normalize-tabular-data 0.1.2__tar.gz → 0.1.3__tar.gz

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/LICENSE

@@ -1,4 +1,4 @@
-Copyright (c) 2026, Members of Service Employees International Union (SEIU)
+Copyright (c) 2026, Service Employees International Union (SEIU)
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: normalize-tabular-data
-Version: 0.1.2
-Summary: TUI for normalizing tabular data with polars
+Version: 0.1.3
+Summary: TUI for normalizing tabular data with Polars
 Author: David Mertz, Ph.D.
 Author-email: David Mertz, Ph.D. <mertz@gnosis.cx>
 License-Expression: BSD-2-Clause
@@ -11,6 +11,7 @@ Requires-Dist: textual>=8.2.8
 Requires-Dist: polars>=1.44,<2
 Requires-Dist: fastexcel>=0.10
 Requires-Dist: xlsxwriter>=0.9
+Requires-Dist: platformdirs>=3.0
 Requires-Python: >=3.14
 Description-Content-Type: text/markdown
 
@@ -119,17 +120,22 @@ Launch with `normalize-tabular-data`. Keys:
 
 ## Operations
 
-- **Normalize dates** — parse a messy date column of *any* input format into
-  canonical UTC `year-month-dayThh:mm:ss` values (second resolution;
-  fractional parts are truncated); unparseable values become null.
+- **Normalize date(t)imes** — parse a messy date column of *any* input
+  format into canonical UTC datetimes at millisecond resolution;
+  unparseable values become null. CSV/TSV/JSONLines export serializes
+  datetime columns as second-resolution `year-month-dayThh:mm:ss`
+  strings; Parquet and Excel keep the real datetime values.
+- **Normalize (d)ates** — the same match-anything parsing, condensed to
+  the UTC calendar date: the column becomes date-only ISO-8601
+  (`year-month-day`); unparseable values become null.
 - **Trim whitespace** — strip edges and collapse internal whitespace runs,
   per selected columns.
 - Rename a column — click its header in the preview and type the new name.
-- **Deduplicate rows** — on all or selected columns, keeping first or last.
 - **Combine columns** — concatenate two or more columns with a separator.
 - **Split column** — break one column into `{col}_1..{col}_k`; a blank
   delimiter splits on runs of whitespace.
 - **Remove columns** — drop selected columns entirely.
+- **Deduplicate rows** — on all or selected columns, keeping first or last.
 
 ## Formats
 
@@ -144,7 +150,9 @@ still runs on the full data.
 Every open file keeps an internal log of the operations performed on it. When
 saving, the dialog offers a **Save sequence of operations?** checkbox
 (unticked by default); with it ticked, a second dialog asks where to write
-the script — suggested `<table>.ntd`, any other extension you type is kept.
+the script — suggested `<table>.ntd`; scripts always use the `.ntd`
+extension, so the name you type is saved with that suffix whatever it
+ends in.
 
 The script is plain ASCII text, one operation per line, e.g.:
 

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/README.md

@@ -103,17 +103,22 @@ Launch with `normalize-tabular-data`. Keys:
 
 ## Operations
 
-- **Normalize dates** — parse a messy date column of *any* input format into
-  canonical UTC `year-month-dayThh:mm:ss` values (second resolution;
-  fractional parts are truncated); unparseable values become null.
+- **Normalize date(t)imes** — parse a messy date column of *any* input
+  format into canonical UTC datetimes at millisecond resolution;
+  unparseable values become null. CSV/TSV/JSONLines export serializes
+  datetime columns as second-resolution `year-month-dayThh:mm:ss`
+  strings; Parquet and Excel keep the real datetime values.
+- **Normalize (d)ates** — the same match-anything parsing, condensed to
+  the UTC calendar date: the column becomes date-only ISO-8601
+  (`year-month-day`); unparseable values become null.
 - **Trim whitespace** — strip edges and collapse internal whitespace runs,
   per selected columns.
 - Rename a column — click its header in the preview and type the new name.
-- **Deduplicate rows** — on all or selected columns, keeping first or last.
 - **Combine columns** — concatenate two or more columns with a separator.
 - **Split column** — break one column into `{col}_1..{col}_k`; a blank
   delimiter splits on runs of whitespace.
 - **Remove columns** — drop selected columns entirely.
+- **Deduplicate rows** — on all or selected columns, keeping first or last.
 
 ## Formats
 
@@ -128,7 +133,9 @@ still runs on the full data.
 Every open file keeps an internal log of the operations performed on it. When
 saving, the dialog offers a **Save sequence of operations?** checkbox
 (unticked by default); with it ticked, a second dialog asks where to write
-the script — suggested `<table>.ntd`, any other extension you type is kept.
+the script — suggested `<table>.ntd`; scripts always use the `.ntd`
+extension, so the name you type is saved with that suffix whatever it
+ends in.
 
 The script is plain ASCII text, one operation per line, e.g.:
 

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "normalize-tabular-data"
-version = "0.1.2"
-description = "TUI for normalizing tabular data with polars"
+version = "0.1.3"
+description = "TUI for normalizing tabular data with Polars"
 readme = "README.md"
 license = "BSD-2-Clause"
 license-files = ["LICENSE"]
@@ -12,6 +12,7 @@ dependencies = [
     "polars>=1.44,<2",
     "fastexcel>=0.10",
     "xlsxwriter>=0.9",
+    "platformdirs>=3.0",
 ]
 
 [[project.authors]]
@@ -27,6 +28,7 @@ dev = [
     "pytest-asyncio>=0.25",
     "ruff>=0.16.10",
     "twine>=7.0.0",
+    "ty>=0.0.84",
 ]
 
 [build-system]

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/pyproject.toml.orig

@@ -1,7 +1,7 @@
 [project]
 name = "normalize-tabular-data"
-version = "0.1.2"
-description = "TUI for normalizing tabular data with polars"
+version = "0.1.3"
+description = "TUI for normalizing tabular data with Polars"
 readme = "README.md"
 license = "BSD-2-Clause"
 license-files = ["LICENSE"]
@@ -15,6 +15,7 @@ dependencies = [
     "polars>=1.44,<2",
     "fastexcel>=0.10",
     "xlsxwriter>=0.9",
+    "platformdirs>=3.0",
 ]
 
 [dependency-groups]
@@ -23,6 +24,7 @@ dev = [
     "pytest-asyncio>=0.25",
     "ruff>=0.16.10",
     "twine>=7.0.0",
+    "ty>=0.0.84",
 ]
 
 [project.scripts]

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/__init__.py

@@ -6,7 +6,7 @@ import argparse
 import sys
 from pathlib import Path
 
-__version__ = "0.1.2"
+__version__ = "0.1.3"
 
 
 def _run_script(path: Path, script_path: Path, output: Path | None) -> int:

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/app.py

@@ -6,6 +6,7 @@ import bisect
 import datetime as _dt
 import random
 from pathlib import Path
+from platformdirs import user_config_dir
 from typing import Any, Iterable
 
 from textual import on
@@ -43,6 +44,7 @@ from normalize_tabular_data.screens import (
 from normalize_tabular_data.widgets import ColumnSidebar, MenuFooter, StepsBar
 
 PREVIEW_ROWS = 250
+APP_CONFIG_NAME = "normalize-tabular-data"
 
 
 class CommandMenu(CommandPalette):
@@ -135,10 +137,19 @@ class OpChooserModal(ModalDialog):
 
     def _marked_title(self, op_key: str, title: str, hotkey: str = "") -> str:
         """Return "Normalize (d)ates"-style markup: the op's hotkey letter
-        highlighted where it sits in the title (keeping its case). Ops
-        without a designated hotkey claim the first unused letter in the
-        title instead; plain title if no letter can be claimed."""
+        highlighted where it sits in the title (keeping its case). A title
+        that already parenthesizes its letter ("Normalize date(t)imes")
+        highlights the letter between the parens; a plain title matches
+        its first occurrence and adds the parens. Ops without a designated
+        hotkey claim the first unused letter in the title instead; plain
+        title if no letter can be claimed."""
         if hotkey:
+            i = title.lower().find(f"({hotkey.lower()})")
+            if i >= 0:
+                # the letter to claim sits inside existing parentheses:
+                # highlight it without adding a second pair
+                self._hotkeys[title[i + 1].lower()] = op_key
+                return f"{title[: i + 1]}[cyan]{title[i + 1]}[/cyan]{title[i + 2:]}"
             i = title.lower().find(hotkey.lower())
         else:
             i = next(
@@ -275,8 +286,15 @@ class NormalizeApp(App[None]):
         self,
         initial_file: str | Path | None = None,
         initial_script: str | Path | None = None,
+        config_dir: Path | None = None,
     ) -> None:
         super().__init__()
+        # where the chosen theme (and any future settings) are stored;
+        # None means the per-user platform directory
+        self._config_dir = config_dir
+        # flips to True at on_mount: only theme changes made after are the
+        # user's and only those are persisted
+        self._app_ready = False
         # filesystem path to open when the app starts (the optional file
         # named on the command line); loaded like a file chosen in-app
         self.initial_file: Path | None = (
@@ -354,6 +372,9 @@ class NormalizeApp(App[None]):
         return list(self.pipeline.current().columns)
 
     def on_mount(self) -> None:
+        # remember this point so theme changes from here on are the
+        # user's own choices and get persisted
+        self._apply_saved_theme()
         self.push_screen(MainScreen())
         self._refresh_steps()
         # the command line's file, if any, is opened once the main screen
@@ -361,6 +382,46 @@ class NormalizeApp(App[None]):
         # crashing if it is unreadable or has an unsupported format
         if self.initial_file is not None:
             self.call_after_refresh(self.load_path, self.initial_file)
+        self._app_ready = True
+
+    # --- theme persistence ---------------------------------------------
+
+    def watch_theme(self, theme_name: str) -> None:
+        """Persist every theme the user chooses (menu or command), so the
+        next launch opens with it again. Skipped while headless, so the
+        pilot tests cannot touch the real configuration; skipped before
+        mount, where only the reactive's default applies."""
+        if self._app_ready and not self.is_headless:
+            try:
+                self._theme_config_path().parent.mkdir(parents=True, exist_ok=True)
+                self._theme_config_path().write_text(
+                    theme_name + "\n", encoding="utf-8"
+                )
+            except OSError:
+                pass  # unwritable config location: run without remembering
+
+    def _apply_saved_theme(self) -> None:
+        """Restore the theme chosen in a previous session, if it still exists.
+
+        Skipped in headless runs that use the real configuration location
+        (pilot tests): an eyeballing developer's own saved theme should
+        not leak into supposedly default-themed tests. A headless run
+        given an explicit config_dir is a deliberate fixture and loads it."""
+        if self.is_headless and self._config_dir is None:
+            return
+        try:
+            saved = self._theme_config_path().read_text(encoding="utf-8").strip()
+        except OSError:
+            return
+        if not saved:
+            return
+        try:
+            self.theme = saved
+        except Exception:
+            pass  # a theme name that no longer exists: keep the default
+
+    def _theme_config_path(self) -> Path:
+        return (self._config_dir or Path(user_config_dir(APP_CONFIG_NAME))) / "theme"
 
     # --- data plumbing ---------------------------------------------------
 

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/io.py

@@ -79,6 +79,24 @@ def excel_sheets(path: Path) -> list[str]:
     return list(fastexcel.read_excel(path).sheet_names)
 
 
+def _second_resolution(df: pl.DataFrame) -> pl.DataFrame:
+    """Datetime columns serialized as canonical UTC second-resolution text.
+
+    Used only for the text formats (CSV/TSV/JSONL): polars pads datetime
+    output with sub-second digits no matter the unit, and the canonical
+    text form carries no fractional part. Parquet and Excel keep the real
+    Datetime values (millisecond resolution)."""
+    datetime_cols = [
+        c for c, dtype in df.schema.items() if isinstance(dtype, pl.Datetime)
+    ]
+    if not datetime_cols:
+        return df
+    return df.with_columns(
+        pl.col(c).dt.truncate("1s").dt.to_string("%Y-%m-%dT%H:%M:%S")
+        for c in datetime_cols
+    )
+
+
 def write_table(
     df: pl.DataFrame,
     path: Path,
@@ -86,11 +104,11 @@ def write_table(
     sheet_name: str = "data",
 ) -> None:
     if fmt == "csv":
-        df.write_csv(path)
+        _second_resolution(df).write_csv(path)
     elif fmt == "tsv":
-        df.write_csv(path, separator="\t")
+        _second_resolution(df).write_csv(path, separator="\t")
     elif fmt == "jsonl":
-        df.write_ndjson(path)
+        _second_resolution(df).write_ndjson(path)
     elif fmt == "parquet":
         df.write_parquet(path)
     elif fmt == "xlsx":

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/ops.py

@@ -48,19 +48,22 @@ def _apply_date_normalize(df: pl.DataFrame, p: dict[str, Any]) -> pl.DataFrame:
     series = df.get_column(col)
     if series.dtype != pl.String:
         series = series.cast(pl.String)
-    # canonical UTC at second resolution: polars has no second-unit
-    # Datetime (ns/us/ms only, all of which print sub-second digits in
-    # every export), so the normalized column is the truncated
-    # `year-month-dayThh:mm:ss` string; fractional parts are dropped and
-    # unparseable values become null
+    # canonical UTC Datetime at millisecond resolution; unparseable -> null
     return df.with_columns(
-        date_parser.parse_series(series)
-        .dt.truncate("1s")
-        .dt.to_string("%Y-%m-%dT%H:%M:%S")
-        .alias(col)
+        date_parser.parse_series(series).dt.cast_time_unit("ms").alias(col)
     )
 
 
+def _apply_date_only(df: pl.DataFrame, p: dict[str, Any]) -> pl.DataFrame:
+    col: str = p["column"]
+    series = df.get_column(col)
+    if series.dtype != pl.String:
+        series = series.cast(pl.String)
+    # date-only ISO-8601: any input format condensed to the UTC calendar
+    # date; unparseable values become null
+    return df.with_columns(date_parser.parse_series(series).dt.date().alias(col))
+
+
 def _apply_trim_collapse(df: pl.DataFrame, p: dict[str, Any]) -> pl.DataFrame:
     cols: list[str] = p["columns"]
     return df.with_columns(
@@ -126,8 +129,8 @@ def _apply_split_column(df: pl.DataFrame, p: dict[str, Any]) -> pl.DataFrame:
 OPS: tuple[Operation, ...] = (
     Operation(
         key="date_normalize",
-        title="Normalize dates",
-        hotkey="d",
+        title="Normalize date(t)imes",
+        hotkey="t",
         params=(
             ParamSpec(
                 "column",
@@ -138,6 +141,20 @@ OPS: tuple[Operation, ...] = (
         ),
         apply=_apply_date_normalize,
     ),
+    Operation(
+        key="date_only",
+        title="Normalize (d)ates",
+        hotkey="d",
+        params=(
+            ParamSpec(
+                "column",
+                "column",
+                "Date column",
+                help="Any input format; condenses to the UTC calendar date",
+            ),
+        ),
+        apply=_apply_date_only,
+    ),
     Operation(
         key="trim_collapse",
         title="Trim whitespace",
@@ -152,22 +169,6 @@ OPS: tuple[Operation, ...] = (
         ),
         apply=_apply_trim_collapse,
     ),
-    Operation(
-        key="dedup_rows",
-        title="Deduplicate rows",
-        hotkey="p",
-        params=(
-            ParamSpec("columns", "column_multi", "Key columns (empty = all)"),
-            ParamSpec(
-                "keep",
-                "choice",
-                "Keep",
-                default="first",
-                choices=("first", "last"),
-            ),
-        ),
-        apply=_apply_dedup_rows,
-    ),
     Operation(
         key="combine_columns",
         title="Combine columns",
@@ -212,6 +213,22 @@ OPS: tuple[Operation, ...] = (
         ),
         apply=_apply_drop_columns,
     ),
+    Operation(
+        key="dedup_rows",
+        title="Deduplicate rows",
+        hotkey="p",
+        params=(
+            ParamSpec("columns", "column_multi", "Key columns (empty = all)"),
+            ParamSpec(
+                "keep",
+                "choice",
+                "Keep",
+                default="first",
+                choices=("first", "last"),
+            ),
+        ),
+        apply=_apply_dedup_rows,
+    ),
 )
 
 OP_REGISTRY: dict[str, Operation] = {op.key: op for op in OPS}

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/screens.py

@@ -363,7 +363,11 @@ class OpParamsModal(ModalDialog):
 
     def compose_body(self) -> ComposeResult:
         self.widgets: dict[str, object] = {}
-        candidates = self._date_candidates() if self.op.key == "date_normalize" else []
+        candidates = (
+            self._date_candidates()
+            if self.op.key in ("date_normalize", "date_only")
+            else []
+        )
         for spec in self.op.params:
             if spec.kind == "column_multi":
                 sel = SelectionList(
@@ -531,8 +535,8 @@ class SaveModal(ModalDialog):
 class ScriptNameModal(ModalDialog):
     """Name the '.ntd' script file when saving a sequence of operations.
 
-    The default carries the .ntd extension; typing a different one saves
-    under that extension instead."""
+    The extension is always .ntd: whatever the user types, the saved name
+    ends in the fixed script suffix."""
 
     dialog_title = "Save script"
 
@@ -547,17 +551,14 @@ class ScriptNameModal(ModalDialog):
             id="script_input",
         )
         yield self.path_input
-        yield Static(
-            f"Default extension: {SCRIPT_SUFFIX} — type another to override",
-            classes="help",
-        )
+        yield Static(f"Scripts always end in {SCRIPT_SUFFIX}", classes="help")
 
     def action_ok(self) -> None:
         raw = self.path_input.value.strip()
         if not raw:
             self.app.notify("Type a script file name or path", severity="error")
             return
-        target = Path(raw).expanduser()
+        target = Path(raw).expanduser().with_suffix(SCRIPT_SUFFIX)
         if target.exists():
             self.app.notify("Script file exists: pick another name", severity="warning")
             return

{normalize_tabular_data-0.1.2 → normalize_tabular_data-0.1.3}/src/normalize_tabular_data/widgets.py

@@ -92,13 +92,13 @@ class ColumnSidebar(Widget):
 
 
 def _dtype_mark(dtype: str) -> str:
-    """Compact 3-wide dtype mark for the sidebar: dt (datetime/date),
-    str (string), int (integer), dec (float/decimal), t/f (boolean);
-    everything else oth. Each mark is padded to 3 columns so column
-    names left-align."""
+    """Compact 3-wide dtype mark for the sidebar: dt (datetime),
+    day (date-only), str (string), int (integer), dec (float/decimal),
+    t/f (boolean); everything else oth. Each mark is padded to 3 columns
+    so column names left-align."""
     for prefix, mark in (
         ("Datetime", "dt "),
-        ("Date", "dt "),
+        ("Date", "day"),
         ("Time", "dt "),
         ("String", "str"),
         ("Categorical", "str"),