From 25374b106d3263987a11bcd7f36fa264599384e7 Mon Sep 17 00:00:00 2001
From: Nishchay Karle <45297081+NishchayKarle@users.noreply.github.com>
Date: Mon, 30 Sep 2024 16:25:01 -0500
Subject: [PATCH] Add usage tracking project name (#3624)

Add option to deanonymize usage tracking data by tagging it with a name using parsl.Config.project_name.
---
 parsl/config.py                         |  5 +++++
 parsl/tests/unit/test_usage_tracking.py | 21 +++++++++++++++++++++
 parsl/usage_tracking/usage.py           |  8 ++++++++
 3 files changed, 34 insertions(+)

diff --git a/parsl/config.py b/parsl/config.py
index c3725eccf8..1358e99d28 100644
--- a/parsl/config.py
+++ b/parsl/config.py
@@ -83,6 +83,9 @@ class Config(RepresentationMixin, UsageInformation):
         Setting this field to 0 will disable usage tracking. Default (this field is not set): usage tracking is not enabled.
         Parsl only collects minimal, non personally-identifiable,
         information used for reporting to our funding agencies.
+    project_name: str, optional
+        Option to deanonymize usage tracking data.
+        If set, this value will be used as the project name in the usage tracking data and placed on the leaderboard.
     initialize_logging : bool, optional
         Make DFK optionally not initialize any logging. Log messages
         will still be passed into the python logging system under the
@@ -118,6 +121,7 @@ def __init__(self,
                  max_idletime: float = 120.0,
                  monitoring: Optional[MonitoringHub] = None,
                  usage_tracking: int = 0,
+                 project_name: Optional[str] = None,
                  initialize_logging: bool = True) -> None:
 
         executors = tuple(executors or [])
@@ -154,6 +158,7 @@ def __init__(self,
         self.max_idletime = max_idletime
         self.validate_usage_tracking(usage_tracking)
         self.usage_tracking = usage_tracking
+        self.project_name = project_name
         self.initialize_logging = initialize_logging
         self.monitoring = monitoring
         self.std_autopath: Optional[Callable] = std_autopath
diff --git a/parsl/tests/unit/test_usage_tracking.py b/parsl/tests/unit/test_usage_tracking.py
index 351355811c..1581249cd2 100644
--- a/parsl/tests/unit/test_usage_tracking.py
+++ b/parsl/tests/unit/test_usage_tracking.py
@@ -43,3 +43,24 @@ def test_invalid_types(level):
     # we can't instantiate TypeCheckError if we're in typeguard 2.x environment
     # because it does not exist... so check name using strings.
     assert ex.type.__name__ in ["TypeCheckError", "TypeError"]
+
+
+@pytest.mark.local
+def test_valid_project_name():
+    """Test valid project_name."""
+    assert (
+        Config(
+            usage_tracking=3,
+            project_name="unit-test",
+        ).project_name == "unit-test"
+    )
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("name", (1, 1.0, True, object()))
+def test_invalid_project_name(name):
+    """Test invalid project_name."""
+    with pytest.raises(Exception) as ex:
+        Config(usage_tracking=3, project_name=name)
+
+    assert ex.type.__name__ in ["TypeCheckError", "TypeError"]
diff --git a/parsl/usage_tracking/usage.py b/parsl/usage_tracking/usage.py
index 3730fcc464..c22eb529fe 100644
--- a/parsl/usage_tracking/usage.py
+++ b/parsl/usage_tracking/usage.py
@@ -114,6 +114,7 @@ def __init__(self, dfk, port=50077,
                                                 sys.version_info.minor,
                                                 sys.version_info.micro)
         self.tracking_level = self.check_tracking_level()
+        self.project_name = self.config.project_name
         self.start_time = None
         logger.debug("Tracking level: {}".format(self.tracking_level))
 
@@ -153,6 +154,9 @@ def construct_start_message(self) -> bytes:
                    'platform.system': platform.system(),
                    'tracking_level': int(self.tracking_level)}
 
+        if self.project_name:
+            message['project_name'] = self.project_name
+
         if self.tracking_level >= 2:
             message['components'] = get_parsl_usage(self.dfk._config)
 
@@ -188,6 +192,10 @@ def construct_end_message(self) -> bytes:
                    'end': end_time,
                    'execution_time': end_time - self.start_time,
                    'components': [dfk_component] + get_parsl_usage(self.dfk._config)}
+
+        if self.project_name:
+            message['project_name'] = self.project_name
+
         logger.debug(f"Usage tracking end message (unencoded): {message}")
 
         return self.encode_message(message)