fix: fix handling of datetime with less than 3 fractional digits for …

…seconds (#14) (#16)

* fix: fix handling of datetime with less than 3 fractional digits for seconds (#14)

* Fix pyright error: parser is not a known member of module dateutil

* Add tests for parse_time

aidial_analytics_realtime/app.py

@@ -1,7 +1,7 @@
 import json
 import logging
 import re
-from datetime import datetime, timezone
+from datetime import datetime
 
 import uvicorn
 from fastapi import Depends, FastAPI, Request
@@ -12,6 +12,7 @@
     create_influx_writer,
 )
 from aidial_analytics_realtime.rates import RatesCalculator
+from aidial_analytics_realtime.time import parse_time
 from aidial_analytics_realtime.topic_model import TopicModel
 from aidial_analytics_realtime.universal_api_utils import merge
 
@@ -173,10 +174,7 @@ async def on_log_message(
         response["upstream_uri"] if "upstream_uri" in response else ""
     )
 
-    timestamp = datetime.fromisoformat(request["time"])
-    if timestamp.tzinfo is None:
-        # The logs may come without the timezone information. We want it to be interpreted as UTC, not local time.
-        timestamp = timestamp.replace(tzinfo=timezone.utc)
+    timestamp = parse_time(request["time"])
 
     match = re.search(RATE_PATTERN, uri)
     if match:

aidial_analytics_realtime/time.py

@@ -0,0 +1,11 @@
+from datetime import datetime, timezone
+
+import dateutil.parser as dateutil_parser
+
+
+def parse_time(time_string: str) -> datetime:
+    timestamp = dateutil_parser.isoparse(time_string)
+    if timestamp.tzinfo is None:
+        # The logs may come without the timezone information. We want it to be interpreted as UTC, not local time.
+        timestamp = timestamp.replace(tzinfo=timezone.utc)
+    return timestamp

noxfile.py

@@ -6,7 +6,7 @@
 SRC = "."
 
 
-@nox.session
+@nox.session(python=["3.10"])
 def lint(session):
     """Runs linters and fixers"""
     session.run("poetry", "install", external=True)
@@ -17,7 +17,7 @@ def lint(session):
     session.run("pyright", SRC)
 
 
-@nox.session
+@nox.session(python=["3.10"])
 def format(session):
     """Runs linters and fixers"""
     session.run("poetry", "install", external=True)

pyproject.toml

@@ -25,6 +25,7 @@ torch = {version = "^2.0.1+cpu", source = "pytorch"}
 bertopic = "^0.15.0"
 llvmlite = "^0.40.1"
 python-dotenv = "^1.0.0"
+python-dateutil = "^2.8.2"
 
 [[tool.poetry.source]]
 name = "pytorch"

tests/test_app.py

@@ -51,10 +51,44 @@ def test_data_request():
                         },
                     }
                 )
-            }
+            },
+            {
+                "message": json.dumps(
+                    {
+                        "apiType": "DialOpenAI",
+                        "chat": {"id": "chat-2"},
+                        "project": {"id": "PROJECT-KEY-2"},
+                        "user": {"id": "", "title": ""},
+                        "request": {
+                            "protocol": "HTTP/1.1",
+                            "method": "POST",
+                            "uri": "/openai/deployments/gpt-4/chat/completions",
+                            "time": "2023-11-24T03:33:40.39",
+                            "body": json.dumps(
+                                {
+                                    "messages": [
+                                        {"role": "system", "content": ""},
+                                        {"role": "user", "content": "Hi!"},
+                                    ],
+                                    "model": "gpt-4",
+                                    "max_tokens": 2000,
+                                    "stream": True,
+                                    "n": 1,
+                                    "temperature": 0.0,
+                                }
+                            ),
+                        },
+                        "response": {
+                            "status": "200",
+                            "body": 'data: {"id":"chatcmpl-2","object":"chat.completion.chunk","created":1700828102,"model":"gpt-4","choices":[{"index":0,"delta":{"role":"assistant","content":"Hi"},"finish_reason":null}]}\n\ndata: {"id":"chatcmpl-2","object":"chat.completion.chunk","created":1700828102,"model":"gpt-4","choices":[{"index":0,"delta":{},"finish_reason":"stop"}],"usage":{"completion_tokens":189,"prompt_tokens":22,"total_tokens":211}}\n\ndata: [DONE]\n',
+                        },
+                    }
+                )
+            },
         ],
     )
     assert response.status_code == 200
     assert write_api_mock.points == [
-        'analytics,deployment=gpt-4,language=undefined,model=gpt-4,project_id=PROJECT-KEY,response_id=chatcmpl-1,title=undefined,topic=TestTopic,upstream=undefined chat_id="chat-1",completion_tokens=189i,number_request_messages=2i,price=0,prompt_tokens=22i,user_hash="undefined" 1692214959997000000'
+        'analytics,deployment=gpt-4,language=undefined,model=gpt-4,project_id=PROJECT-KEY,response_id=chatcmpl-1,title=undefined,topic=TestTopic,upstream=undefined chat_id="chat-1",completion_tokens=189i,number_request_messages=2i,price=0,prompt_tokens=22i,user_hash="undefined" 1692214959997000000',
+        'analytics,deployment=gpt-4,language=undefined,model=gpt-4,project_id=PROJECT-KEY-2,response_id=chatcmpl-2,title=undefined,topic=TestTopic,upstream=undefined chat_id="chat-2",completion_tokens=189i,number_request_messages=2i,price=0,prompt_tokens=22i,user_hash="undefined" 1700796820390000000',
     ]

tests/test_time.py

@@ -0,0 +1,64 @@
+from datetime import datetime, timedelta, timezone
+
+import pytest
+
+from aidial_analytics_realtime.time import parse_time
+
+
+@pytest.mark.parametrize(
+    "time_string, expected",
+    [
+        (
+            "2011-12-03T10:15:30",
+            datetime(2011, 12, 3, 10, 15, 30, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30+01:00",
+            datetime(
+                2011, 12, 3, 10, 15, 30, tzinfo=timezone(timedelta(hours=1))
+            ),
+        ),
+        (
+            "2011-12-03T10:15:30.1",
+            datetime(2011, 12, 3, 10, 15, 30, 100000, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.12",
+            datetime(2011, 12, 3, 10, 15, 30, 120000, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.123",
+            datetime(2011, 12, 3, 10, 15, 30, 123000, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.1234",
+            datetime(2011, 12, 3, 10, 15, 30, 123400, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.12345",
+            datetime(2011, 12, 3, 10, 15, 30, 123450, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.123456",
+            datetime(2011, 12, 3, 10, 15, 30, 123456, tzinfo=timezone.utc),
+        ),
+        (
+            "2011-12-03T10:15:30.1234567",
+            datetime(2011, 12, 3, 10, 15, 30, 123456, tzinfo=timezone.utc),
+        ),  # Python's datetime supports up to microsecond precision
+    ],
+)
+def test_parse_time(time_string: str, expected: datetime):
+    assert parse_time(time_string) == expected
+
+
+@pytest.mark.parametrize(
+    "time_string",
+    [
+        "2011-12-03T10:15:30.",  # No fractional part
+        "2011-12-03T10:15:30+01:00[Europe/Paris]",  # Named timezones are not supported
+    ],
+)
+def test_parse_time_should_fail(time_string: str):
+    with pytest.raises(ValueError):
+        parse_time(time_string)