awen-dlmax 0.1.0__tar.gz

awen_dlmax-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Ross Hollyman
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

awen_dlmax-0.1.0/NOTICE

@@ -0,0 +1,13 @@
+DLMAX — Dynamic Linear Models in JAX
+Copyright 2026 Ross Hollyman
+
+This product includes software developed by Ross Hollyman.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+If you use DLMAX or its methods in academic work, please cite the
+accompanying paper (see CITATION.cff).

awen_dlmax-0.1.0/PKG-INFO

@@ -0,0 +1,265 @@
+Metadata-Version: 2.4
+Name: awen-dlmax
+Version: 0.1.0
+Summary: Dynamic Linear Models in JAX
+Author-email: Ross Hollyman <hello@awenforecasting.com>
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://github.com/ross-h1/DLMAX
+Project-URL: Repository, https://github.com/ross-h1/DLMAX
+Project-URL: Issues, https://github.com/ross-h1/DLMAX/issues
+Keywords: dynamic linear models,state space models,bayesian forecasting,model averaging,time series,forecasting,jax,kalman filter
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+Requires-Dist: jax>=0.4
+Requires-Dist: numpyro>=0.15
+Requires-Dist: numpy>=1.26
+Requires-Dist: scipy>=1.11
+Requires-Dist: pandas>=2.0
+Requires-Dist: matplotlib>=3.8
+Requires-Dist: h5py>=3.16.0
+Provides-Extra: gpu
+Requires-Dist: jax[cuda12]; sys_platform == "linux" and extra == "gpu"
+Provides-Extra: xarray
+Requires-Dist: xarray>=2023.1; extra == "xarray"
+Requires-Dist: h5netcdf>=1.0; extra == "xarray"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: xarray>=2023.1; extra == "dev"
+Requires-Dist: ruff>=0.6; extra == "dev"
+Requires-Dist: jupyter; extra == "dev"
+Dynamic: license-file
+
+# DLMAX — Dynamic Linear Models in JAX
+
+DLMAX is a JAX-based framework for **Dynamic Linear Models (DLMs)** with
+**Discount Model Averaging (DMA)**. It fits and forecasts large panels of time
+series under a Bayesian state-space formulation, averaging over a universe of
+candidate models whose discount factors control how quickly each adapts. The
+whole forward filter is written in JAX, so it vectorises across series and runs
+on CPU or GPU.
+
+DLMAX is the engine behind the paper *Forecasting: Fast and Slow*. The name
+nods to the idea that a panel is best served not by one model but by a
+*universe* of fast- and slow-adapting models, combined online by their
+realised predictive performance.
+
+> **Status: alpha (v0.1.0).** The numerics are tested and stable — the
+> forecasts here are the ones behind the paper — but this is a `0.x` release and
+> the public API may change between minor versions, per
+> [semantic versioning](https://semver.org/#spec-item-4). Pin an exact version
+> if you depend on it: `pip install awen-dlmax==0.1.0`.
+>
+> Bug reports and questions are welcome via
+> [GitHub Issues](https://github.com/ross-h1/DLMAX/issues). If you hit a
+> numerical problem, the most useful report includes the model spec, the shape
+> of the panel, and whether 64-bit floats were enabled.
+
+## Features
+
+- **Component-based model builder** — compose `LocalLevel`, `LocalTrend`,
+  `Fourier` (seasonal), and `Regressors` (exogenous) terms into a DLM.
+- **Discount Model Averaging** — maintain a universe of models differing in
+  their discount factors and combine them by online predictive score.
+- **Online-learned discounts** — the default "wing" grid does not fix the
+  discount factors at all: each family carries three wingmen and the centre is
+  moved by online gradient descent, so the adaptation rate is learned per
+  series rather than chosen. This is the method of *Forecasting: Fast and Slow*.
+- **Panel-scale** — vectorised across series; length-grouped batching handles
+  unbalanced panels; optional Dask distribution across a cluster.
+- **Streaming API** — `fit` once, then `update` incrementally as new
+  observations arrive, with HDF5 persistence of filter state.
+- **Probabilistic forecasts** — Student-t predictive intervals at arbitrary
+  horizons and coverage levels.
+- **DataFrame interface** — long or wide input, accepted interchangeably
+  (see [Input format](#input-format)).
+
+## Installation
+
+DLMAX requires Python ≥ 3.11. The distribution is published as
+**`awen-dlmax`**; the import name is `DLMAX`.
+
+```bash
+pip install awen-dlmax
+```
+
+For a GPU build of JAX (Linux + CUDA 12):
+
+```bash
+pip install "awen-dlmax[gpu]"
+```
+
+To reproduce published results, pin the exact version they were produced under
+— `pip install awen-dlmax==0.1.0`.
+
+DLMAX runs in double precision; enable 64-bit JAX before importing anything
+that builds a model:
+
+```python
+import jax
+jax.config.update("jax_enable_x64", True)
+```
+
+## Input format
+
+Series are supplied as a pandas DataFrame in either of two shapes, detected
+automatically.
+
+**Long** — one row per observation, three columns:
+
+| unique_id | ds | y |
+|---|---|---|
+| series_a | 2015-01-01 | 10.4 |
+| series_a | 2015-02-01 | 11.1 |
+| series_b | 2015-01-01 | 27.3 |
+
+`unique_id` names the series, `ds` is the timestamp (or an integer index), `y`
+is the observation. Series may be of different lengths and need not share a
+calendar.
+
+**Wide** — one row per period, one column per series:
+
+| | series_a | series_b |
+|---|---|---|
+| **2015-01-01** | 10.4 | 27.3 |
+| **2015-02-01** | 11.1 | 28.0 |
+
+The index is the timestamp and the column names are the series ids. A series
+that starts late or ends early is simply `NaN` in those cells; missing values
+are carried through the filter as missing rather than imputed.
+
+Both give identical results — wide input is converted to long internally, and
+forecasts are returned keyed the way the input was.
+
+## Quick start
+
+```python
+import jax
+jax.config.update("jax_enable_x64", True)
+
+import numpy as np
+import pandas as pd
+from DLMAX import AutoFFS
+
+# A small monthly panel in long format: columns (unique_id, ds, y).
+rng = np.random.default_rng(0)
+dates = pd.date_range("2015-01-01", periods=72, freq="MS")
+frames = []
+for sid in ["series_a", "series_b", "series_c"]:
+    trend = np.linspace(0, 5, len(dates))
+    seasonal = 2.0 * np.sin(2 * np.pi * np.arange(len(dates)) / 12)
+    noise = rng.normal(0, 0.5, len(dates))
+    frames.append(pd.DataFrame({
+        "unique_id": sid,
+        "ds": dates,
+        "y": 10 + trend + seasonal + noise,
+    }))
+df = pd.concat(frames, ignore_index=True)
+
+# Fit a model universe (monthly seasonality) and forecast 12 steps ahead
+# with 80% and 95% predictive intervals.
+model = AutoFFS(season_length=12)
+model.fit(df, freq="MS")
+forecast = model.predict(h=12, level=[80, 95])
+
+print(forecast.head())
+```
+
+`predict` returns a long-format DataFrame with the point forecast and the
+requested interval bounds per `(unique_id, ds)`.
+
+### Streaming updates
+
+When the next period's observations arrive, `update` advances the filter over
+the new rows only — no refit. Every fitted series must appear, and the panel
+advances together.
+
+```python
+# next month's observation for each series
+next_ds = dates[-1] + pd.offsets.MonthBegin(1)
+df_new = pd.DataFrame({
+    "unique_id": ["series_a", "series_b", "series_c"],
+    "ds": [next_ds] * 3,
+    "y": [16.1, 15.4, 15.9],
+})
+
+model.update(df_new)            # one filter pass over the new rows only
+forecast = model.predict(h=12)  # forecast from the advanced state
+```
+
+`fit(history)` then `update(new)` gives the same answer as `fit(history + new)`
+— the filter is sequential, so how the data is fed in cannot change the result.
+
+## Key concepts
+
+- **`AutoFFS`** — the high-level estimator. Builds the model universe sized to
+  your data and exposes `fit` / `update` / `predict` for forecasting forward,
+  and `cross_validation` for rolling-origin backtests.
+- **Model universe** — the set of candidate DLMs being averaged. The default is
+  the **wing grid** (below). Pass `blocks=[...]` for a different model set, or
+  `blocks=[StaticBlock(...)]` for a fixed-discount universe.
+- **Discount factor** — governs how quickly a model forgets: 1 never forgets,
+  smaller adapts faster. The wing *learns* these. Separately, `dma_pdr` /
+  `dma_mdr` govern how fast the *averaging* re-weights toward
+  better-performing models.
+- **`AutoFFSUniverse`** — the same model with state persisted to disk. Use it
+  when state must outlive the process, the panel does not fit in memory, or
+  series arrive over time (`add_series`). It agrees with `AutoFFS` bitwise.
+
+### The wing grid, in six words
+
+These terms recur throughout the API and the source:
+
+| term | meaning |
+|---|---|
+| **grid** | the whole candidate model set for a series |
+| **family** | one structure in that set — a choice of trend damping, seasonality on/off, and error law (additive or multiplicative) |
+| **wing** | the mechanism that learns a discount rather than fixing it: a family carries trial discounts either side of a centre, and the centre moves toward whichever is scoring better |
+| **wingman** | one of those trials. Three per family: one at the centre, one either side |
+| **worker** | a single model — one family at one wingman offset. A 16-family grid has 48 workers |
+| **union DMA** | the top-level averaging, run over *every* worker of *every* block at once, rather than combining each block separately and then combining the combinations |
+
+So a forecast is: each worker filters the series, the wing moves each family's
+discount centre using how its own wingmen scored, and the union DMA weights all
+workers by recent predictive performance to produce the reported forecast.
+
+## Awkward data
+
+Real panels have series that cross zero, stop and start, or sit flat through
+their warmup window. [`docs/INVARIANTS.md`](docs/INVARIANTS.md) documents what
+DLMAX does in each case and why — missing values are carried as *no
+information* rather than imputed, the variance link is symmetric about zero,
+and the diffuse prior survives a partly-missing warmup. Worth reading if your
+panel has any of those, and before changing the core filter.
+
+## Development
+
+```bash
+git clone https://github.com/ross-h1/DLMAX.git
+cd DLMAX
+uv sync --extra dev      # or: pip install -e ".[dev]"
+pytest                   # runs the fast suite
+pytest -m slow           # also runs the slower numerical tests
+```
+
+Some tests exercise a private M1-competition dataset and are skipped unless
+`DLMAX_M1M_PATH` points at a local copy; the default suite needs no external
+data.
+
+## Citation
+
+If you use DLMAX or its methods in academic work, please cite the accompanying
+paper. See [`CITATION.cff`](CITATION.cff), or use GitHub's *Cite this
+repository* button.
+
+## License
+
+Apache License 2.0 — see [`LICENSE`](LICENSE) and [`NOTICE`](NOTICE).