Merge pull request #26 from metanorma/fix/isolate-term

remove field of application and second line from IEV retrieved entry:…
glossarist · Nov 22, 2021 · 62469c5 · 62469c5
2 parents 6e863bd + 743dd59
commit 62469c5
Show file tree

Hide file tree

Showing 4 changed files with 338 additions and 4 deletions.
diff --git a/lib/iev.rb b/lib/iev.rb
@@ -16,10 +16,13 @@ module Iev
   #
   def self.get(code, lang)
     url = "http://www.electropedia.org/iev/iev.nsf/"\
-      "display?openform&ievref=#{code}"
+          "display?openform&ievref=#{code}"
     doc = Nokogiri::HTML OpenURI.open_uri(url), nil, "UTF-8"
     xpath = "//table/tr/td/div/font[.=\"#{lang}\"]/../../"\
-      "following-sibling::td[2]"
-    doc.at(xpath)&.text&.strip
+            "following-sibling::td[2]"
+    a = doc&.at(xpath)&.children&.to_xml
+    a&.sub(%r{<br/>.*$}, "")
+      &.sub(%r{, &lt;.*$}, "")
+      &.gsub(%r{<[^>]*>}, "")&.strip
   end
 end
diff --git a/lib/iev/version.rb b/lib/iev/version.rb
@@ -1,3 +1,3 @@
 module Iev
-  VERSION = "0.2.6".freeze
+  VERSION = "0.3.0".freeze
 end
diff --git a/spec/examples/103_08_14.html b/spec/examples/103_08_14.html
@@ -0,0 +1,324 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
+<html>
+<head>
+<TITLE>IEC 60050 - International Electrotechnical Vocabulary - Details for IEV number 103-08-14: "<i>p</i>-fractile"</TITLE>
+<meta name="Description" content="Definition of <i>p</i>-fractile by Electropedia. Meaning of <i>p</i>-fractile. Translations of <i>p</i>-fractile. Equivalent terms for <i>p</i>-fractile." />
+<meta name="Keywords" content="<i>p</i>-fractile, &lt;of a probability distribution&gt;, <i>p</i>-quantile, &lt;of a probability distribution&gt;,fractile d'ordre <i>p</i>, &lt;d'une loi de probabilité&gt; m, quantile d'ordre <i>p</i>, &lt;d'une loi de probabilité&gt; m,(لتوزيع احتمالى) p- كمية , (لتوزيع احتمالى) p- كسرية ,<i>p</i>-Quantil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n, <i>p</i>-Fraktil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n,fractil de orden <i>p</i> (de una ley de probabilidad),<i>p</i>-frattile (di una distribuzione di probabilità), quantile di ordine <i>p</i> (di una distribuzione di probabilità),<i>p</i>－分位点, &lt;確率分布の&gt;, <i>p</i>－分位点, &lt;確率分布の&gt;, <i>p</i>-フラクタイル, &lt;確率分布の&gt;, <i>p</i>－クォンタイル, &lt;確率分布の&gt;,kwantyl rzędu <i>p</i>, fraktyl rzędu <i>p</i>,fractil-<i>p</i> (de uma distribuição de probabilidade), quantil-<i>p</i> (de uma distribuição de probabilidade),<i>p</i>-fraktil (av en sannolikhetsfördelning),<i>p</i> 분위수, &lt;확률분포&gt;,<i>p</i>-квантил, &lt;расподеле вероватноће&gt; м јд,p-分位数, &lt;一个概率分布的&gt;, dictionary, definition, meaning, translation, term, terminology, IEC 60050, Electropedia, IEV 103-08-14,IEV, International, Electrotechnical, Vocabulary, VEI, Vocabulaire, Electrotechnique, Internationale, IEC, electric, electronic" />
+<meta name="Copyright" content="IEC - International Electrotechnical Commission" />
+<meta name="Language" content="English, French, Arabic, Chinese, Czech, Finnish, German, Italian, Japanese, Norwegian, Polish, Portuguese, Russian, Serbian, Slovenian, Spanish, Swedish" />
+<SCRIPT language="JavaScript">   
+
+function printWindow() { bV = parseInt(navigator.appVersion); if (bV >= 4) window.print(); }
+
+function popup(name) { msgWindow=open(name,'popup','history=yes,toolbar=no,scrollbars=yes,resizable=yes,width=700,height=750'); }
+
+</SCRIPT>  
+
+<STYLE TYPE="text/css">    
+a:link {color: blue ; text-decoration: none} 
+a:visited { color: blue  ; text-decoration: none } 
+a:active { color: blue ; text-decoration: none } 
+a:hover { color: blue ; text-decoration: none; background-color: #C0C0C0}
+table {font-family: arial} 
+td { font-size: 11pt; font-family: Arial, Helvetica, sans-serif }
+h6 {page-break-before: always;} 
+</STYLE>
+
+<STYLE TYPE="text/css" media="print"> 
+.hide { display: none } 
+</STYLE>
+<!-- Google Analytics -->
+<script type="text/javascript">
+  var _gaq = _gaq || [];
+  _gaq.push(['_setAccount', 'UA-34400412-1']);
+  _gaq.push(['_trackPageview']);
+  (function() {
+    var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
+    ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
+    var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
+  })();
+</script>
+<!-- End Google Analytics -->
+<meta name="Description" content="<i>p</i>-fractile, &lt;of a probability distribution&gt;, <i>p</i>-quantile, &lt;of a probability distribution&gt;,fractile d'ordre <i>p</i>, &lt;d'une loi de probabilité&gt; m, quantile d'ordre <i>p</i>, &lt;d'une loi de probabilité&gt; m,(لتوزيع احتمالى) p- كمية , (لتوزيع احتمالى) p- كسرية ,<i>p</i>-Quantil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n, <i>p</i>-Fraktil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n,fractil de orden <i>p</i> (de una ley de probabilidad),<i>p</i>-frattile (di una distribuzione di probabilità), quantile di ordine <i>p</i> (di una distribuzione di probabilità),<i>p</i>－分位点, &lt;確率分布の&gt;, <i>p</i>－分位点, &lt;確率分布の&gt;, <i>p</i>-フラクタイル, &lt;確率分布の&gt;, <i>p</i>－クォンタイル, &lt;確率分布の&gt;,kwantyl rzędu <i>p</i>, fraktyl rzędu <i>p</i>,fractil-<i>p</i> (de uma distribuição de probabilidade), quantil-<i>p</i> (de uma distribuição de probabilidade),<i>p</i>-fraktil (av en sannolikhetsfördelning),<i>p</i> 분위수, &lt;확률분포&gt;,<i>p</i>-квантил, &lt;расподеле вероватноће&gt; м јд,p-分位数, &lt;一个概率分布的&gt;" >
+<STYLE TYPE="text/css">
+@font-face {    font-family: 'liberation_serifitalic';    
+src: url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_italic_macroman/LiberationSerif-Italic-webfont.eot');    
+src: url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_italic_macroman/LiberationSerif-Italic-webfont.eot?#iefix') format('embedded-opentype'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_italic_macroman/LiberationSerif-Italic-webfont.woff') format('woff'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_italic_macroman/LiberationSerif-Italic-webfont.ttf') format('truetype'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_italic_macroman/LiberationSerif-Italic-webfont.svg#liberation_serifitalic') format('svg');    
+}
+
+@font-face {    font-family: 'liberation_serifregular';    
+src: url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_regular_macroman/LiberationSerif-Regular-webfont.eot');    
+src: url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_regular_macroman/LiberationSerif-Regular-webfont.eot?#iefix') format('embedded-opentype'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_regular_macroman/LiberationSerif-Regular-webfont.woff') format('woff'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_regular_macroman/LiberationSerif-Regular-webfont.ttf') format('truetype'),         
+url('/font/Liberation-Serif-fontfacekit/web%20fonts/liberationserif_regular_macroman/LiberationSerif-Regular-webfont.svg#liberation_serifregular') format('svg');    
+font-weight: normal;    font-style: normal;}
+* {font-family:  liberation_serifregular;}body {font-family:  liberation_serifregular;}td {font-family:  liberation_serifregular;}i {font-family:  liberation_serifitalic;}#copy  td { font-size:11px; }#copy  a { font-size:11px; text-decoration: none;}#copy  {border-top: 1px #c0c0c0 solid;}hr {height: 1px; color: #c0c0c0;}</STYLE><script type="text/javascript" src="/mathjax/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
+<script language="JavaScript" type="text/javascript">
+<!-- 
+document._domino_target = "_self";
+function _doClick(v, o, t, h) {
+  var form = document._Display;
+  if (form.onsubmit) {
+     var retVal = form.onsubmit();
+     if (typeof retVal == "boolean" && retVal == false)
+       return false;
+  }
+  var target = document._domino_target;
+  if (o.href != null) {
+    if (o.target != null)
+       target = o.target;
+  } else {
+    if (t != null)
+      target = t;
+  }
+  form.target = target;
+  form.__Click.value = v;
+  if (h != null)
+    form.action += h;
+  form.submit();
+  return false;
+}
+// -->
+</script>
+</head>
+<body text="#000000" bgcolor="#FFFFFF" leftmargin=10 topmargin=0>
+
+<form method="post" action="/iev/iev.nsf/display?OpenForm&amp;Seq=1&amp;ievref=103-08-14" name="_Display">
+<input type="hidden" name="__Click" value="0" /><style>
+.Top {
+  margin: auto;
+  height: 109px;
+  padding: 0px;
+font-family: Arial, Helvetica;
+}
+.logolink {
+  position: relative;
+  width: 260px;
+  height: 60px;
+  top: 20px;
+  left: 8px;
+  float: left;
+  font-family: Arial, Helvetica, sans-serif;
+}
+.logolink h1 {
+  color: #000;
+  font-size: 17px;
+  font-weight: bold;
+  letter-spacing: 2px;
+float:left;
+top:-60px;
+margin:0;
+margin-left:80px;
+position:relative;
+  font-family: Arial, Helvetica, sans-serif;
+}
+.top {
+  outline: none;
+  margin: 0px;
+  padding: 0px;
+  font-family: Arial, Helvetica, sans-serif;
+  font-size: 12px;
+}
+</style><div class="Top">
+  <div class="logolink"><a title="Back to homepage" href="http://www.iec.ch/">
+  <img title="www.iec.ch" alt=" www.iec.ch" src="/img/logo.jpg"></a>
+    <h1>International</h1>
+    <h1>Electrotechnical</h1>
+    <h1>Commission</h1>
+  </div>
+</div>
+<table width="100%" border="0" cellspacing="0" cellpadding="0">
+<tr valign="top"><td width="1%" valign="middle"><img width="65" height="1" src="/icons/ecblank.gif" border="0" alt="" /><br />
+</td><td width="1%"><img width="19" height="1" src="/icons/ecblank.gif" border="0" alt="" /><br />
+</td><td width="46%" valign="middle"><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /><br />
+</td><td width="1%"><img width="19" height="1" src="/icons/ecblank.gif" border="0" alt="" /><br />
+</td><td width="54%"><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /><br />
+</td></tr>
+</table>
+<div align="right"><font size="2" face="Arial"> </font><a href="&amp;login"></a></div>
+<table style="width: 100%; margin-bottom: 10px;" cellpadding=0 width="100%" border="0" cellspacing="0" cellpadding="0">
+<tr valign="top"><td style="width: 144px;" width="15%"><img src="/iev/iev.nsf/30c867bbcbcee00dc1257e81003e697e/$Body/0.E5A?OpenElement&amp;FieldElemFormat=gif" width="144" height="20" /></td><td class="hide" style="border-top: 1px #c0c0c0 solid; width: 100% !important; " width="85%"><div align="right"><font size="2" face="Arial">Queries, comments, suggestions? Please </font><a href="mailto:[email protected]"><font size="2" face="Arial">contact us</font></a><font size="2" face="Arial">.</font></div></td></tr>
+</table>
+<br />
+
+<input type="button" value="   Home   " class="hide" style="font-size: 11px;" onclick="window.open('/iev/iev.nsf/6d6bdd8667c378f7c12581fa003d80e7?OpenForm', document._domino_target);" />
+<input type="button" onclick="history.back()" value="   Back   " class="hide" style="font-size: 11px;" />
+<input type="button" onclick="printWindow()" value="   Print     " class="hide" style="font-size: 11px;" /><style>
+	td.col1 {width:57px;}
+	td.col11 {width:55px;}
+	td.col2 {width:21px;}
+	td.col21 {width:19px;width:25px; text-align:left;}
+	td.col3 {width:567px;}
+	td.col4 {width:57px;}
+	table.table1 {width:710px;}
+</style>
+<font face=""><br>
+
+    <br />
+
+    <table class ="table1" cellpadding="1" cellspacing="0" border="0" cellpadding="0">
+        <tr valign="top">
+            <td class="col1"><b><font size="2" color="#800080" face="Arial">Area</font></b></td>
+	<td class="col2"><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+	<td class="col3"><b><font size="2" face="Arial"></font></b><a href=index?openform&part=103>Mathematics - Functions</a> / Probability </td>
+            <td class="col4"><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><b><font size="2" color="#800080" face="Arial">IEV ref</font></b></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td><b><b>103-08-14</b></b></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+        </tr>
+
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">en</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td><b><i>p</i>-fractile</b>, &lt;of a probability distribution&gt;<br><b><i>p</i>-quantile</b>, &lt;of a probability distribution&gt;</td>
+        </tr>
+        <tr valign="top">
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> for a number <i>p</i> between 0 and 1, value of a random variable for which the distribution function equals <i>p</i> or jumps from a value less than or equal to <i>p</i> to a value greater than <i>p</i>   </td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">fr</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <b>fractile d'ordre <i>p</i></b>, &lt;d'une loi de probabilité&gt; m<br><b>quantile d'ordre <i>p</i></b>, &lt;d'une loi de probabilité&gt; m</td>
+        </tr>
+        <tr valign="top">
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> pour un nombre <i>p</i> compris entre 0 et 1, valeur d'une variable aléatoire pour laquelle la fonction de répartition prend une valeur <i>p</i> ou saute d'une valeur inférieure ou égale à <i>p</i> à une valeur supérieure à <i>p</i> </td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">ar</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td style="direction:rtl; text-align:left" widthx="567"> (لتوزيع احتمالى) p- كمية <br>(لتوزيع احتمالى) p- كسرية </td>
+        </tr>
+
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">de</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <i>p</i>-Quantil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n<br><i>p</i>-Fraktil, &lt;einer Wahrscheinlichkeitsverteilung&gt; n</td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">es</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> fractil de orden <i>p</i> (de una ley de probabilidad)</td>
+        </tr>
+
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">it</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <i>p</i>-frattile (di una distribuzione di probabilità)<br>quantile di ordine <i>p</i> (di una distribuzione di probabilità)</td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080" face="Arial">ko</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <i>p</i> 분위수, &lt;확률분포&gt;</td>
+        </tr>
+
+        <tr valign="top">
+            <td width="701" colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">ja</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td lang="ja-jp"> <i>p</i>－分位点, &lt;確率分布の&gt;<br><i>p</i>－分位点, &lt;確率分布の&gt;<br><i>p</i>-フラクタイル, &lt;確率分布の&gt;<br><i>p</i>－クォンタイル, &lt;確率分布の&gt;</td>
+        </tr>
+
+
+
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080" face="Arial">pl</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> kwantyl rzędu <i>p</i><br>fraktyl rzędu <i>p</i></td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080" face="Arial">pt</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> fractil-<i>p</i> (de uma distribuição de probabilidade)<br>quantil-<i>p</i> (de uma distribuição de probabilidade)</td>
+        </tr>
+
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">sr</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <i>p</i>-квантил, &lt;расподеле вероватноће&gt; м јд</td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080">sv</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> <i>p</i>-fraktil (av en sannolikhetsfördelning)</td>
+        </tr>
+
+        <tr valign="top">
+            <td colspan="4"><hr></td>
+        </tr>
+        <tr valign="top">
+            <td><div align="center"><font color="#800080" face="Arial">zh</font></div></td>
+            <td><img width="1" height="1" src="/icons/ecblank.gif" border="0" alt="" /></td>
+            <td> p-分位数, &lt;一个概率分布的&gt;</td>
+        </tr>
+    </table>
+
+</font><hr>
+<table width="100%" border="0" cellspacing="0" cellpadding="0">
+<tr valign="top"><td width="25%"><a href="/iev/iev.nsf/display?openform&amp;ievref=103-08-14&amp;login"></a><font color="#800080">Publication date:</font>   2009-12  </td><td width="75%"><div align="right"><a href="http://www.iec.ch/about/copyright/copyright-db_entry.htm" target="_blank"><font color="#003399">Copyright</font></a> © <a href="http://www.iec.ch/about/co/office-e.htm" target="_blank"><font color="#003399">IEC</font></a> 2021. All Rights Reserved.</div></td></tr>
+</table>
+</form>
+</body>
+</html>
diff --git a/spec/iev_spec.rb b/spec/iev_spec.rb
@@ -22,6 +22,13 @@
     expect(File.exist?("testcache2")).to be_truthy
   end
 
+  # <td><b><i>p</i>-fractile</b>, &lt;of a probability distribution&gt;<br><b><i>p</i>-quantile</b>, &lt;of a probability distribution&gt;</td>
+  it "strips extraneous information from term" do
+    mock_open_uri("103-08-14")
+    term = @db.fetch "103-08-14", "en"
+    expect(term).to eq "p-fractile"
+  end
+
   it "return empty string if code not found" do
     mock_open_uri("111-11-11")
     term = @db.fetch "111-11-11", "en"