Merge pull request #1 from Baroquem/dev

Merge changes for initial version

.gitignore

@@ -1 +1,3 @@
 node_modules
+yarn.lock
+.DS_Store

client/package.json

@@ -1,5 +1,5 @@
 {
-  "name": "pubmed-search",
+  "name": "client",
   "version": "0.1.0",
   "private": true,
   "dependencies": {
@@ -8,15 +8,22 @@
     "@testing-library/user-event": "^7.1.2",
     "axios": "^0.19.0",
     "bootstrap": "^4.4.1",
+    "core-js": "^3.6.4",
+    "es6-promise": "^4.2.8",
+    "isomorphic-fetch": "^2.2.1",
     "mongodb": "^3.4.0",
     "mongodb-client-encryption": "^1.0.0",
     "mongoose": "^5.8.1",
     "react": "^16.12.0",
     "react-dom": "^16.12.0",
     "react-router-dom": "^5.1.2",
-    "react-scripts": "3.3.0",
-    "react-spinners": "^0.8.0"
+    "react-scripts": "^3.4.0",
+    "react-spinners": "^0.8.0",
+    "serve": "^11.3.0"
   },
+  "bundledDependencies": [
+    "serve"
+  ],
   "scripts": {
     "start": "react-scripts start",
     "build": "react-scripts build",

client/src/App.js

@@ -7,6 +7,9 @@ import Bookbag from './Bookbag';
 import SearchPage from './SearchPage';
 import SearchTips from './SearchTips';
 
+require('es6-promise').polyfill();
+require('isomorphic-fetch');
+
 class App extends Component {
   // initialize state
   state = {

client/src/SearchTips.js

@@ -14,14 +14,15 @@ const SearchTips = () => {
       <ol>
         <li>By default, words entered into the search box are treated as an OR boolean search. For example, typing <span className="example">aspirin tylenol</span> will return a list of all article citations that include the term "aspirin" <i>or</i> the term "tylenol."
         </li>
+        <li>To perform a boolean AND search, enclose each search term individually in double quotes. Typing <span className="example">"aspirin" "tylenol"</span>will return a list of all article citations that include the term "aspirin" <i>and</i> the term "tylenol." (Note, however, that this type of search will only match exact terms. <span className="example">"puppies" "kittens"</span> will return matches only for the exact words 'puppies' and 'kittens', not 'puppy' and 'kitten'.)</li>
         <li>You can search for a phrase by entering a set of words in quotation marks. E.g., typing <span className="example">"patient care"</span> will result in a search for the words <i>patient</i> and <i>care</i> together, not individually.</li>
         <li>You can exclude terms from your search by prefacing them with a hypen (-). If you wanted to modify the previous search so that you only saw articles about patient care that did not involve cancer, you could type <span className="example">"patient care" -cancer</span>.</li>
         <li>Clicking one of the highlighted keywords in an article citation from a previous search result will immediately perform a keyword search for that term. This type of search returns a list of article citations that include that term <i>only</i> as a subject keyword.</li>
       </ol>
       <h4>Notes on search</h4>
       <ol>
         <li>Searches are case-insensitive and diacritic-insensitive. Regular text searches (i.e., non-keyword searches) look for query terms in a citation's title, abstract, and keywords, and are ranked by relevance.</li>
-        <li>The search engine uses a stemmer. Thus, for example, a search for "puppies" will include results for "puppy" as well.</li>
+        <li>The search engine uses a stemmer. Thus, for example, a search for "puppies" will include results for "puppy" as well (unless performing a boolean AND search).</li>
         <li>Searches that are too broad (e.g., "cancer"), will result in an error. You will see the message <span className="example">Your search could not be completed. Your topic may be too broad; try narrowing your search terms.</span>. This is a limitation of the system due to computer memory constraints. If you encounter this error, redo your search with a more limited query, e.g. "pancreatic cancer."</li>
         <li>More complex searches may take longer to produce results.</li>
       </ol>

client/src/index.js

@@ -1,3 +1,5 @@
+import "core-js/stable";
+import "regenerator-runtime/runtime";
 import "bootstrap/dist/css/bootstrap.min.css";
 import React from 'react';
 import ReactDOM from 'react-dom';

scripts/date_range_from_pubmed_xml.rb

@@ -0,0 +1,100 @@
+# Given a PubMed abstract index XML file, get the range of years
+# covered by the article citations
+require 'nokogiri'
+
+class CitationParser < Nokogiri::XML::SAX::Document
+
+
+  def initialize(filename)
+    @filename = filename
+  end
+
+  def start_document
+    @text = ''
+    @years = []
+    @in_pub_date = false
+    @capture = false
+  end
+
+  def start_element(element, attrs)
+    case element
+    when 'PubDate'
+      @in_pub_date = true
+    when 'Year'
+      @capture = @in_pub_date
+    end
+  end
+
+  def characters(str)
+    @text += str if @capture
+  end
+
+  def end_element(element)
+    case element
+    when 'Year'
+      if @capture
+        @years << Integer(@text)
+        @text = ''
+        @capture = false
+      end
+    when 'PubDate'
+      @in_pub_date = false
+    end
+  end
+
+  def end_document
+    #puts "Range found: #{@years.min}-#{@years.max} for #{@filename} (avg: #{mean(@years)}, mode: #{mode(@years)}, median: #{median(@years)})"
+    year_report = " ("
+    categories(@years).each do |k, v|
+      year_report += "#{k}: #{v}, "
+    end
+    year_report += ")"
+    puts "Range found: #{@years.min}-#{@years.max} for #{@filename}" + year_report
+  end
+
+  def mean(years)
+    @years.reduce(:+) / @years.count
+  end
+
+  def median(years)
+    sorted = years.sort
+    if years.count % 2 == 0
+      0.5 * (sorted[years.length / 2 - 1] + sorted[years.length / 2])
+    else
+      sorted[years.length / 2]
+    end
+  end
+
+  def mode(years)
+    frequencies = years.reduce(Hash.new(0)) { |h,v| h[v] += 1; h }
+    years.max_by { |y| frequencies[y] }
+  end
+
+  def categories(years)
+    {
+      :early => years.select { |y| y < 1960 }.count,
+      :_1960s => years.select { |y| y >= 1960 && y < 1970 }.count,
+      :_1970s => years.select { |y| y >= 1970 && y < 1980 }.count,
+      :_1980s => years.select { |y| y >= 1980 && y < 1990 }.count,
+      :_1990s => years.select { |y| y >= 1990 && y < 2000 }.count,
+      :_2000s => years.select { |y| y >= 2000 && y < 2010 }.count,
+      :_2010s => years.select { |y| y >= 2010 }.count,
+    }
+  end
+end
+
+
+# Main program
+if ARGV[0]
+  # PRocess the single file specified
+  files = [ARGV[0]]
+else
+  files = Dir['source_files/originals/*']
+end
+puts "\n\nWorking..."
+files.sort.each do |f|
+  parser = Nokogiri::XML::SAX::Parser.new(CitationParser.new(File.basename(f)))
+  #puts "Processing #{f}"
+  parser.parse_file f
+end
+puts "\nDone\n\n"

server/package.json

@@ -1,5 +1,5 @@
 {
-  "name": "backend",
+  "name": "server",
   "version": "1.0.0",
   "description": "",
   "main": "index.js",