From 30cd30236a804ce706c87ff83554c39817b2f36f Mon Sep 17 00:00:00 2001
From: Stephen Hess <stephen.k.hess@gmail.com>
Date: Wed, 5 Jul 2017 08:36:43 -0400
Subject: [PATCH] don't call libpostal if sources=whosonfirst

---
 sanitizer/_text.js           | 26 ++++++++-------
 test/unit/sanitizer/_text.js | 61 ++++++++++++++++++++++++++++++++++++
 2 files changed, 75 insertions(+), 12 deletions(-)

diff --git a/sanitizer/_text.js b/sanitizer/_text.js
index 874a9b17..d7d1227b 100644
--- a/sanitizer/_text.js
+++ b/sanitizer/_text.js
@@ -1,29 +1,31 @@
-var check = require('check-types'),
-  text_analyzer = require('pelias-text-analyzer');
+const check = require('check-types');
+const text_analyzer = require('pelias-text-analyzer');
+const _ = require('lodash');
 
 // validate texts, convert types and apply defaults
 function sanitize( raw, clean ){
-
   // error & warning messages
-  var messages = { errors: [], warnings: [] };
+  const messages = { errors: [], warnings: [] };
 
   // invalid input 'text'
   // must call `!check.nonEmptyString` since `check.emptyString` returns
   //  `false` for `undefined` and `null`
   if( !check.nonEmptyString( raw.text ) ){
     messages.errors.push('invalid param \'text\': text length, must be >0');
-  }
 
-  // valid input 'text'
-  else {
-    // valid text
+  } else {
     clean.text = raw.text;
 
-    // parse text with query parser
-    var parsed_text = text_analyzer.parse(clean.text);
-    if (check.assigned(parsed_text)) {
-      clean.parsed_text = parsed_text;
+    // only call libpostal if there are other sources besides whosonfirst
+    //  since placeholder will take care of it later
+    if (!_.isEqual(clean.sources, ['whosonfirst'])) {
+      // parse text with query parser
+      const parsed_text = text_analyzer.parse(clean.text);
+      if (check.assigned(parsed_text)) {
+        clean.parsed_text = parsed_text;
+      }
     }
+
   }
 
   return messages;
diff --git a/test/unit/sanitizer/_text.js b/test/unit/sanitizer/_text.js
index c29f0a1d..a8a46137 100644
--- a/test/unit/sanitizer/_text.js
+++ b/test/unit/sanitizer/_text.js
@@ -142,6 +142,67 @@ module.exports.tests.text_parser = function(test, common) {
 
   });
 
+  test('sources=whosonfirst should not call text_analyzer and set clean.text from raw.text', (t) => {
+    const sanitizer = proxyquire('../../../sanitizer/_text', {
+      'pelias-text-analyzer': { parse: query => t.fail('should not have been called') }
+    });
+
+    const raw = {
+      text: 'raw clean.text'
+    };
+    const clean = {
+      sources: ['whosonfirst'],
+      text: 'original clean.text'
+    };
+
+    const expected_clean = {
+      sources: ['whosonfirst'],
+      text: 'raw clean.text'
+    };
+
+    const messages = sanitizer(raw, clean);
+
+    t.deepEquals(clean, expected_clean);
+    t.deepEquals(messages, { errors: [], warnings: [] });
+    t.end();
+
+  });
+
+  test('sources with whosonfirst + others should call analyzer', (t) => {
+    const sanitizer = proxyquire('../../../sanitizer/_text', {
+      'pelias-text-analyzer': { parse: function(query) {
+        return {
+          key1: 'value 1',
+          key2: 'value 2'
+        };
+      }
+    }});
+
+    const raw = {
+      text: 'raw text'
+    };
+    const clean = {
+      sources: ['whosonfirst', 'another source'],
+      text: 'clean text'
+    };
+
+    const expected_clean = {
+      sources: ['whosonfirst', 'another source'],
+      text: 'raw text',
+      parsed_text: {
+        key1: 'value 1',
+        key2: 'value 2'
+      }
+    };
+
+    const messages = sanitizer(raw, clean);
+
+    t.deepEquals(clean, expected_clean);
+    t.deepEquals(messages, { errors: [], warnings: [] });
+    t.end();
+
+  });
+
 };
 
 module.exports.all = function (tape, common) {