Get test data file directly in rebalance script (#70107)

Rather than looking through tests.json to find the data file, look for it explicitly
within a given target to avoid problems processing data in tests.json.

hacking/shippable/rebalance.py

@@ -7,9 +7,14 @@
 """
 CLI tool that analyses a Shippable run's test result and re-balances the test targets into new groups.
 
-Before running this script you must run download.py like
+Before running this script you must run download.py like:
+
+    ./download.py https://app.shippable.com/github/<team>/<repo>/runs/<run_num> --test-results --job-number x --job-number y
+
+Or to get all job results from a run:
+
+    ./download.py https://app.shippable.com/github/<team>/<repo>/runs/<run_num> --test-results --all
 
-./download.py https://app.shippable.com/github/<team>/<repo>/runs/<run_num> --test-results --job-number x --job-number y
 
 Set the dir <team>/<repo>/<run_num> as the value of '-p/--test-path' for this script.
 """
@@ -23,6 +28,8 @@ import operator
 import os
 import re
 
+from glob import glob
+
 try:
     import argcomplete
 except ImportError:
@@ -72,29 +79,31 @@ def get_raw_test_targets(args, test_path):
     target_times = {}
 
     for job_id in os.listdir(test_path):
-        json_path = os.path.join(test_path, job_id, 'test.json')
-        if not os.path.isfile(json_path):
+        json_path = os.path.join(test_path, job_id, 'test', 'testresults', 'data')
+
+        # Some tests to do not have a data directory
+        if not os.path.exists(json_path):
+            continue
+
+        json_file = glob(os.path.join(json_path, '*integration-*.json'))[0]
+        if not os.path.isfile(json_file):
             if args.verbose:
-                print("The test json file '%s' does not exist or is not a file, skipping test job run" % json_path)
+                print("The test json file '%s' does not exist or is not a file, skipping test job run" % json_file)
             continue
 
-        with open(json_path, mode='rb') as fd:
+        with open(json_file, mode='rb') as fd:
             test_info = json.loads(fd.read().decode('utf-8'))
 
-        for test in test_info:
-            if not test.get('path', '').endswith('.json') or 'contents' not in test.keys():
-                continue
-
-            info = json.loads(test['contents'])
+        targets = test_info.get('targets', {})
 
-            for target_name, target_info in info.get('targets', {}).items():
-                target_runtime = int(target_info.get('run_time_seconds', 0))
+        for target_name, target_info in targets.items():
+            target_runtime = int(target_info.get('run_time_seconds', 0))
 
-                # If that target already is found and has a higher runtime than the current one, ignore this entry.
-                if target_times.get(target_name, 0) > target_runtime:
-                    continue
+            # If that target already is found and has a higher runtime than the current one, ignore this entry.
+            if target_times.get(target_name, 0) > target_runtime:
+                continue
 
-                target_times[target_name] = target_runtime
+            target_times[target_name] = target_runtime
 
     return dict(sorted(target_times.items(), key=lambda i: i[1], reverse=True))