ISO date conversion fix

- Fix to ISO date conversion where string is already in ISO-format.
- Ray dependency fix where checking for location of temporary directory.

Author: turukawa

CHANGELOG

@@ -1,5 +1,6 @@
-- 1.2.1: Minor change in Ray dependency
-- 1.2.0: Upgrading Pydantic from version 1.8 to version 2.10
+- 1.2.2: Fix to date ISO conversion where date str already in ISO-format, plus Ray dependency fix.
+- 1.2.1: Minor change in Ray dependency.
+- 1.2.0: Upgrading Pydantic from version 1.8 to version 2.10.
 - 1.1.3: Improving exception error messages to be more helpful.
 - 1.1.2: Fixes for `CATEGORY` parser for ambiguous term names.
 - 1.1.1: Fixes for action parser where field names contain ambiguous characters.

docs/changelog.md

@@ -3,11 +3,20 @@ title: Change log
 summary: Version history, including for legacy versions.
 authors:
   - Gavin Chait
-date: 2025-03-13
+date: 2025-09-11
 tags: wrangling, crosswalks, versions
 ---
 # Change log
 
+## Version 1.2.2 (2025-09-11)
+
+- Fix to ISO date conversion where string is already in ISO-format.
+- Ray dependency fix where checking for location of temporary directory.
+
+## Version 1.2.1 (2025-07-19)
+
+- Minor change in Ray dependency.
+
 ## Version 1.2.0 (2025-03-13)
 
 - Upgrading Pydantic from version 1 to 2. This is not backwardly compatible.

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "whyqd"
-version = "1.2.1"
+version = "1.2.2"
 description = "data wrangling simplicity, complete audit transparency, and at speed"
 authors = ["Gavin Chait <gchait@whythawk.com>"]
 license = "BSD-3-Clause"

whyqd/VERSION

@@ -1 +1 @@
-1.2.1
+1.2.2

whyqd/config/ray_init.py

@@ -10,7 +10,8 @@
 def clear_spillway() -> bool:
     # https://stackoverflow.com/a/56151260/295606
     # only do this if ray is not initialised, otherwise it'll crash. horribly.
-    WHYQD_SPILLWAY = ray._common.utils.get_user_temp_dir()
+    # WHYQD_SPILLWAY = ray._common.utils.get_user_temp_dir()
+    WHYQD_SPILLWAY = ray._private.utils.get_user_temp_dir()
     if not Path(WHYQD_SPILLWAY).exists():
         return False
     for path in Path(WHYQD_SPILLWAY).iterdir():
@@ -50,8 +51,8 @@ def ray_start(**kwargs):
             kwargs["object_store_memory"] = settings.WHYQD_MEMORY
         if not kwargs.get("_temp_dir") and settings.WHYQD_SPILLWAY:
             # TODO: On some environments, I'm getting weird timeouts
-            if not ray._common.utils.get_user_temp_dir().endswith(settings.WHYQD_SPILLWAY):
-                kwargs["_temp_dir"] = f"{ray._common.utils.get_user_temp_dir()}{settings.WHYQD_SPILLWAY}"
+            if not ray._private.utils.get_user_temp_dir().endswith(settings.WHYQD_SPILLWAY):
+                kwargs["_temp_dir"] = f"{ray._private.utils.get_user_temp_dir()}{settings.WHYQD_SPILLWAY}"
             pass
         # if not kwargs.get("_system_config") and settings.WHYQD_SPILLWAY:
         #     kwargs["_system_config"] = {

whyqd/parsers/datasource.py

@@ -625,6 +625,17 @@ def check_date_format(self, *, date_type: str, date_value: str) -> bool:
     ### PANDAS TYPE PARSERS
     ###################################################################################################
 
+    def _date_is_isoformat(self, x: str) -> bool:
+        """
+        Checks a string to see whether it can be coerced as an ISO-formatted date `YYYY-MM-DD`
+        https://stackoverflow.com/a/61569783
+        """
+        try:
+            datetime.fromisoformat(x.replace('Z', '+00:00'))
+        except (ValueError, TypeError):
+            return False
+        return True
+
     def _parse_dates(self, x: Union[None, str], dayfirst: bool = True) -> Union[pd.NaT, date.isoformat]:
         """
         This is the hard-won 'trust nobody', certainly not Americans, date parser. This requires that the user
@@ -652,6 +663,9 @@ def _parse_dates(self, x: Union[None, str], dayfirst: bool = True) -> Union[pd.N
         if isinstance(x, str) and len(x) <= 10 and ":" in x:
             # This specific variation of a date is interpreted as a time, so ...
             x = re.sub(r"[\\/,\.:;]", "-", x)
+        # Check if already in ISO Format
+        if self._date_is_isoformat(x):
+            return str(x).strip()[:10]
         # Check if to_datetime can handle things
         if not pd.isnull(pd.to_datetime(x, errors="coerce", dayfirst=dayfirst)):
             return date.isoformat(pd.to_datetime(x, errors="coerce", dayfirst=dayfirst))