From b15bde431d04c19c7c89b34e495cf19b62684a2e Mon Sep 17 00:00:00 2001
From: Argenis Leon <argenisleon@gmail.com>
Date: Thu, 21 Nov 2019 20:16:16 -0600
Subject: [PATCH] Fix json output

---
 optimus/ml/distancecluster.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/optimus/ml/distancecluster.py b/optimus/ml/distancecluster.py
index 409ff465..3929f908 100644
--- a/optimus/ml/distancecluster.py
+++ b/optimus/ml/distancecluster.py
@@ -1,7 +1,10 @@
+import simplejson as json
+
 from pyspark.sql import functions as F
 
 from optimus import Optimus
 from optimus.helpers.columns import name_col
+from optimus.helpers.json import json_converter
 from optimus.ml import keycollision
 from optimus.ml.contants import FINGERPRINT_COL, CLUSTER_COL, CLUSTER_SIZE_COL, RECOMMENDED_COL, COUNT_COL, \
     LEVENSHTEIN_DISTANCE
@@ -69,6 +72,8 @@ def levenshtein_json(df, input_col):
         for iv in v:
             a.append(kv_dict[iv])
 
+    # json.dump(data, outfile, indent=4, ensure_ascii=False, default=json_converter)
+    result = json.dumps(result, ignore_nan=True, default=json_converter)
     return result