parse number as int and float as appropiate

usc-isi-i2 · Jan 25, 2018 · 2274223 · 2274223
1 parent 3b462e1
commit 2274223
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 4 deletions.
diff --git a/etk/core.py b/etk/core.py
@@ -2021,11 +2021,11 @@ def parse_number(d, config=None):
             return None
 
         if isinstance(text, numbers.Number):
-            return float(text)
+            return str(text)
 
         try:
             text = text.strip().replace('\n', '').replace('\t', '')
-            num = float(text)
+            num = str(float(text)) if '.' in text else str(int(text))
             return num
         except:
             pass

diff --git a/etk/unit_tests/test_content_extractions.py b/etk/unit_tests/test_content_extractions.py
@@ -499,7 +499,12 @@ def test_extract_as_is_post_filter_3(self):
                 {
                     "description": "Noncombatant Status Asserted",
                     "id": "assertedcontestednoncombatantnoncombatantnotstatusstatus",
-                    "size": "red34"
+                    "size": "34.0"
+                },
+                {
+                    "description": "Noncombatant Status Asserted",
+                    "id": "assertedcontestednoncombatantnoncombatantnotstatusstatus",
+                    "size": "redme34"
                 }
             ]
         }
@@ -539,7 +544,9 @@ def test_extract_as_is_post_filter_3(self):
         }
         c = Core(extraction_config=e_config)
         r = c.process(doc)
-        self.assertEqual(r['knowledge_graph']['actor_size'][0]['value'], 54.0)
+        self.assertTrue(len(r['knowledge_graph']['actor_size']) == 2)
+        self.assertEqual(r['knowledge_graph']['actor_size'][0]['value'], '54')
+        self.assertEqual(r['knowledge_graph']['actor_size'][1]['value'], '34.0')
 
     def test_extract_as_is_artbitrary_path(self):
         doc = {