Tracking now behaves better during timeouts

Author: erezsh

data_diff/diff_tables.py

@@ -14,7 +14,15 @@
 
 from .tracking import create_end_event_json, create_start_event_json, send_event_json, is_tracking_enabled
 from .sql import Select, Checksum, Compare, Count, TableName, Time, Value
-from .utils import CaseAwareMapping, CaseInsensitiveDict, safezip, split_space, CaseSensitiveDict, ArithString
+from .utils import (
+    CaseAwareMapping,
+    CaseInsensitiveDict,
+    safezip,
+    split_space,
+    CaseSensitiveDict,
+    ArithString,
+    run_as_daemon,
+)
 from .databases.base import Database
 from .databases.database_types import (
     DbPath,
@@ -334,7 +342,7 @@ def diff_tables(self, table1: TableSegment, table2: TableSegment) -> DiffResult:
         if is_tracking_enabled():
             options = dict(self)
             event_json = create_start_event_json(options)
-            send_event_json(event_json)
+            run_as_daemon(send_event_json, event_json)
 
         self.stats["diff_count"] = 0
         start = time.monotonic()

data_diff/tracking.py

@@ -16,6 +16,7 @@
 START_EVENT = "os_diff_run_start"
 END_EVENT = "os_diff_run_end"
 TOKEN = "ccb8c3a6-3b6f-445c-ad67-994efa7bd020"
+TIMEOUT = 8
 
 DEFAULT_PROFILE = os.path.expanduser("~/.datadiff.toml")
 
@@ -106,7 +107,7 @@ def send_event_json(event_json):
     data = json.dumps(event_json).encode()
     try:
         req = urllib.request.Request(TRACK_URL, data=data, headers=headers)
-        with urllib.request.urlopen(req) as f:
+        with urllib.request.urlopen(req, timeout=TIMEOUT) as f:
             res = f.read()
             if f.code != 200:
                 raise RuntimeError(res)

data_diff/utils.py

@@ -7,6 +7,7 @@
 from uuid import UUID
 import operator
 import string
+import threading
 
 alphanums = string.digits + string.ascii_lowercase
 
@@ -245,3 +246,10 @@ def match_regexps(regexps: Dict[str, Any], s: str) -> Sequence[tuple]:
         m = re.match(regexp + "$", s)
         if m:
             yield m, v
+
+
+def run_as_daemon(threadfunc, *args):
+    th = threading.Thread(target=threadfunc, args=args)
+    th.daemon = True
+    th.start()
+    return th