From dbb688906aa7fb262bd47236c96cbed93778250a Mon Sep 17 00:00:00 2001
From: jtyoung84 <104453205+jtyoung84@users.noreply.github.com>
Date: Wed, 8 May 2024 09:32:38 -0700
Subject: [PATCH] feat: adds data enums

---
 .../data_name_patterns.py                     | 67 +++++++++++++++++++
 tests/test_data_name_patterns.py              | 39 +++++++++++
 2 files changed, 106 insertions(+)
 create mode 100644 src/aind_data_schema_models/data_name_patterns.py
 create mode 100644 tests/test_data_name_patterns.py

diff --git a/src/aind_data_schema_models/data_name_patterns.py b/src/aind_data_schema_models/data_name_patterns.py
new file mode 100644
index 0000000..c650105
--- /dev/null
+++ b/src/aind_data_schema_models/data_name_patterns.py
@@ -0,0 +1,67 @@
+"""Module for defining our data naming conventions"""
+
+from datetime import datetime
+from enum import Enum
+
+
+class RegexParts(str, Enum):
+    """Regular expression components to be re-used elsewhere"""
+
+    DATE = r"\d{4}-\d{2}-\d{2}"
+    TIME = r"\d{2}-\d{2}-\d{2}"
+
+
+class DataRegex(str, Enum):
+    """Regular expression patterns for different kinds of data and their properties"""
+
+    DATA = f"^(?P<label>.+?)_(?P<c_date>{RegexParts.DATE.value})_(?P<c_time>{RegexParts.TIME.value})$"
+    RAW = (
+        f"^(?P<platform_abbreviation>.+?)_(?P<subject_id>.+?)_(?P<c_date>{RegexParts.DATE.value})_(?P<c_time>"
+        f"{RegexParts.TIME.value})$"
+    )
+    DERIVED = (
+        f"^(?P<input>.+?_{RegexParts.DATE.value}_{RegexParts.TIME.value})_(?P<process_name>.+?)_(?P<c_date>"
+        f"{RegexParts.DATE.value})_(?P<c_time>{RegexParts.TIME.value})"
+    )
+    ANALYZED = (
+        f"^(?P<project_abbreviation>.+?)_(?P<analysis_name>.+?)_(?P<c_date>"
+        f"{RegexParts.DATE.value})_(?P<c_time>{RegexParts.TIME.value})$"
+    )
+    NO_UNDERSCORES = "^[^_]+$"
+    NO_SPECIAL_CHARS = '^[^<>:;"/|? \\_]+$'
+    NO_SPECIAL_CHARS_EXCEPT_SPACE = '^[^<>:;"/|?\\_]+$'
+
+
+class DataLevel(str, Enum):
+    """Data level name"""
+
+    DERIVED = "derived"
+    RAW = "raw"
+    SIMULATED = "simulated"
+
+
+class Group(str, Enum):
+    """Data collection group name"""
+
+    BEHAVIOR = "behavior"
+    EPHYS = "ephys"
+    MSMA = "MSMA"
+    OPHYS = "ophys"
+
+
+def datetime_to_name_string(dt: datetime) -> str:
+    """Take a date and time object, format it as a string"""
+    return dt.strftime("%Y-%m-%d_%H-%M-%S")
+
+
+def datetime_from_name_string(d: str, t: str) -> datetime:
+    """Take date and time strings, generate date and time objects"""
+    d = datetime.strptime(d, "%Y-%m-%d").date()
+    t = datetime.strptime(t, "%H-%M-%S").time()
+    return datetime.combine(d, t)
+
+
+def build_data_name(label: str, creation_datetime: datetime) -> str:
+    """Construct a valid data description name"""
+    dt_str = datetime_to_name_string(creation_datetime)
+    return f"{label}_{dt_str}"
diff --git a/tests/test_data_name_patterns.py b/tests/test_data_name_patterns.py
new file mode 100644
index 0000000..44ea190
--- /dev/null
+++ b/tests/test_data_name_patterns.py
@@ -0,0 +1,39 @@
+"""Tests classes in data_name_patterns module"""
+
+import unittest
+
+from aind_data_schema_models.data_name_patterns import (
+    datetime_to_name_string,
+    datetime_from_name_string,
+    build_data_name,
+    RegexParts,
+    DataRegex,
+    DataLevel,
+    Group,
+)
+
+
+class TestRegexParts(unittest.TestCase):
+    """Tests methods in RegexParts class"""
+
+    def test_patterns_success(self):
+        """Tests that the regex patterns match successfully."""
+
+        input_date = "2020-10-19"
+        input_time = "08-30-59"
+
+        self.assertRegex(input_date, RegexParts.DATE)
+        self.assertRegex(input_time, RegexParts.TIME)
+
+    def test_patterns_fail(self):
+        """Tests that the regex patterns match unsuccessfully."""
+
+        deformed_date = "10/19/2020"
+        deformed_time = "8:30:59"
+
+        self.assertNotRegex(deformed_date, RegexParts.DATE)
+        self.assertNotRegex(deformed_time, RegexParts.TIME)
+
+
+if __name__ == "__main__":
+    unittest.main()