refactor stencil.merger namespace

[erdos]

Rework and simplify merger.clj file. This namespace is responsible for extracting stencil expressions from the template files while maintaining a correct XML tree.

The file is the oldest part of the library (from 2018) and even though we barely had to touch it, it still shows signs of immaturity: complex code, bad naming, hard to follow test cases, missing docs, not very good performance.

The plan is to rewrite it to utilize transducers more with a trampoline-style executor around it. The principles of the rewrite are going to be the following:

• Simplify code by breaking it up to smaller methods and create an easy-to-follow pattern to be used around the whole file (with the trampoline style)
• Write comments for cases that are not obvious
• Get to almost 100% test coverage on merging algorithm
• Do some benchmarking to make sure there is no visible performance degradation.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Please upload report for BASE (master@fc2cfbd). Learn more about missing BASE report.

Additional details and impacted files

@@            Coverage Diff            @@
##             master     #152   +/-   ##
=========================================
  Coverage          ?   93.78%           
=========================================
  Files             ?       30           
  Lines             ?     2077           
  Branches          ?       92           
=========================================
  Hits              ?     1948           
  Misses            ?       37           
  Partials          ?       92           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[erdos]

Performance testing with document from issue #81

(comment
  ;; clj -A:build
  (require '[clojure.repl.deps :refer [add-lib]])
  (add-lib 'criterium/criterium {:mvn/version "0.4.6"}) 
  (use 'criterium.core)
  (require 'stencil.merger)

  (->> "/home/erdos/Downloads/word/document.xml"
       (clojure.java.io/file)
       (clojure.java.io/reader)
       (stencil.merger/parse-to-tokens-seq)
       (dorun)
       (bench))
  )

Output on current branch:

Evaluation count : 240 in 60 samples of 4 calls.
             Execution time mean : 305.535731 ms
    Execution time std-deviation : 31.925299 ms
   Execution time lower quantile : 293.731132 ms ( 2.5%)
   Execution time upper quantile : 348.607057 ms (97.5%)
                   Overhead used : 5.464008 ns

Found 5 outliers in 60 samples (8.3333 %)
	low-severe	1 (1.6667 %)
	low-mild	4 (6.6667 %)
 Variance from outliers : 72.0506 % Variance is severely inflated by outliers

Baseline on master branch:

Evaluation count : 60 in 60 samples of 1 calls.
             Execution time mean : 1.526966 sec
    Execution time std-deviation : 50.359723 ms
   Execution time lower quantile : 1.480701 sec ( 2.5%)
   Execution time upper quantile : 1.655090 sec (97.5%)
                   Overhead used : 5.454800 ns

Found 4 outliers in 60 samples (6.6667 %)
	low-severe	2 (3.3333 %)
	low-mild	2 (3.3333 %)
 Variance from outliers : 19.0423 % Variance is moderately inflated by outliers

To summarize, there is a significant (5X) speedup. Std-deviance increased, possibly due to the GC kicking in because of the higher memory usage by the larger number of tests in a unit of time.