readabs 0.2.2__tar.gz → 0.2.3__tar.gz

{readabs-0.2.2 → readabs-0.2.3}/CHANGELOG.md

@@ -1,3 +1,16 @@
+Version 0.2.3 released 19-Jun-2026 (Canberra Australia)
+
+ - Pandas 3.0 compatibility: removed the deprecated `copy=` keyword from the
+   `set_axis()` call in `qtly_to_monthly()` (Copy-on-Write makes it redundant).
+   This was the only deprecated pandas usage in the package; the conversion is
+   unchanged and the caller's input series is still left untouched.
+ - Refreshed the modules' built-in self-tests for current ABS structures: the
+   labour-force (6202.0) and CPI (6401.0) table identifiers are now 8 digits, and
+   the trimmed-mean CPI example series is updated to the monthly series. The
+   local-zip test in `read_abs_cat` now skips cleanly when the fixture is absent.
+
+---
+
 Version 0.2.2 released 04-Jun-2026 (Canberra Australia)
 
  - A *selector* (in `select_one()`, `select()` and `select_and_splice()` sources)

{readabs-0.2.2 → readabs-0.2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: readabs
-Version: 0.2.2
+Version: 0.2.3
 Summary: Get ABS timeseries data in pandas DataFrames
 Project-URL: Repository, https://github.com/bpalmer4/readabs
 Project-URL: Homepage, https://github.com/bpalmer4/readabs

readabs-0.2.3/TODO.md

@@ -0,0 +1,23 @@
+# TODO
+
+## Type-checker issues (pre-existing, not runtime bugs)
+
+These are flagged by `pyright src/readabs` and/or `mypy src/readabs`. All 14
+module self-tests pass; these are type-checker friction only.
+
+- [ ] **`grab_abs_url.py` — `url` declared twice.** `grab_abs_url()` has an explicit
+  `url: str = ""` parameter *and* `**kwargs: Unpack[ReadArgs]`, where `ReadArgs`
+  also declares `url: NotRequired[str]` (`read_support.py:29`). This produces three
+  errors: the param/TypedDict overlap (line 37/40) and the `check_kwargs`/`get_args`
+  arg-type mismatches (lines 78-79). `read_support.py:33` already documents that
+  `url` is meant to be handled separately — fix by dropping `url` from `ReadArgs`.
+
+- [ ] **`grab_abs_url.py:359` — `ExcelFile.parse` not in stubs.** `excel.parse(sheet_name)`
+  works at runtime but the pandas type stubs don't expose `.parse` on `ExcelFile`.
+  Add a targeted `# type: ignore` / `# pyright: ignore`, or switch to `pd.read_excel`.
+
+- [ ] **`recalibrate.py:78,80` — `Series | DataFrame` union not narrowed by `.shape`.**
+  `result` is typed `Series | DataFrame`; the `len(data.shape) == ...` guards don't
+  narrow the runtime class, so `result.columns` (line 78) and `result.name` (line 80)
+  each flag on the type that lacks the attribute. Narrow the branches with
+  `isinstance` so both checkers agree.