align matches for correctness and readability

[thatbudakguy]

using a matcher based on edit distance ratios results in matches that may include irrelevant material at the end; alignment is necessary to remove this material from the match.

additionally, alignment allows for padding gaps with spaces for easier comparison of the results, and can aid in colorizing the results to highlight differences (#54). it could play into the structured output (#39).

[thatbudakguy]

candidates for libraries to do this:

• nwalign3: doesn't look particularly maintained but has a C backend no support for utf-8
• minineedle: very small, pure python, maybe maintained? no custom matrices
• uta-align: small, C-backed, not passing its tests? no docs, no custom matrices?
• parasail-python: backed by popular SIMD C library, somewhat larger no support for utf-8
• abydos: huge, well-maintained kitchen-sink NLP library only scoring, no alignments
• cltk: pure python, arbitrary matrices & utf-8 supported
• gotoh2: C-backed, supported scoring matrices via csv
• lingpy: C-backed, supports custom scorer dict – using this one

[thatbudakguy]

rather than attempting to "trim" low-similarity regions from the two sequences, it might be better to switch to Smith-Waterman or another local alignment, since ultimately we're concerned with all of the high-similarity regions within the two sequences wherever they are (not truly a global alignment).

currently, using Needleman-Wunsch results in alignments like:

由也千乘之國可使治其賦也不知其仁也求也何如子曰求也千室之邑百乘之家可使為之宰也不知其仁也赤也何如子曰赤也束帶立於朝可使與賓客言　　　　　　　　　　　　　　　　　　　　　　也　
由也千乘之國可使治其賦　　　　　　　也　　　　求也千室之邑百乘之家可使為之宰　　　　　　　　　　　　赤也束帶立於朝可使與賓客言也又曰子謂子產有君子之道四焉其行己也恭其事上也敬

where both outlying 也 are undesirable, and we certainly want to avoid the second one since it's far outside the main high-scoring region. the optimal alignment is:

由也千乘之國可使治其賦也不知其仁也求也何如子曰求也千室之邑百乘之家可使為之宰也不知其仁也赤也何如子曰赤也束帶立於朝可使與賓客言也　
由也千乘之國可使治其賦也　　　　　　　　　　　求也千室之邑百乘之家可使為之宰　　　　　　　　　　　　赤也束帶立於朝可使與賓客言也