P281424/AMR
1
0
Fork 0
mirror of https://github.com/msberends/AMR.git synced 2025-07-08 22:41:52 +02:00
Code Releases Activity

(v1.3.0.9015) as.mo() speedup for valid taxonomic names

Browse Source
This commit is contained in:
dr. M.S. (Matthijs) Berends
2020-09-03 20:59:21 +02:00
parent c4b87fe241
commit 68e9cb78e9
104 changed files with 542 additions and 529 deletions
Download Patch File Download Diff File Expand all files Collapse all files

116
docs/articles/benchmarks.html
View File

@ -39,7 +39,7 @@
</button>
<span class="navbar-brand">
<a class="navbar-link" href="../index.html">AMR (for R)</a>
<span class="version label label-default" data-toggle="tooltip" data-placement="bottom" title="Latest development version">1.3.0.9006</span>
<span class="version label label-default" data-toggle="tooltip" data-placement="bottom" title="Latest development version">1.3.0.9015</span>
</span>
</div>
@ -79,7 +79,7 @@
<a href="../articles/datasets.html">
<span class="fa fa-database"></span>
Download our reference data sets for own use
Data sets for download / own use
</a>
</li>
<li>
@ -225,28 +225,40 @@
<span class="fu"><a href="../reference/as.mo.html">as.mo</a></span>(<span class="st">"MRSA"</span>), <span class="co"># Methicillin Resistant S. aureus</span>
<span class="fu"><a href="../reference/as.mo.html">as.mo</a></span>(<span class="st">"VISA"</span>), <span class="co"># Vancomycin Intermediate S. aureus</span>
<span class="fu"><a href="../reference/as.mo.html">as.mo</a></span>(<span class="st">"VRSA"</span>), <span class="co"># Vancomycin Resistant S. aureus</span>
<span class="fu"><a href="../reference/as.mo.html">as.mo</a></span>(<span class="fl">22242419</span>), <span class="co"># Catalogue of Life ID</span>
times = <span class="fl">10</span>)
<span class="fu"><a href="https://rdrr.io/r/base/print.html">print</a></span>(<span class="kw">S.aureus</span>, unit = <span class="st">"ms"</span>, signif = <span class="fl">2</span>)
<span class="co"># Unit: milliseconds</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># as.mo("sau") 8.7 9.3 13 9.8 12 40 10</span>
<span class="co"># as.mo("stau") 160.0 180.0 200 200.0 210 220 10</span>
<span class="co"># as.mo("STAU") 160.0 180.0 190 190.0 200 210 10</span>
<span class="co"># as.mo("staaur") 9.8 12.0 15 12.0 12 42 10</span>
<span class="co"># as.mo("STAAUR") 8.4 8.7 13 10.0 12 37 10</span>
<span class="co"># as.mo("S. aureus") 13.0 16.0 38 18.0 45 150 10</span>
<span class="co"># as.mo("S aureus") 12.0 17.0 21 17.0 18 48 10</span>
<span class="co"># as.mo("Staphylococcus aureus") 7.1 8.7 12 9.7 11 38 10</span>
<span class="co"># as.mo("Staphylococcus aureus (MRSA)") 880.0 920.0 930 930.0 960 980 10</span>
<span class="co"># as.mo("Sthafilokkockus aaureuz") 400.0 430.0 450 440.0 460 500 10</span>
<span class="co"># as.mo("MRSA") 8.6 12.0 20 12.0 37 42 10</span>
<span class="co"># as.mo("VISA") 15.0 17.0 20 18.0 19 40 10</span>
<span class="co"># as.mo("VRSA") 13.0 14.0 19 17.0 19 46 10</span>
<span class="co"># as.mo(22242419) 140.0 140.0 160 150.0 170 210 10</span>
<span class="co"># expr min lq mean median uq max</span>
<span class="co"># as.mo("sau") 12.0 12.0 24.0 15.0 40.0 43.0</span>
<span class="co"># as.mo("stau") 170.0 170.0 190.0 180.0 210.0 250.0</span>
<span class="co"># as.mo("STAU") 160.0 180.0 200.0 190.0 220.0 230.0</span>
<span class="co"># as.mo("staaur") 9.4 11.0 21.0 13.0 40.0 48.0</span>
<span class="co"># as.mo("STAAUR") 9.0 13.0 34.0 14.0 43.0 140.0</span>
<span class="co"># as.mo("S. aureus") 16.0 18.0 20.0 19.0 21.0 25.0</span>
<span class="co"># as.mo("S aureus") 15.0 16.0 20.0 18.0 21.0 39.0</span>
<span class="co"># as.mo("Staphylococcus aureus") 1.1 1.1 1.4 1.6 1.6 1.7</span>
<span class="co"># as.mo("Staphylococcus aureus (MRSA)") 870.0 920.0 950.0 940.0 980.0 1000.0</span>
<span class="co"># as.mo("Sthafilokkockus aaureuz") 390.0 410.0 440.0 440.0 460.0 490.0</span>
<span class="co"># as.mo("MRSA") 11.0 12.0 30.0 13.0 40.0 130.0</span>
<span class="co"># as.mo("VISA") 16.0 18.0 30.0 20.0 46.0 69.0</span>
<span class="co"># as.mo("VRSA") 14.0 19.0 33.0 33.0 47.0 51.0</span>
<span class="co"># neval</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
<span class="co"># 10</span>
</pre></div>
<p><img src="benchmarks_files/figure-html/unnamed-chunk-4-1.png" width="562.5"></p>
<p>In the table above, all measurements are shown in milliseconds (thousands of seconds). A value of 5 milliseconds means it can determine 200 input values per second. It case of 100 milliseconds, this is only 10 input values per second.</p>
<p>In the table above, all measurements are shown in milliseconds (thousands of seconds). A value of 5 milliseconds means it can determine 200 input values per second. It case of 100 milliseconds, this is only 10 input values per second. It is clear that accepted taxonomic names are extremely fast, but some variations can take up to 500-1000 times as much time.</p>
<p>To improve performance, two important calculations take almost no time at all: <strong>repetitive results</strong> and <strong>already precalculated results</strong>.</p>
<div id="repetitive-results" class="section level3">
<h3 class="hasAnchor">
@ -255,50 +267,46 @@
<div class="sourceCode" id="cb3"><pre class="downlit">
<span class="co"># take all MO codes from the example_isolates data set</span>
<span class="kw">x</span> <span class="op">&lt;-</span> <span class="kw">example_isolates</span><span class="op">$</span><span class="kw">mo</span> <span class="op">%&gt;%</span>
<span class="co"># keep only the unique ones</span>
<span class="fu"><a href="https://rdrr.io/r/base/unique.html">unique</a></span>() <span class="op">%&gt;%</span>
<span class="co"># pick 50 of them at random</span>
<span class="fu"><a href="https://rdrr.io/r/base/sample.html">sample</a></span>(<span class="fl">50</span>) <span class="op">%&gt;%</span>
<span class="co"># paste that 10,000 times</span>
<span class="fu"><a href="https://rdrr.io/r/base/rep.html">rep</a></span>(<span class="fl">10000</span>) <span class="op">%&gt;%</span>
<span class="co"># scramble it</span>
<span class="co"># and copy them a thousand times</span>
<span class="fu"><a href="https://rdrr.io/r/base/rep.html">rep</a></span>(<span class="fl">1000</span>) <span class="op">%&gt;%</span>
<span class="co"># then scramble them</span>
<span class="fu"><a href="https://rdrr.io/r/base/sample.html">sample</a></span>()
<span class="co"># got indeed 50 times 10,000 = half a million?</span>
<span class="co"># as the example_isolates has 2,000 rows, we should have 2 million items</span>
<span class="fu"><a href="https://rdrr.io/r/base/length.html">length</a></span>(<span class="kw">x</span>)
<span class="co"># [1] 500000</span>
<span class="co"># [1] 2000000</span>
<span class="co"># and how many unique values do we have?</span>
<span class="fu"><a href="https://dplyr.tidyverse.org/reference/n_distinct.html">n_distinct</a></span>(<span class="kw">x</span>)
<span class="co"># [1] 50</span>
<span class="co"># [1] 90</span>
<span class="co"># now let's see:</span>
<span class="kw">run_it</span> <span class="op">&lt;-</span> <span class="fu">microbenchmark</span>(<span class="fu"><a href="../reference/mo_property.html">mo_name</a></span>(<span class="kw">x</span>),
times = <span class="fl">10</span>)
<span class="fu"><a href="https://rdrr.io/r/base/print.html">print</a></span>(<span class="kw">run_it</span>, unit = <span class="st">"ms"</span>, signif = <span class="fl">3</span>)
<span class="co"># Unit: milliseconds</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># mo_name(x) 1750 1790 1830 1810 1850 1950 10</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># mo_name(x) 90.3 101 120 102 141 202 10</span>
</pre></div>
<p>So transforming 500,000 values (!!) of 50 unique values only takes 1.81 seconds. You only lose time on your unique input values.</p>
<p>So getting official taxonomic names of 2,000,000 (!!) items consisting of 90 unique values only takes 0.102 seconds. You only lose time on your unique input values.</p>
</div>
<div id="precalculated-results" class="section level3">
<h3 class="hasAnchor">
<a href="#precalculated-results" class="anchor"></a>Precalculated results</h3>
<p>What about precalculated results? If the input is an already precalculated result of a helper function like <code><a href="../reference/mo_property.html">mo_name()</a></code>, it almost doesn’t take any time at all (see ‘C’ below):</p>
<div class="sourceCode" id="cb4"><pre class="downlit">
<span class="kw">run_it</span> <span class="op">&lt;-</span> <span class="fu">microbenchmark</span>(A = <span class="fu"><a href="../reference/mo_property.html">mo_name</a></span>(<span class="st">"B_STPHY_AURS"</span>),
<span class="kw">run_it</span> <span class="op">&lt;-</span> <span class="fu">microbenchmark</span>(A = <span class="fu"><a href="../reference/mo_property.html">mo_name</a></span>(<span class="st">"STAAUR"</span>),
B = <span class="fu"><a href="../reference/mo_property.html">mo_name</a></span>(<span class="st">"S. aureus"</span>),
C = <span class="fu"><a href="../reference/mo_property.html">mo_name</a></span>(<span class="st">"Staphylococcus aureus"</span>),
times = <span class="fl">10</span>)
<span class="fu"><a href="https://rdrr.io/r/base/print.html">print</a></span>(<span class="kw">run_it</span>, unit = <span class="st">"ms"</span>, signif = <span class="fl">3</span>)
<span class="co"># Unit: milliseconds</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># A 6.08 6.23 10.40 6.56 7.03 44.90 10</span>
<span class="co"># B 11.70 12.00 12.70 12.70 13.60 13.90 10</span>
<span class="co"># C 1.05 1.11 1.19 1.13 1.25 1.55 10</span>
<span class="co"># A 7.08 7.29 8.00 8.25 8.49 9.22 10</span>
<span class="co"># B 12.30 13.50 14.20 14.50 14.70 14.80 10</span>
<span class="co"># C 2.14 2.26 7.35 2.38 2.51 52.30 10</span>
</pre></div>
<p>So going from <code><a href="../reference/mo_property.html">mo_name("Staphylococcus aureus")</a></code> to <code>"Staphylococcus aureus"</code> takes 0.0011 seconds - it doesn’t even start calculating <em>if the result would be the same as the expected resulting value</em>. That goes for all helper functions:</p>
<p>So going from <code><a href="../reference/mo_property.html">mo_name("Staphylococcus aureus")</a></code> to <code>"Staphylococcus aureus"</code> takes 0.0024 seconds - it doesn’t even start calculating <em>if the result would be the same as the expected resulting value</em>. That goes for all helper functions:</p>
<div class="sourceCode" id="cb5"><pre class="downlit">
<span class="kw">run_it</span> <span class="op">&lt;-</span> <span class="fu">microbenchmark</span>(A = <span class="fu"><a href="../reference/mo_property.html">mo_species</a></span>(<span class="st">"aureus"</span>),
B = <span class="fu"><a href="../reference/mo_property.html">mo_genus</a></span>(<span class="st">"Staphylococcus"</span>),
@ -311,15 +319,15 @@
times = <span class="fl">10</span>)
<span class="fu"><a href="https://rdrr.io/r/base/print.html">print</a></span>(<span class="kw">run_it</span>, unit = <span class="st">"ms"</span>, signif = <span class="fl">3</span>)
<span class="co"># Unit: milliseconds</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># A 0.886 1.010 1.040 1.020 1.06 1.25 10</span>
<span class="co"># B 1.010 1.030 1.150 1.040 1.27 1.64 10</span>
<span class="co"># C 0.885 1.030 1.110 1.060 1.26 1.29 10</span>
<span class="co"># D 0.812 0.822 1.000 1.000 1.05 1.43 10</span>
<span class="co"># E 0.827 0.989 1.070 1.030 1.23 1.35 10</span>
<span class="co"># F 0.887 0.994 1.070 1.040 1.08 1.35 10</span>
<span class="co"># G 0.812 0.839 0.969 0.916 1.04 1.32 10</span>
<span class="co"># H 0.815 1.020 1.090 1.050 1.30 1.37 10</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># A 1.29 1.38 1.64 1.47 1.84 2.28 10</span>
<span class="co"># B 1.27 1.62 1.76 1.69 1.82 2.71 10</span>
<span class="co"># C 1.28 1.32 1.56 1.48 1.77 2.09 10</span>
<span class="co"># D 1.29 1.46 1.68 1.66 1.77 2.24 10</span>
<span class="co"># E 1.26 1.39 5.34 1.64 1.77 39.00 10</span>
<span class="co"># F 1.26 1.33 1.58 1.44 1.80 2.14 10</span>
<span class="co"># G 1.32 1.51 1.65 1.68 1.75 2.05 10</span>
<span class="co"># H 1.31 1.43 1.71 1.68 1.86 2.49 10</span>
</pre></div>
<p>Of course, when running <code><a href="../reference/mo_property.html">mo_phylum("Firmicutes")</a></code> the function has zero knowledge about the actual microorganism, namely <em>S. aureus</em>. But since the result would be <code>"Firmicutes"</code> anyway, there is no point in calculating the result. And because this package ‘knows’ all phyla of all known bacteria (according to the Catalogue of Life), it can just return the initial value immediately.</p>
</div>
@ -347,14 +355,14 @@
times = <span class="fl">100</span>)
<span class="fu"><a href="https://rdrr.io/r/base/print.html">print</a></span>(<span class="kw">run_it</span>, unit = <span class="st">"ms"</span>, signif = <span class="fl">4</span>)
<span class="co"># Unit: milliseconds</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># en 12.88 13.52 16.72 14.63 16.00 55.03 100</span>
<span class="co"># de 13.79 14.51 18.36 15.11 16.67 136.90 100</span>
<span class="co"># nl 17.72 18.59 22.30 20.15 21.87 54.69 100</span>
<span class="co"># es 13.78 14.38 19.16 15.35 16.86 49.96 100</span>
<span class="co"># it 13.83 14.40 18.57 15.24 16.32 58.12 100</span>
<span class="co"># fr 13.72 14.47 19.67 15.21 17.46 52.22 100</span>
<span class="co"># pt 13.73 14.43 17.76 15.10 16.85 51.69 100</span>
<span class="co"># expr min lq mean median uq max neval</span>
<span class="co"># en 13.29 13.54 17.53 13.70 14.93 58.25 100</span>
<span class="co"># de 14.25 14.46 19.09 14.69 16.23 58.96 100</span>
<span class="co"># nl 17.89 18.46 24.37 19.05 21.14 70.25 100</span>
<span class="co"># es 14.05 14.41 18.08 14.72 16.11 57.07 100</span>
<span class="co"># it 14.07 14.38 19.18 14.63 16.40 58.14 100</span>
<span class="co"># fr 13.98 14.42 17.30 14.57 15.31 56.81 100</span>
<span class="co"># pt 13.95 14.38 17.78 14.60 16.32 57.53 100</span>
</pre></div>
<p>Currently supported are German, Dutch, Spanish, Italian, French and Portuguese.</p>
</div>