1
|
|
|
class ModelOutput(object): |
|
|
|
|
2
|
|
|
"""Abstract base class for topic models. |
3
|
|
|
|
4
|
|
|
Ensures consistent interface across models, for base result display capabilities. |
5
|
|
|
|
6
|
|
|
Attributes |
7
|
|
|
---------- |
8
|
|
|
_doc_topic_matrix : mapping of document ids to weights for topic indices |
9
|
|
|
matrix storing the relative topic weights for each document |
10
|
|
|
_topic_term_matrix : mapping of terms to each topic |
11
|
|
|
""" |
12
|
|
|
def __init__(self, vectorized_corpus=None, model_func=None, |
13
|
|
|
vocab=None, term_frequency=None, topic_term_matrix=None, |
14
|
|
|
doc_lengths=None, doc_topic_matrix=None, **kwargs): |
15
|
|
|
if vectorized_corpus and model_func: |
|
|
|
|
16
|
|
|
self._vocab = vectorized_corpus.id_term_map |
17
|
|
|
self._doc_lengths = vectorized_corpus.doc_lengths |
18
|
|
|
self._term_frequency = vectorized_corpus.term_frequency |
19
|
|
|
self._topic_term_matrix, self._doc_topic_matrix = model_func( |
20
|
|
|
vectorized_corpus, **kwargs) |
|
|
|
|
21
|
|
|
|
22
|
|
|
elif (vocab and term_frequency and topic_term_matrix and doc_lengths and |
|
|
|
|
23
|
|
|
doc_topic_matrix): |
|
|
|
|
24
|
|
|
self._vocab = vocab |
25
|
|
|
self._term_frequency = term_frequency |
26
|
|
|
self._topic_term_matrix = topic_term_matrix |
27
|
|
|
self._doc_lengths = doc_lengths |
28
|
|
|
self._doc_topic_matrix = doc_topic_matrix |
29
|
|
|
else: |
30
|
|
|
raise ValueError("Must provide either vectorized corpus and model func, " |
31
|
|
|
"or term data and doc data.") |
32
|
|
|
|
33
|
|
|
@property |
|
|
|
|
34
|
|
|
def vocab(self): |
|
|
|
|
35
|
|
|
return self._vocab |
|
|
|
|
36
|
|
|
|
37
|
|
|
@property |
|
|
|
|
38
|
|
|
def term_frequency(self): |
|
|
|
|
39
|
|
|
return self._term_frequency |
|
|
|
|
40
|
|
|
|
41
|
|
|
@property |
|
|
|
|
42
|
|
|
def topic_term_matrix(self): |
|
|
|
|
43
|
|
|
return self._topic_term_matrix |
|
|
|
|
44
|
|
|
|
45
|
|
|
@property |
|
|
|
|
46
|
|
|
def doc_lengths(self): |
|
|
|
|
47
|
|
|
return self._doc_lengths |
|
|
|
|
48
|
|
|
|
49
|
|
|
@property |
|
|
|
|
50
|
|
|
def doc_topic_matrix(self): |
|
|
|
|
51
|
|
|
return self._doc_topic_matrix |
|
|
|
|
52
|
|
|
|
53
|
|
|
|
54
|
|
|
|
The coding style of this project requires that you add a docstring to this code element. Below, you find an example for methods:
If you would like to know more about docstrings, we recommend to read PEP-257: Docstring Conventions.