy as np
-from gzip import open as gzip_open
+import pandas as pd
 
-# Importing and vectorizing text data
-with gzip_open('dataset/SearchSnippets.txt.gz', 'rb') as file:
-    texts = file.readlines()
+# Importing data
+df = pd.read_csv(
+    'dataset/SearchSnippets.txt.gz', header=None, names=['texts'])
+texts = df['texts'].str.strip().tolist()
 
 # Vectorizing documents, obtaining full vocabulary and biterms
-X, vocab = btm.get_words_freqs(texts)
-docs_vec = btm.get_vectorized_docs(X)
-biterms = btm.get_biterms(X)
+X, vocabulary, vocab_dict = btm.get_words_freqs(texts)
+docs_vec = btm.get_vectorized_docs(texts, vocabulary)
+biterms = btm.get_biterms(docs_vec)
 
 # Initializing and running model
-model = btm.BTM(X, vocab, T=8, W=vocab.size, M=20, alpha=50/8, beta=0.01)
+model = btm.BTM(
+    X, vocabulary, seed=12321, T=8, W=vocabulary.size, M=20, alpha=50/8, beta=0.01)
 model.fit(biterms, iterations=20)
 p_zd = model.transform(docs_vec)
 

docs/source/tutorial.rst

@@ -10,19 +10,21 @@ Here is a simple example of package usage:
 
     import bitermplus as btm
     import numpy as np
-    from gzip import open as gzip_open
+    import pandas as pd
 
-    # Importing and vectorizing text data
-    with gzip_open('dataset/SearchSnippets.txt.gz', 'rb') as file:
-        texts = file.readlines()
+    # Importing data
+    df = pd.read_csv(
+        'dataset/SearchSnippets.txt.gz', header=None, names=['texts'])
+    texts = df['texts'].str.strip().tolist()
 
     # Vectorizing documents, obtaining full vocabulary and biterms
-    X, vocab = btm.get_words_freqs(texts)
-    docs_vec = btm.get_vectorized_docs(X)
-    biterms = btm.get_biterms(X)
+    X, vocabulary, vocab_dict = btm.get_words_freqs(texts)
+    docs_vec = btm.get_vectorized_docs(texts, vocabulary)
+    biterms = btm.get_biterms(docs_vec)
 
     # Initializing and running model
-    model = btm.BTM(X, vocab, T=8, W=vocab.size, M=20, alpha=50/8, beta=0.01)
+    model = btm.BTM(
+        X, vocabulary, seed=12321, T=8, W=vocabulary.size, M=20, alpha=50/8, beta=0.01)
     model.fit(biterms, iterations=20)
     p_zd = model.transform(docs_vec)
 

src/bitermplus/__init__.py

@@ -1,4 +1,4 @@
-__version__ = '0.5.10'
+__version__ = '0.6.0'
 
 from bitermplus._btm import BTM
 from bitermplus._util import *