fix(toggl): disambiguate entries that seem like duplicates to Harvest

since Harvest data doesn't contain the start/end times of time entries, entries
on the same day, for the same person, with the same total duration and description/notes
can appear as duplicates

this fixes the notes for all such seemingly duplicate entries by appending a counter to
allow Harvest to disambiguate

Author: thekaveman

compiler_admin/services/toggl.py

@@ -128,6 +128,20 @@ def convert_to_harvest(
     info = project_info()
     source["Project"] = source["Project"].apply(lambda x: info.get(key=x, default=x))
 
+    # find duplicates based on a subset of columns
+    cols = ["Date", "Hours", "First name", "Last name", "Notes"]
+    is_duplicate = source.duplicated(subset=cols, keep=False)
+
+    if is_duplicate.any():
+        # Create a counter for the duplicate rows
+        counter = source[is_duplicate].groupby(cols).cumcount() + 1
+        group_size = source[is_duplicate].groupby(cols)["Notes"].transform("size")
+
+        # Update the 'Notes' column with the counter
+        source.loc[is_duplicate, "Notes"] = (
+            source.loc[is_duplicate, "Notes"] + " (" + counter.astype(str) + "/" + group_size.astype(str) + ")"
+        )
+
     files.write_csv(output_path, source, columns=output_cols)
 
 

tests/services/test_toggl.py

@@ -1,30 +1,27 @@
-from datetime import datetime, timedelta, date
-from io import BytesIO, StringIO
 import math
 import sys
+from datetime import date, datetime, timedelta
+from io import BytesIO, StringIO
 from pathlib import Path
 from tempfile import NamedTemporaryFile
 
 import pandas as pd
 import pytest
 
 import compiler_admin.services.toggl
+from compiler_admin.services.toggl import CONVERTERS, HARVEST_COLUMNS, JUSTWORKS_COLUMNS, TOGGL_COLUMNS
+from compiler_admin.services.toggl import __name__ as MODULE
 from compiler_admin.services.toggl import (
-    CONVERTERS,
-    __name__ as MODULE,
     _get_first_name,
     _get_last_name,
     _prepare_input,
     _str_timedelta,
     convert_to_harvest,
     convert_to_justworks,
     download_time_entries,
+    files,
     lock_time_entries,
     summarize,
-    TOGGL_COLUMNS,
-    HARVEST_COLUMNS,
-    JUSTWORKS_COLUMNS,
-    files,
 )
 
 
@@ -209,6 +206,28 @@ def test_convert_to_harvest_sample(toggl_file, harvest_file, mock_google_user_in
     assert output_df["Client"].eq("Test Client 123").all()
 
 
+def test_convert_to_harvest_with_duplicates(mock_google_user_info):
+    # Test that seemingly duplicate time entries are disambiguated
+
+    mock_google_user_info.return_value = {"First Name": "Test", "Last Name": "User"}
+    csv_input = """Email,Project,Task,Client,Start date,Start time,Duration,Description
+test@example.com,Compiler,Backend,ACME,2025-11-18,09:00:00,01:00:00,A task
+test@example.com,Compiler,Backend,ACME,2025-11-18,10:00:00,01:00:00,A task
+test@example.com,Compiler,Backend,ACME,2025-11-18,11:00:00,02:00:00,Another task
+"""
+    input_buffer = StringIO(csv_input)
+    output_buffer = StringIO()
+
+    convert_to_harvest(source_path=input_buffer, output_path=output_buffer, client_name="ACME")
+
+    output_buffer.seek(0)
+    df = pd.read_csv(output_buffer)
+
+    assert df.loc[0, "Notes"] == "[Backend] A task (1/2)"
+    assert df.loc[1, "Notes"] == "[Backend] A task (2/2)"
+    assert df.loc[2, "Notes"] == "[Backend] Another task"
+
+
 def test_convert_to_justworks_mocked(toggl_file, spy_files):
     convert_to_justworks(toggl_file)