handle addressit case where parsed_text.street is produced and parsed_text.name is not

sanitiser/_tokenizer.js

@@ -27,9 +27,26 @@ function sanitize( raw, clean ){
 
   // if the text parser has run then we only tokenize the 'name' section
   // of the 'parsed_text' object, ignoring the 'admin' parts.
-  if( clean.hasOwnProperty('parsed_text') && clean.parsed_text.hasOwnProperty('name') ){
+  if( clean.hasOwnProperty('parsed_text') ) {
     inputParserRanSuccessfully = true;
-    text = clean.parsed_text.name; // use this string instead
+
+    // parsed_text.name is set, this is the highest priority, use this string
+    if( clean.parsed_text.hasOwnProperty('name') ){
+      text = clean.parsed_text.name; // use this string instead
+    }
+
+    // else handle the case where parsed_text.street was produced but
+    // no parsed_text.name is produced.
+    // additionally, handle the case where parsed_text.number is present
+    // note: the addressit module may also produce parsed_text.unit info
+    // for now, we discard that information as we don't have an appropriate
+    else if( clean.parsed_text.hasOwnProperty('street') ){
+      text = [
+        clean.parsed_text.number,
+        clean.parsed_text.street
+      ].filter(function(el){return el;})
+      .join(' '); // remove empty elements
+    }
   }
 
   // always set 'clean.tokens*' arrays for consistency and to avoid upstream errors.

test/unit/sanitiser/_tokenizer.js

@@ -81,6 +81,38 @@ module.exports.tests.sanity_checks = function(test, common) {
     t.deepEquals(messages.errors, [], 'no errors');
     t.deepEquals(messages.warnings, [], 'no warnings');
 
+    t.end();
+  });
+  test('favor clean.parsed_text street data over clean.text', function(t) {
+
+    var clean = { parsed_text: { number: '190', street: 'foo st' }, text: 'bar' };
+    var messages = sanitiser({}, clean);
+
+    // favor clean.parsed_text.name over clean.text
+    t.deepEquals(clean.tokens, [ '190', 'foo', 'st' ], 'use street name + number');
+    t.deepEquals(clean.tokens_complete, [ '190', 'foo', 'st' ], 'use street name + number');
+    t.deepEquals(clean.tokens_incomplete, [], 'no tokens');
+
+    // no errors/warnings produced
+    t.deepEquals(messages.errors, [], 'no errors');
+    t.deepEquals(messages.warnings, [], 'no warnings');
+
+    t.end();
+  });
+  test('favor clean.parsed_text.name over clean.parsed_text street data', function(t) {
+
+    var clean = { parsed_text: { number: '190', street: 'foo st', name: 'foo' }, text: 'bar' };
+    var messages = sanitiser({}, clean);
+
+    // favor clean.parsed_text.name over all other variables
+    t.deepEquals(clean.tokens, [ 'foo' ], 'use clean.parsed_text.name');
+    t.deepEquals(clean.tokens_complete, [ 'foo' ], 'use clean.parsed_text.name');
+    t.deepEquals(clean.tokens_incomplete, [], 'no tokens');
+
+    // no errors/warnings produced
+    t.deepEquals(messages.errors, [], 'no errors');
+    t.deepEquals(messages.warnings, [], 'no warnings');
+
     t.end();
   });
 };