Speedup of the TextRank algorithm

[miso-belica]

This is the speedup inspired by #139 but I not doing so aggressive optimizations because I want to prevent the tests to fail.

[codecov]

Codecov Report

Merging #140 into master will increase coverage by 0.59%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #140      +/-   ##
==========================================
+ Coverage   80.68%   81.28%   +0.59%     
==========================================
  Files          37       37              
  Lines        1719     1774      +55     
==========================================
+ Hits         1387     1442      +55     
  Misses        332      332

Impacted Files	Coverage Δ
sumy/summarizers/text_rank.py	95.65% <100%> (+0.26%)	⬆️
sumy/summarizers/sum_basic.py	100% <0%> (ø)	⬆️
sumy/summarizers/luhn.py	98.18% <0%> (+0.06%)	⬆️
sumy/nlp/tokenizers.py	89.87% <0%> (+0.12%)	⬆️
sumy/summarizers/reduction.py	97.36% <0%> (+0.14%)	⬆️
sumy/models/tf.py	97.29% <0%> (+0.15%)	⬆️
sumy/utils.py	84.61% <0%> (+0.24%)	⬆️
sumy/summarizers/lsa.py	90.54% <0%> (+0.26%)	⬆️
sumy/models/dom/_sentence.py	96.42% <0%> (+0.27%)	⬆️
... and 7 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2613de4...591cd5e. Read the comment docs.