"""
utils
-----
Utility functions for data loading and cleaning.
Contents:
_check_str_similarity,
_check_str_args,
graph_lda_topic_evals,
english_names_list
"""
from difflib import SequenceMatcher
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from gensim import corpora
from gensim.models import CoherenceModel
from gensim.models.ldamulticore import LdaMulticore
from tqdm.auto import tqdm
[docs]def _check_str_similarity(str_1, str_2):
"""Checks the similarity of two strings."""
return SequenceMatcher(None, str_1, str_2).ratio()
[docs]def _check_str_args(arguments, valid_args):
"""
Checks whether a str argument is valid, and makes suggestions if not.
"""
if isinstance(arguments, str):
if arguments in valid_args:
return arguments
suggestions = []
for v in valid_args:
similarity_score = round(_check_str_similarity(str_1=arguments, str_2=v), 2)
arg_and_score = (v, similarity_score)
suggestions.append(arg_and_score)
ordered_suggestions = sorted(suggestions, key=lambda x: x[1], reverse=True)
print(f"'{arguments}' is not a valid argument for the given function.")
print(f"The closest valid options to '{arguments}' are:")
for item in ordered_suggestions[:5]:
print(item)
return
elif isinstance(arguments, list):
# Check arguments, and remove them if they're invalid.
for a in arguments:
_check_str_args(arguments=a, valid_args=valid_args)
return arguments
[docs]def graph_lda_topic_evals(
corpus=None,
num_topic_words=10,
topic_nums_to_compare=None,
metrics=True,
verbose=True,
**kwargs,
):
"""
Graphs metrics for the given models over the given number of topics.
Parameters
----------
corpus : list of lists (default=None)
The text corpus over which analysis should be done.
num_topic_words : int (default=10)
The number of keywords that should be extracted.
topic_nums_to_compare : list (default=None)
The number of topics to compare metrics over.
Note: None selects all numbers from 1 to num_topic_words.
metrics : str or bool (default=True: all metrics)
The metrics to include.
Options:
stability: model stability based on Jaccard similarity.
coherence: how much the words associated with model topics co-occur.
verbose : bool (default=True)
Whether to show a tqdm progress bar for the query.
**kwargs : keyword arguments
Arguments correspoding to gensim.models.ldamulticore.LdaMulticore.
Returns
-------
ax : matplotlib axis
A graph of the given metrics for each of the given models based on each topic number.
"""
assert (
metrics == "stability" or metrics == "coherence" or metrics == True
), "An invalid value has been passed to the 'metrics' argument - please choose from 'stability', 'coherence', or True for both."
if metrics == True:
metrics = ["stability", "coherence"]
if not isinstance(corpus[0], list):
corpus = [c.split() for c in corpus]
def jaccard_similarity(topic_1, topic_2):
"""
Derives the Jaccard similarity of two topics.
Notes
-----
Jaccard similarity:
- A statistic used for comparing the similarity and diversity of sample sets.
- J(A,B) = (A ∩ B)/(A ∪ B).
- Goal is low Jaccard scores for coverage of the diverse elements.
"""
# Fix for cases where there are not enough texts for clustering models.
if topic_1 == [] and topic_2 != []:
topic_1 = topic_2
if topic_1 != [] and topic_2 == []:
topic_2 = topic_1
if topic_1 == [] and topic_2 == []:
topic_1, topic_2 = ["_None"], ["_None"]
intersection = set(topic_1).intersection(set(topic_2))
num_intersect = float(len(intersection))
union = set(topic_1).union(set(topic_2))
num_union = float(len(union))
return num_intersect / num_union
plt.figure()
dirichlet_dict = corpora.Dictionary(corpus)
bow_corpus = [dirichlet_dict.doc2bow(text) for text in corpus]
# Add an extra topic so that metrics can be calculated all inputs.
if topic_nums_to_compare == None:
topic_nums_to_compare = list(range(num_topic_words + 1)[1:])
else:
topic_nums_to_compare.append(topic_nums_to_compare[-1] + 1)
LDA_models = {}
LDA_topics = {}
disable = not verbose
for i in tqdm(
iterable=topic_nums_to_compare, desc="LDA models ran", disable=disable
):
LDA_models[i] = LdaMulticore(
corpus=bow_corpus, id2word=dirichlet_dict, num_topics=i, **kwargs
)
shown_topics = LDA_models[i].show_topics(
num_topics=i, num_words=num_topic_words, formatted=False
)
LDA_topics[i] = [[word[0] for word in topic[1]] for topic in shown_topics]
LDA_stability = {}
for i in range(0, len(topic_nums_to_compare) - 1):
jaccard_sims = []
for t1, topic1 in enumerate( # pylint: disable=unused-variable
LDA_topics[topic_nums_to_compare[i]]
):
sims = []
for t2, topic2 in enumerate( # pylint: disable=unused-variable
LDA_topics[topic_nums_to_compare[i + 1]]
):
sims.append(jaccard_similarity(topic1, topic2))
jaccard_sims.append(sims)
LDA_stability[topic_nums_to_compare[i]] = jaccard_sims
mean_stabilities = [
np.array(LDA_stability[i]).mean() for i in topic_nums_to_compare[:-1]
]
coherences = [
CoherenceModel(
model=LDA_models[i],
texts=corpus,
dictionary=dirichlet_dict,
coherence="c_v",
).get_coherence()
for i in topic_nums_to_compare[:-1]
]
# Limit topic numbers to the number of keywords.
coh_sta_diffs = [
coherences[i] - mean_stabilities[i]
for i in range(len(topic_nums_to_compare))[:-1]
]
coh_sta_max = max(coh_sta_diffs)
coh_sta_max_idxs = [i for i, j in enumerate(coh_sta_diffs) if j == coh_sta_max]
# Choose less topics in case there's more than one max.
ideal_topic_num_index = coh_sta_max_idxs[0]
ideal_topic_num = topic_nums_to_compare[ideal_topic_num_index]
ax = sns.lineplot(
x=topic_nums_to_compare[:-1], y=mean_stabilities, label="Average Topic Overlap"
)
ax = sns.lineplot(
x=topic_nums_to_compare[:-1], y=coherences, label="Topic Coherence"
)
ax.axvline(x=ideal_topic_num, label="Ideal Number of Topics", color="black")
ax.axvspan(
xmin=ideal_topic_num - 1, xmax=ideal_topic_num + 1, alpha=0.5, facecolor="grey"
)
# Set plot limits.
y_max = max(max(mean_stabilities), max(coherences)) + (
0.10 * max(max(mean_stabilities), max(coherences))
)
ax.set_ylim([0, y_max])
ax.set_xlim([topic_nums_to_compare[0], topic_nums_to_compare[-1] - 1])
ax.axes.set_title("Method Metrics per Number of Topics", fontsize=25)
ax.set_ylabel("Metric Level", fontsize=20)
ax.set_xlabel("Number of Topics", fontsize=20)
plt.legend(fontsize=20)
return ax