pyrregular 0.2.1__tar.gz → 0.2.2__tar.gz

{pyrregular-0.2.1 → pyrregular-0.2.2}/.github/workflows/build.yml

@@ -4,15 +4,15 @@ on:
   pull_request:
     branches: [ main ]
   workflow_dispatch:
-  release:
-    types: [ published ]
+  workflow_call:
 
 jobs:
   build:
     strategy:
+      fail-fast: false
       matrix:
         os: [ubuntu-latest]
-        python-version: ["3.11"]
+        python-version: ["3.10", "3.11", "3.12", "3.13"]
     runs-on: ${{ matrix.os }}
 
     steps:
@@ -35,6 +35,11 @@ jobs:
         run: |
           pip install ".[test]"
 
-      - name: Test with pytest
+      - name: Test with pytest (offline)
         run: |
-          pytest
+          pytest -m "not network"
+
+      - name: Test with pytest (dataset downloads)
+        if: matrix.python-version == '3.11'
+        run: |
+          pytest -m network

{pyrregular-0.2.1 → pyrregular-0.2.2}/.github/workflows/python-publish.yml

@@ -8,7 +8,11 @@ permissions:
   contents: read
 
 jobs:
+  test:
+    uses: ./.github/workflows/build.yml
+
   deploy:
+    needs: test
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
@@ -33,3 +37,10 @@ jobs:
         with:
           user: __token__
           password: ${{ secrets.PYPI_API_TOKEN }}
+
+  docs:
+    needs: test
+    uses: ./.github/workflows/sphinx.yml
+    permissions:
+      contents: write
+      pages: write

{pyrregular-0.2.1 → pyrregular-0.2.2}/.github/workflows/sphinx.yml

@@ -1,9 +1,8 @@
 name: "docs"
 
 on:
-  release:
-    types: [published]
   workflow_dispatch:
+  workflow_call:
 
 permissions:
   contents: write
@@ -39,7 +38,7 @@ jobs:
         make clean html SPHINXOPTS='-a'
 
     - name: Deploy to GitHub Pages
-      if: github.ref == 'refs/heads/main'
+      if: github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/v')
       uses: peaceiris/actions-gh-pages@v3
       with:
         github_token: ${{ secrets.GITHUB_TOKEN }}

{pyrregular-0.2.1 → pyrregular-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyrregular
-Version: 0.2.1
+Version: 0.2.2
 Summary: Irregular time series made easy
 Author: Francesco Spinnato, Cristiano Landi
 Author-email: francesco.spinnato@unipi.it
@@ -8,6 +8,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 Requires-Dist: awkward

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyproject.toml

@@ -9,3 +9,4 @@ write_to = "pyrregular/_version.py"
 addopts = "-ra -q"
 testpaths = ["tests"]
 norecursedirs = ["tests/_deprecated"]
+markers = ["network: downloads datasets from Hugging Face"]

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.2.1'
-__version_tuple__ = version_tuple = (0, 2, 1)
+__version__ = version = '0.2.2'
+__version_tuple__ = version_tuple = (0, 2, 2)
 
-__commit_id__ = commit_id = 'g49dbc9f9d'
+__commit_id__ = commit_id = 'ga0f3c03bc'

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/accessor.py

@@ -2,7 +2,7 @@ import numpy as np
 import pandas as pd
 import xarray as xr
 
-from pyrregular.conversion_utils import _ak_dropnan, _fill_time_index, _reset_time_index
+from pyrregular.conversion_utils import _ak_dropnan, _reset_time_index
 from pyrregular.io_utils import save_to_file
 
 
@@ -212,7 +212,7 @@ class IrregularAccessor:
                     columns[name] = _repeat(coord.values, idx[coord.dims[0]])
         return pd.DataFrame(columns)
 
-    def to_hdf5(self, filename, compression="gzip", compression_opts=None):
+    def to_hdf5(self, filename, compression="gzip", compression_opts=1):
         save_to_file(
             data_array=self._da,
             filename=filename,

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/conversion_utils.py

@@ -134,7 +134,7 @@ def _to_tslearn(X):
     return X.swapaxes(1, 2)
 
 
-def _fill_time_index(arr):
+def _fill_time_index(arr, nan_delta=1):
     T = deepcopy(arr)
 
     # time delta
@@ -143,6 +143,12 @@ def _fill_time_index(arr):
     # mean time delta
     delta_mean = np.nanmean(a_diff, axis=2, keepdims=True)
 
+    nans = np.isnan(delta_mean)
+    if np.any(
+        nans
+    ):  # this can happen if there is only 1 valid value in the time series
+        delta_mean[nans] = nan_delta
+
     # last timestep
     last_valid_t = np.nanmax(T, axis=2, keepdims=True)
 

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/io_utils.py

@@ -159,6 +159,30 @@ def _get_metadata(
     return dates, ids, signals, n_records
 
 
+def _check_static(keys, id_name, static_names, hint):
+    """Raise if an id has more than one value for a static column.
+
+    read_csv identifies an id by the tuple (id, *static values), so a column that
+    is not really static would silently split the id into several ones.
+    """
+    first = {}
+    for key in keys:
+        other = first.setdefault(key[0], key)
+        if other != key:
+            columns = [
+                n
+                for n, a, b in zip(static_names, other[1:], key[1:])
+                if a is not b and a != b
+            ]
+            names = ", ".join(repr(c) for c in columns)
+            verb = "is" if len(columns) == 1 else "are"
+            raise ValueError(
+                f"{names} {verb} not static for {id_name} {key[0]!r} (values "
+                f"{other[1:]} and {key[1:]}). A static column must have one value "
+                f"per id. {hint}"
+            )
+
+
 def read_csv(
     filenames: str | list | dict,
     ts_id: str = "ts_id",
@@ -193,6 +217,27 @@ def read_csv(
         **kwargs,
     )
 
+    # a static column must have one value per id, otherwise the id would be split
+    _check_static(
+        ids,
+        ts_id,
+        dims["ts_id"],
+        hint='If it changes over time, move it to dims["time_id"] or use it as a '
+        "signal; if each value should be its own series, include it in the series id.",
+    )
+    _check_static(
+        signals,
+        signal_id,
+        dims["signal_id"],
+        hint="If it differs between series, move it to dims['ts_id'].",
+    )
+    _check_static(
+        dates,
+        time_id,
+        dims["time_id"],
+        hint="If it differs between series, move it to dims['ts_id'] or use it as a signal.",
+    )
+
     # date strings are parsed before sorting, so the time axis follows real time
     parsed_dates = None
     if time_index_as_datetime and all(isinstance(d[0], str) for d in dates):
@@ -289,6 +334,5 @@ def read_csv(
 
 def get_current_aoe_time():
     aoe_tz = timezone(timedelta(hours=-12))
-    current_time_utc = datetime.utcnow()
-    current_time_aoe = current_time_utc.replace(tzinfo=timezone.utc).astimezone(aoe_tz)
+    current_time_aoe = datetime.now(timezone.utc).astimezone(aoe_tz)
     return current_time_aoe.isoformat()

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/models/ncde.py

@@ -15,38 +15,10 @@ import optax
 from sklearn.base import BaseEstimator, ClassifierMixin
 from sklearn.model_selection import train_test_split
 
+from pyrregular.conversion_utils import _fill_time_index
 from pyrregular.models.nodes import _standardize
 
 
-def _fill_time_index(arr, nan_delta=1):
-    T = deepcopy(arr)
-
-    # time delta
-    a_diff = T[:, :, 1:] - T[:, :, :-1]
-
-    # mean time delta
-    delta_mean = np.nanmean(a_diff, axis=2, keepdims=True)
-
-    nans = np.isnan(delta_mean)
-    if np.any(
-        nans
-    ):  # this can happen if there is only 1 valid value in the time series
-        delta_mean[nans] = nan_delta
-
-    # last timestep
-    last_valid_t = np.nanmax(T, axis=2, keepdims=True)
-
-    # find where the nans are
-    nan_mask = np.isnan(T)
-
-    # where nans are there is an increasing value from 1 to the last nan
-    nan_indices = np.cumsum(nan_mask, axis=2)
-
-    replacement = (last_valid_t + delta_mean * nan_indices)[nan_mask]
-    T[nan_mask] = replacement
-    return T
-
-
 def _fill_all_nans_signals(X):
     X_ = deepcopy(X)
     mask = np.sum(np.isnan(X_), axis=2) >= X_.shape[2] - 2

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular/repository.py

@@ -25,7 +25,8 @@ def download_dataset_from_huggingface(
         if api_token is None:
             raise ValueError("You need to provide an API token to download the dataset")
         downloader = pooch.HTTPDownloader(
-            **dict(headers={"Authorization": f"Bearer {api_token}"})
+            progressbar=progressbar,
+            headers={"Authorization": f"Bearer {api_token}"},
         )
     else:
         downloader = pooch.HTTPDownloader(progressbar=progressbar)
@@ -35,7 +36,9 @@ def download_dataset_from_huggingface(
 def load_dataset_from_file(name, api_token=None):
     if ".h5" not in name:
         name += ".h5"
-    file = download_dataset_from_huggingface(name, api_token)
+    file = download_dataset_from_huggingface(
+        name, use_api_token=api_token is not None, api_token=api_token
+    )
     return file
 
 

{pyrregular-0.2.1 → pyrregular-0.2.2}/pyrregular.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyrregular
-Version: 0.2.1
+Version: 0.2.2
 Summary: Irregular time series made easy
 Author: Francesco Spinnato, Cristiano Landi
 Author-email: francesco.spinnato@unipi.it
@@ -8,6 +8,7 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 Requires-Dist: awkward

pyrregular-0.2.2/pyrregular.egg-info/scm_version.json

@@ -0,0 +1,8 @@
+{
+  "tag": "0.2.2",
+  "distance": 0,
+  "node": "ga0f3c03bc664933363e4a80dbd04764d680873f4",
+  "dirty": false,
+  "branch": "HEAD",
+  "node_date": "2026-10-04"
+}

{pyrregular-0.2.1 → pyrregular-0.2.2}/setup.cfg

@@ -10,6 +10,7 @@ classifiers =
 	Programming Language :: Python :: 3.10
 	Programming Language :: Python :: 3.11
 	Programming Language :: Python :: 3.12
+	Programming Language :: Python :: 3.13
 
 [options]
 packages = find:

{pyrregular-0.2.1 → pyrregular-0.2.2}/tests/test_dataset_loading.py

@@ -3,6 +3,8 @@ import pytest
 from pyrregular import load_dataset
 from tests.constants import TEST_CASES_FAST as TEST_CASES
 
+pytestmark = pytest.mark.network
+
 
 @pytest.fixture(params=TEST_CASES)
 def loaded_dataset(request):

{pyrregular-0.2.1 → pyrregular-0.2.2}/tests/test_offline.py

@@ -355,3 +355,55 @@ def test_version():
     import pyrregular
 
     assert isinstance(pyrregular.__version__, str) and pyrregular.__version__
+
+
+def test_read_csv_rejects_non_static_columns(tmp_path):
+    # label changes inside series a: it is not static, so read_csv must not split a
+    rows = [
+        ("a", "s", 1, 1.0, "x"),
+        ("a", "s", 2, 2.0, "x"),
+        ("a", "s", 3, 3.0, "y"),
+        ("b", "s", 1, 5.0, "z"),
+    ]
+    path = tmp_path / "static.csv"
+    pd.DataFrame(
+        rows, columns=["ts_id", "signal_id", "time_id", "value_id", "label"]
+    ).to_csv(path, index=False)
+    with pytest.raises(ValueError, match="'label' is not static for ts_id 'a'"):
+        read_csv(
+            path,
+            dims={"ts_id": ["label"], "signal_id": [], "time_id": []},
+            time_index_as_datetime=False,
+        )
+
+
+def test_api_token_is_sent(monkeypatch):
+    # capture the downloader instead of downloading
+    from pyrregular import repository
+
+    seen = {}
+
+    def fake_fetch(name, downloader):
+        seen["name"], seen["downloader"] = name, downloader
+        return "fake/path.h5"
+
+    monkeypatch.delenv("HF_TOKEN", raising=False)
+    monkeypatch.setattr(repository.REPOSITORY, "fetch", fake_fetch)
+    assert (
+        repository.load_dataset_from_file("Garment", api_token="hf_x") == "fake/path.h5"
+    )
+    assert seen["name"] == "Garment.h5"
+    assert seen["downloader"].kwargs["headers"] == {"Authorization": "Bearer hf_x"}
+
+
+def test_fill_time_index():
+    from pyrregular.conversion_utils import _fill_time_index
+
+    T = np.array(
+        [
+            [[0.0, 1.0, 3.0, nan, nan]],  # mean step 1.5
+            [[5.0, nan, nan, nan, nan]],  # one timestamp: no step, use nan_delta
+        ]
+    )
+    assert_equal_nan(_fill_time_index(T)[:, 0, :], [[0, 1, 3, 4.5, 6], [5, 6, 7, 8, 9]])
+    assert_equal_nan(_fill_time_index(T, nan_delta=2)[1, 0], [5, 7, 9, 11, 13])

pyrregular-0.2.1/pyrregular.egg-info/scm_version.json

@@ -1,8 +0,0 @@
-{
-  "tag": "0.2.1",
-  "distance": 0,
-  "node": "g49dbc9f9dd41bb03575d82efc1680f376c9282bc",
-  "dirty": false,
-  "branch": "HEAD",
-  "node_date": "2026-10-03"
-}