Redact the URL query parameters from the urllib3.connectionpool logs (#341)

* Redact the URL query parameters from the urllib3.connectionpool logs

Signed-off-by: Mubashir Kazia <[email protected]>

* Fix code formatting

Signed-off-by: Mubashir Kazia <[email protected]>

* Add str check for the log record message arg dict values

Signed-off-by: Mubashir Kazia <[email protected]>

---------

Signed-off-by: Mubashir Kazia <[email protected]>

Author: mkazia-db

src/databricks/sql/__init__.py

@@ -8,6 +8,39 @@
 
 paramstyle = "named"
 
+import re
+
+
+class RedactUrlQueryParamsFilter(logging.Filter):
+    pattern = re.compile(r"(\?|&)([\w-]+)=([^&]+)")
+    mask = r"\1\2=<REDACTED>"
+
+    def __init__(self):
+        super().__init__()
+
+    def redact(self, string):
+        return re.sub(self.pattern, self.mask, str(string))
+
+    def filter(self, record):
+        record.msg = self.redact(str(record.msg))
+        if isinstance(record.args, dict):
+            for k in record.args.keys():
+                record.args[k] = (
+                    self.redact(record.args[k])
+                    if isinstance(record.arg[k], str)
+                    else record.args[k]
+                )
+        else:
+            record.args = tuple(
+                (self.redact(arg) if isinstance(arg, str) else arg)
+                for arg in record.args
+            )
+
+        return True
+
+
+logging.getLogger("urllib3.connectionpool").addFilter(RedactUrlQueryParamsFilter())
+
 
 class DBAPITypeObject(object):
     def __init__(self, *values):