romda 3.0.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (50) hide show
  1. romda-3.0.0/LICENSE +21 -0
  2. romda-3.0.0/PKG-INFO +197 -0
  3. romda-3.0.0/README.md +155 -0
  4. romda-3.0.0/pyproject.toml +108 -0
  5. romda-3.0.0/romda.egg-info/PKG-INFO +197 -0
  6. romda-3.0.0/romda.egg-info/SOURCES.txt +48 -0
  7. romda-3.0.0/romda.egg-info/dependency_links.txt +1 -0
  8. romda-3.0.0/romda.egg-info/requires.txt +21 -0
  9. romda-3.0.0/romda.egg-info/top_level.txt +1 -0
  10. romda-3.0.0/setup.cfg +4 -0
  11. romda-3.0.0/src/__init__.py +28 -0
  12. romda-3.0.0/src/bias_estimators/__init__.py +12 -0
  13. romda-3.0.0/src/bias_estimators/aux.py +526 -0
  14. romda-3.0.0/src/bias_estimators/bias.py +526 -0
  15. romda-3.0.0/src/bias_estimators/constantbias.py +105 -0
  16. romda-3.0.0/src/bias_estimators/esn.py +420 -0
  17. romda-3.0.0/src/data_assimilation.py +68 -0
  18. romda-3.0.0/src/estimators/__init__.py +71 -0
  19. romda-3.0.0/src/estimators/base.py +431 -0
  20. romda-3.0.0/src/estimators/deterministic.py +288 -0
  21. romda-3.0.0/src/estimators/ensembles.py +984 -0
  22. romda-3.0.0/src/estimators/inflation.py +49 -0
  23. romda-3.0.0/src/models/__init__.py +51 -0
  24. romda-3.0.0/src/models/data_driven/__init__.py +20 -0
  25. romda-3.0.0/src/models/data_driven/autoencoders/__init__.py +320 -0
  26. romda-3.0.0/src/models/data_driven/autoencoders/pod.py +492 -0
  27. romda-3.0.0/src/models/data_driven/autoencoders/pod_utils.py +576 -0
  28. romda-3.0.0/src/models/data_driven/esn.py +1212 -0
  29. romda-3.0.0/src/models/data_driven/esn_config.py +652 -0
  30. romda-3.0.0/src/models/data_driven/linear_model.py +212 -0
  31. romda-3.0.0/src/models/data_driven/pod_esn.py +581 -0
  32. romda-3.0.0/src/observations.py +778 -0
  33. romda-3.0.0/src/plotting/__init__.py +4 -0
  34. romda-3.0.0/src/plotting/annular.py +557 -0
  35. romda-3.0.0/src/plotting/cylinder.py +664 -0
  36. romda-3.0.0/src/plotting/plot_results.py +272 -0
  37. romda-3.0.0/src/plotting/pod.py +318 -0
  38. romda-3.0.0/src/plotting/utils_plotting.py +145 -0
  39. romda-3.0.0/src/utils.py +1085 -0
  40. romda-3.0.0/tests/test_bias_estimators.py +99 -0
  41. romda-3.0.0/tests/test_deterministic.py +54 -0
  42. romda-3.0.0/tests/test_estimator_config.py +89 -0
  43. romda-3.0.0/tests/test_estimator_da.py +118 -0
  44. romda-3.0.0/tests/test_history_reset_last_states.py +36 -0
  45. romda-3.0.0/tests/test_inflation.py +34 -0
  46. romda-3.0.0/tests/test_linear_model_kf.py +65 -0
  47. romda-3.0.0/tests/test_models_shim.py +27 -0
  48. romda-3.0.0/tests/test_observations.py +75 -0
  49. romda-3.0.0/tests/test_pod_esn.py +104 -0
  50. romda-3.0.0/tests/test_pod_spod.py +176 -0
romda-3.0.0/LICENSE ADDED
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2022-2026 Andrea Nóvoa
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
romda-3.0.0/PKG-INFO ADDED
@@ -0,0 +1,197 @@
1
+ Metadata-Version: 2.4
2
+ Name: romda
3
+ Version: 3.0.0
4
+ Summary: Real-time reduced-order modelling and bias-aware data assimilation
5
+ Author-email: Andrea Nóvoa <a.novoa@imperial.ac.uk>
6
+ License: MIT
7
+ Project-URL: Homepage, https://andreanovoa.github.io/real-time-bias-aware-DA/
8
+ Project-URL: Repository, https://github.com/andreanovoa/real-time-bias-aware-DA
9
+ Project-URL: Issues, https://github.com/andreanovoa/real-time-bias-aware-DA/issues
10
+ Keywords: data-assimilation,bias-estimation,real-time,ensemble-kalman-filter
11
+ Classifier: Development Status :: 4 - Beta
12
+ Classifier: Intended Audience :: Science/Research
13
+ Classifier: Programming Language :: Python :: 3
14
+ Classifier: Programming Language :: Python :: 3.10
15
+ Classifier: Programming Language :: Python :: 3.11
16
+ Classifier: Programming Language :: Python :: 3.12
17
+ Classifier: Programming Language :: Python :: 3.13
18
+ Classifier: Programming Language :: Python :: 3.14
19
+ Requires-Python: >=3.10
20
+ Description-Content-Type: text/markdown
21
+ License-File: LICENSE
22
+ Requires-Dist: dynamodels<1.0,>=0.2
23
+ Requires-Dist: echostatenetwork[opt]<1.0,>=0.1
24
+ Requires-Dist: matplotlib<4.0,>=3.9
25
+ Requires-Dist: numpy<3.0,>=2.0
26
+ Requires-Dist: pillow
27
+ Requires-Dist: pyts>=0.13
28
+ Requires-Dist: pyyaml
29
+ Requires-Dist: requests>=2.32
30
+ Requires-Dist: scipy<2.0,>=1.13
31
+ Requires-Dist: tqdm>=4.67
32
+ Requires-Dist: typeguard
33
+ Provides-Extra: dev
34
+ Requires-Dist: pytest>=8.4; extra == "dev"
35
+ Requires-Dist: ruff; extra == "dev"
36
+ Provides-Extra: notebooks
37
+ Requires-Dist: ipykernel>=6.30; extra == "notebooks"
38
+ Requires-Dist: ipywidgets>=8.1; extra == "notebooks"
39
+ Requires-Dist: jupyter>=1.1; extra == "notebooks"
40
+ Requires-Dist: nbconvert>=7.0; extra == "notebooks"
41
+ Dynamic: license-file
42
+
43
+ # romda — real-time reduced-order modelling and bias-aware data assimilation
44
+
45
+ `romda` is an open-source Python package for real-time data assimilation with
46
+ reduced-order models: ensemble Kalman filters (including the regularized bias-aware
47
+ EnKF), physical and data-driven forecast models (ESN, POD-ESN), bias estimators, and
48
+ POD/SPOD decompositions — with applications to thermoacoustics and fluid flows.
49
+
50
+ 📖 **Documentation:** https://andreanovoa.github.io/real-time-bias-aware-DA/
51
+
52
+ > This repository was formerly the `real-time-DA` package. The code used in the papers listed
53
+ > below is preserved at the release tags [v1.0](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.0), [v1.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.1) and [v2.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v2.1).
54
+
55
+
56
+ ---
57
+
58
+ ## 🚀 Getting started
59
+
60
+ 1. **Install from PyPI**
61
+ ```
62
+ pip install romda
63
+ ```
64
+
65
+ or, to work on the source, clone and install in editable mode:
66
+ ```
67
+ git clone https://github.com/andreanovoa/real-time-bias-aware-DA
68
+ cd real-time-bias-aware-DA
69
+ conda create -n romda python=3.12 && conda activate romda
70
+ pip install -e ".[dev,notebooks]" --use-pep517 # dev = pytest, notebooks = jupyter
71
+ ```
72
+
73
+ 2. (Optional) **Run the test suite**
74
+ ```
75
+ python -m pytest tests/ # unit and integration tests
76
+ python -m pytest scripts/tutorials/test_tutorials.py # execute the tutorial notebooks
77
+ ```
78
+
79
+ Quick example — bias-aware state and parameter estimation on a Van der Pol twin experiment:
80
+ ```python
81
+ import numpy as np
82
+ from romda.bias_estimators import ConstantBias
83
+ from romda.estimators import rBA_EnKF
84
+ from romda.models.physical import VdP
85
+ from romda.observations import Observations
86
+
87
+ truth = Observations(model=VdP, t_start=0.6, t_stop=0.8, Nt_obs=30,
88
+ add_noise=True, manual_bias='linear')
89
+ Cdd = np.diag((0.05 * np.max(abs(truth.y_obs), axis=0)) ** 2)
90
+
91
+ ensemble = rBA_EnKF(parent_model=VdP(dt=truth.dt), parent_bias=ConstantBias,
92
+ m=10, std_phi=0.1, std_alpha=dict(zeta=(40., 60.)))
93
+
94
+ for d, t_d in zip(truth.y_obs, truth.t_obs):
95
+ ensemble.forecast_step(t_end=t_d) # advance model (and bias) to the observation time
96
+ ensemble.analysis_step(d=d, Cdd=Cdd) # Bayesian update
97
+ ```
98
+ (The ensemble forecast runs in a multiprocessing pool, so in a script keep this under `if __name__ == "__main__":`.)
99
+
100
+ Check out the [Tutorials folder](https://github.com/andreanovoa/real-time-bias-aware-DA/tree/main/scripts/tutorials), which includes several jupyter notebooks aiming to ease the understanding of the repository.
101
+
102
+
103
+ ---
104
+
105
+
106
+ ## 🌟 What is available?
107
+ Data assimilation methods [`romda.estimators`](src/estimators/)
108
+ * EnKF — ensemble Kalman filter
109
+ * EnSRKF — ensemble square-root Kalman filter
110
+ * rBA-EnKF — regularized bias-aware EnKF
111
+ * KalmanFilter — the classical (linear) Kalman filter
112
+
113
+ Physical models [`romda.models.physical`](https://github.com/andreanovoa/dynamodels) (from the `dynamodels` package)
114
+ * Rijke tube model (dimensional with Galerkin projection)
115
+ * Van der Pol oscillator
116
+ * Lorenz 63
117
+ * Lorenz 96
118
+ * Azimuthal thermoacoustics model
119
+ * Kuramoto-Sivashinsky equation
120
+
121
+ Data-driven models [`romda.models.data_driven`](src/models/data_driven/)
122
+ * `ESN_model` — Echo State Network as a forecasting model
123
+ * `POD_ESN` — POD dimensionality reduction + ESN forecaster
124
+ * `LinearModel` — linear map with process noise (Kalman-filter tutorials)
125
+ * Projectors ([`autoencoders/`](src/models/data_driven/autoencoders/)): `POD`, `SPOD` — (Spectral) Proper Orthogonal Decomposition (Sieber 2016, Towne 2018)
126
+
127
+ Reservoir core: `EchoStateNetwork` — from the external [`echostatenetwork`](https://github.com/andreanovoa/EchoStateNetwork) package, re-exported by `romda.models.data_driven`; `pod_utils` (in `autoencoders/`) — standalone POD/SPOD algorithms
128
+
129
+ Bias estimators [`romda.bias_estimators`](src/bias_estimators/)
130
+ * Echo State Network
131
+ * Constant bias
132
+ * No bias (unbiased limit)
133
+
134
+
135
+ ---
136
+ ## 📂 Structure
137
+ ```
138
+ .
139
+ ├── data/ # Dataset files
140
+ ├── docs/ # Documentation site sources + media used in the notebooks
141
+ ├── results/ # Generated results
142
+ ├── scripts/
143
+ │ └── tutorials/ # Teaching notebooks + notebook execution tests
144
+ ├── src/ # Source code (the romda package: `from romda... import ...`)
145
+ │ ├── observations.py # Truth / observation generation
146
+ │ ├── data_assimilation.py # run_da_loop: the shared sequential DA driver
147
+ │ ├── utils.py
148
+ │ ├── estimators/ # Estimator hierarchy (the DA layer)
149
+ │ │ ├── base.py # Estimator base class (forecast_step, bias wiring)
150
+ │ │ ├── ensembles.py # EnsembleEstimator: EnKF, EnSRKF, rBA_EnKF
151
+ │ │ ├── deterministic.py # DeterministicEstimator: KalmanFilter
152
+ │ │ └── inflation.py # Covariance inflation
153
+ │ ├── bias_estimators/ # Bias model classes
154
+ │ │ ├── bias.py # Bias base class
155
+ │ │ ├── esn.py # ESN_bias
156
+ │ │ └── constantbias.py # ConstantBias, NoBias
157
+ │ ├── models/ # Model layer
158
+ │ │ ├── __init__.py # re-exports the dynamodels package (Model base,
159
+ │ │ │ # integrators, physical models) + pickle aliases
160
+ │ │ └── data_driven/ # ESN_model, POD_ESN, LinearModel, esn_config
161
+ │ │ └── autoencoders/ # Projector hierarchy: POD, SPOD (+ pod_utils)
162
+ │ └── plotting/ # Visualization helpers
163
+ ├── tests/ # Unit and integration test suite
164
+ ├── mkdocs.yml # Documentation site configuration
165
+ ├── pyproject.toml # Package setup
166
+ └── README.md # This file
167
+ ```
168
+
169
+
170
+ ---
171
+
172
+ ## 📚 Main publications from this repository
173
+
174
+ ##### Journal papers
175
+
176
+ - [x] Nóvoa, Noiray, Dawson & Magri (2024). A real-time digital twin of azimuthal thermoacoustic instabilities. Journal of Fluid Mechanics. [Published paper](https://doi.org/10.1017/jfm.2024.1052) | 🏷️ [v1.0](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.0).
177
+ - [x] Nóvoa, Racca & Magri (2023). Inferring unknown unknowns. Computer Methods in Applied Mechanics and Engineering. [Published paper](https://doi.org/10.1016/j.cma.2023.116502) | [_Legacy_ repository](https://github.com/MagriLab/rBA-EnKF).
178
+ - [x] Nóvoa & Magri (2022). Real-time thermoacoustic data assimilation. Journal of Fluid Mechanics. [Published paper](https://doi.org/10.1017/jfm.2022.653) | [_Legacy_ repository](https://github.com/MagriLab/Real-time-TA-DA).
179
+
180
+ ##### Conference papers and proceedings
181
+ - [x] Nóvoa & Magri (2025). Online model learning with data-assimilated reservoir computers. [Preprint](https://doi.org/10.48550/arXiv.2504.16767) | 🏷️ [v1.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.1).
182
+ - [x] Nóvoa & Magri (2024). Real-time digital twins of multiphysics and turbulent flows. [Paper](https://web.stanford.edu/group/ctr/ctrsp24/ii11_NOVOA.pdf).
183
+ - [x] Nóvoa & Magri (2022). Bias-aware thermoacoustic data assimilation. In_ 51st International Congress and Exposition on Noise Control Engineering. [Paper](https://az659834.vo.msecnd.net/eventsairwesteuprod/production-inconference-public/808b4f8c38f944d188db8a326a98c65c). | [_Legacy_ repository](https://github.com/MagriLab/IN22-Bias-aware-TADA).
184
+
185
+ ##### PhD thesis
186
+ - [x] Nóvoa (2024). Real-time data assimilation in nonlinear dynamcal systems. University of Cambridge. [Thesis](https://doi.org/10.17863/CAM.113001).
187
+
188
+ <!-- ##### Conference presentations _(incomplete list)_
189
+ - **APS-DFD 2024, Salt Lake City:** [Abstract](https://meetings.aps.org/Meeting/DFD24/Session/C02.14) | [Poster](https://github.com/user-attachments/files/17966063/APS-poster-final-version.pdf).
190
+ - **APS-DFD 2023, Washington DC:** [Abstract](https://meetings.aps.org/Meeting/DFD23/Session/L30.8).
191
+ - **EFMC14 2022, Athens:** [Abstract](https://euromech.org/conferences/proceedings.htm).
192
+ - **APS-DFD 2022, Phoenix:** [Abstract](https://meetings.aps.org/Meeting/DFD22/Session/G12.4). -->
193
+
194
+ ---
195
+ ## 🤝 Contributing
196
+
197
+ Contributions, bug reports, and feature requests are welcome! Please open an issue or submit a pull request. For questions or collaborations, please reach out to [A. Nóvoa](https://andreanovoa.github.io/).
romda-3.0.0/README.md ADDED
@@ -0,0 +1,155 @@
1
+ # romda — real-time reduced-order modelling and bias-aware data assimilation
2
+
3
+ `romda` is an open-source Python package for real-time data assimilation with
4
+ reduced-order models: ensemble Kalman filters (including the regularized bias-aware
5
+ EnKF), physical and data-driven forecast models (ESN, POD-ESN), bias estimators, and
6
+ POD/SPOD decompositions — with applications to thermoacoustics and fluid flows.
7
+
8
+ 📖 **Documentation:** https://andreanovoa.github.io/real-time-bias-aware-DA/
9
+
10
+ > This repository was formerly the `real-time-DA` package. The code used in the papers listed
11
+ > below is preserved at the release tags [v1.0](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.0), [v1.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.1) and [v2.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v2.1).
12
+
13
+
14
+ ---
15
+
16
+ ## 🚀 Getting started
17
+
18
+ 1. **Install from PyPI**
19
+ ```
20
+ pip install romda
21
+ ```
22
+
23
+ or, to work on the source, clone and install in editable mode:
24
+ ```
25
+ git clone https://github.com/andreanovoa/real-time-bias-aware-DA
26
+ cd real-time-bias-aware-DA
27
+ conda create -n romda python=3.12 && conda activate romda
28
+ pip install -e ".[dev,notebooks]" --use-pep517 # dev = pytest, notebooks = jupyter
29
+ ```
30
+
31
+ 2. (Optional) **Run the test suite**
32
+ ```
33
+ python -m pytest tests/ # unit and integration tests
34
+ python -m pytest scripts/tutorials/test_tutorials.py # execute the tutorial notebooks
35
+ ```
36
+
37
+ Quick example — bias-aware state and parameter estimation on a Van der Pol twin experiment:
38
+ ```python
39
+ import numpy as np
40
+ from romda.bias_estimators import ConstantBias
41
+ from romda.estimators import rBA_EnKF
42
+ from romda.models.physical import VdP
43
+ from romda.observations import Observations
44
+
45
+ truth = Observations(model=VdP, t_start=0.6, t_stop=0.8, Nt_obs=30,
46
+ add_noise=True, manual_bias='linear')
47
+ Cdd = np.diag((0.05 * np.max(abs(truth.y_obs), axis=0)) ** 2)
48
+
49
+ ensemble = rBA_EnKF(parent_model=VdP(dt=truth.dt), parent_bias=ConstantBias,
50
+ m=10, std_phi=0.1, std_alpha=dict(zeta=(40., 60.)))
51
+
52
+ for d, t_d in zip(truth.y_obs, truth.t_obs):
53
+ ensemble.forecast_step(t_end=t_d) # advance model (and bias) to the observation time
54
+ ensemble.analysis_step(d=d, Cdd=Cdd) # Bayesian update
55
+ ```
56
+ (The ensemble forecast runs in a multiprocessing pool, so in a script keep this under `if __name__ == "__main__":`.)
57
+
58
+ Check out the [Tutorials folder](https://github.com/andreanovoa/real-time-bias-aware-DA/tree/main/scripts/tutorials), which includes several jupyter notebooks aiming to ease the understanding of the repository.
59
+
60
+
61
+ ---
62
+
63
+
64
+ ## 🌟 What is available?
65
+ Data assimilation methods [`romda.estimators`](src/estimators/)
66
+ * EnKF — ensemble Kalman filter
67
+ * EnSRKF — ensemble square-root Kalman filter
68
+ * rBA-EnKF — regularized bias-aware EnKF
69
+ * KalmanFilter — the classical (linear) Kalman filter
70
+
71
+ Physical models [`romda.models.physical`](https://github.com/andreanovoa/dynamodels) (from the `dynamodels` package)
72
+ * Rijke tube model (dimensional with Galerkin projection)
73
+ * Van der Pol oscillator
74
+ * Lorenz 63
75
+ * Lorenz 96
76
+ * Azimuthal thermoacoustics model
77
+ * Kuramoto-Sivashinsky equation
78
+
79
+ Data-driven models [`romda.models.data_driven`](src/models/data_driven/)
80
+ * `ESN_model` — Echo State Network as a forecasting model
81
+ * `POD_ESN` — POD dimensionality reduction + ESN forecaster
82
+ * `LinearModel` — linear map with process noise (Kalman-filter tutorials)
83
+ * Projectors ([`autoencoders/`](src/models/data_driven/autoencoders/)): `POD`, `SPOD` — (Spectral) Proper Orthogonal Decomposition (Sieber 2016, Towne 2018)
84
+
85
+ Reservoir core: `EchoStateNetwork` — from the external [`echostatenetwork`](https://github.com/andreanovoa/EchoStateNetwork) package, re-exported by `romda.models.data_driven`; `pod_utils` (in `autoencoders/`) — standalone POD/SPOD algorithms
86
+
87
+ Bias estimators [`romda.bias_estimators`](src/bias_estimators/)
88
+ * Echo State Network
89
+ * Constant bias
90
+ * No bias (unbiased limit)
91
+
92
+
93
+ ---
94
+ ## 📂 Structure
95
+ ```
96
+ .
97
+ ├── data/ # Dataset files
98
+ ├── docs/ # Documentation site sources + media used in the notebooks
99
+ ├── results/ # Generated results
100
+ ├── scripts/
101
+ │ └── tutorials/ # Teaching notebooks + notebook execution tests
102
+ ├── src/ # Source code (the romda package: `from romda... import ...`)
103
+ │ ├── observations.py # Truth / observation generation
104
+ │ ├── data_assimilation.py # run_da_loop: the shared sequential DA driver
105
+ │ ├── utils.py
106
+ │ ├── estimators/ # Estimator hierarchy (the DA layer)
107
+ │ │ ├── base.py # Estimator base class (forecast_step, bias wiring)
108
+ │ │ ├── ensembles.py # EnsembleEstimator: EnKF, EnSRKF, rBA_EnKF
109
+ │ │ ├── deterministic.py # DeterministicEstimator: KalmanFilter
110
+ │ │ └── inflation.py # Covariance inflation
111
+ │ ├── bias_estimators/ # Bias model classes
112
+ │ │ ├── bias.py # Bias base class
113
+ │ │ ├── esn.py # ESN_bias
114
+ │ │ └── constantbias.py # ConstantBias, NoBias
115
+ │ ├── models/ # Model layer
116
+ │ │ ├── __init__.py # re-exports the dynamodels package (Model base,
117
+ │ │ │ # integrators, physical models) + pickle aliases
118
+ │ │ └── data_driven/ # ESN_model, POD_ESN, LinearModel, esn_config
119
+ │ │ └── autoencoders/ # Projector hierarchy: POD, SPOD (+ pod_utils)
120
+ │ └── plotting/ # Visualization helpers
121
+ ├── tests/ # Unit and integration test suite
122
+ ├── mkdocs.yml # Documentation site configuration
123
+ ├── pyproject.toml # Package setup
124
+ └── README.md # This file
125
+ ```
126
+
127
+
128
+ ---
129
+
130
+ ## 📚 Main publications from this repository
131
+
132
+ ##### Journal papers
133
+
134
+ - [x] Nóvoa, Noiray, Dawson & Magri (2024). A real-time digital twin of azimuthal thermoacoustic instabilities. Journal of Fluid Mechanics. [Published paper](https://doi.org/10.1017/jfm.2024.1052) | 🏷️ [v1.0](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.0).
135
+ - [x] Nóvoa, Racca & Magri (2023). Inferring unknown unknowns. Computer Methods in Applied Mechanics and Engineering. [Published paper](https://doi.org/10.1016/j.cma.2023.116502) | [_Legacy_ repository](https://github.com/MagriLab/rBA-EnKF).
136
+ - [x] Nóvoa & Magri (2022). Real-time thermoacoustic data assimilation. Journal of Fluid Mechanics. [Published paper](https://doi.org/10.1017/jfm.2022.653) | [_Legacy_ repository](https://github.com/MagriLab/Real-time-TA-DA).
137
+
138
+ ##### Conference papers and proceedings
139
+ - [x] Nóvoa & Magri (2025). Online model learning with data-assimilated reservoir computers. [Preprint](https://doi.org/10.48550/arXiv.2504.16767) | 🏷️ [v1.1](https://github.com/andreanovoa/real-time-bias-aware-DA/releases/tag/v1.1).
140
+ - [x] Nóvoa & Magri (2024). Real-time digital twins of multiphysics and turbulent flows. [Paper](https://web.stanford.edu/group/ctr/ctrsp24/ii11_NOVOA.pdf).
141
+ - [x] Nóvoa & Magri (2022). Bias-aware thermoacoustic data assimilation. In_ 51st International Congress and Exposition on Noise Control Engineering. [Paper](https://az659834.vo.msecnd.net/eventsairwesteuprod/production-inconference-public/808b4f8c38f944d188db8a326a98c65c). | [_Legacy_ repository](https://github.com/MagriLab/IN22-Bias-aware-TADA).
142
+
143
+ ##### PhD thesis
144
+ - [x] Nóvoa (2024). Real-time data assimilation in nonlinear dynamcal systems. University of Cambridge. [Thesis](https://doi.org/10.17863/CAM.113001).
145
+
146
+ <!-- ##### Conference presentations _(incomplete list)_
147
+ - **APS-DFD 2024, Salt Lake City:** [Abstract](https://meetings.aps.org/Meeting/DFD24/Session/C02.14) | [Poster](https://github.com/user-attachments/files/17966063/APS-poster-final-version.pdf).
148
+ - **APS-DFD 2023, Washington DC:** [Abstract](https://meetings.aps.org/Meeting/DFD23/Session/L30.8).
149
+ - **EFMC14 2022, Athens:** [Abstract](https://euromech.org/conferences/proceedings.htm).
150
+ - **APS-DFD 2022, Phoenix:** [Abstract](https://meetings.aps.org/Meeting/DFD22/Session/G12.4). -->
151
+
152
+ ---
153
+ ## 🤝 Contributing
154
+
155
+ Contributions, bug reports, and feature requests are welcome! Please open an issue or submit a pull request. For questions or collaborations, please reach out to [A. Nóvoa](https://andreanovoa.github.io/).
@@ -0,0 +1,108 @@
1
+ [build-system]
2
+ requires = ["setuptools>=64", "wheel"]
3
+ build-backend = "setuptools.build_meta"
4
+
5
+ [project]
6
+ name = "romda"
7
+ version = "3.0.0"
8
+ description = "Real-time reduced-order modelling and bias-aware data assimilation"
9
+ readme = "README.md"
10
+ requires-python = ">=3.10"
11
+ license = {text = "MIT"}
12
+ authors = [
13
+ {name = "Andrea Nóvoa", email = "a.novoa@imperial.ac.uk"}
14
+ ]
15
+ keywords = ["data-assimilation", "bias-estimation", "real-time", "ensemble-kalman-filter"]
16
+ classifiers = [
17
+ "Development Status :: 4 - Beta",
18
+ "Intended Audience :: Science/Research",
19
+ "Programming Language :: Python :: 3",
20
+ "Programming Language :: Python :: 3.10",
21
+ "Programming Language :: Python :: 3.11",
22
+ "Programming Language :: Python :: 3.12",
23
+ "Programming Language :: Python :: 3.13",
24
+ "Programming Language :: Python :: 3.14",
25
+ ]
26
+
27
+ dependencies = [
28
+ "dynamodels>=0.2,<1.0",
29
+ "echostatenetwork[opt]>=0.1,<1.0", # [opt] = scikit-optimize for ESN hyperparameter search
30
+ "matplotlib>=3.9,<4.0",
31
+ "numpy>=2.0,<3.0",
32
+ "pillow",
33
+ "pyts>=0.13",
34
+ "pyyaml",
35
+ "requests>=2.32",
36
+ "scipy>=1.13,<2.0",
37
+ "tqdm>=4.67",
38
+ "typeguard",
39
+ ]
40
+
41
+ [project.optional-dependencies]
42
+ dev = [
43
+ "pytest>=8.4",
44
+ "ruff",
45
+ ]
46
+ notebooks = [
47
+ "ipykernel>=6.30",
48
+ "ipywidgets>=8.1",
49
+ "jupyter>=1.1",
50
+ "nbconvert>=7.0",
51
+ ]
52
+
53
+ [project.urls]
54
+ Homepage = "https://andreanovoa.github.io/real-time-bias-aware-DA/"
55
+ Repository = "https://github.com/andreanovoa/real-time-bias-aware-DA"
56
+ Issues = "https://github.com/andreanovoa/real-time-bias-aware-DA/issues"
57
+
58
+ [tool.setuptools]
59
+ # The src directory IS the romda package (src/__init__.py is romda/__init__.py)
60
+ package-dir = {"romda" = "src"}
61
+ packages = [
62
+ "romda",
63
+ "romda.bias_estimators",
64
+ "romda.estimators",
65
+ "romda.models",
66
+ "romda.models.data_driven",
67
+ "romda.models.data_driven.autoencoders",
68
+ "romda.plotting",
69
+ ]
70
+
71
+ [tool.pytest.ini_options]
72
+ testpaths = ["tests"]
73
+ python_files = ["test_*.py", "*_test.py"]
74
+ python_classes = ["Test*"]
75
+ python_functions = ["test_*"]
76
+ addopts = [
77
+ "-v",
78
+ "--strict-markers",
79
+ ]
80
+
81
+ [tool.mypy]
82
+ python_version = "3.14.3"
83
+ warn_return_any = true
84
+ warn_unused_configs = true
85
+ disallow_untyped_defs = false
86
+ disallow_incomplete_defs = false
87
+ check_untyped_defs = true
88
+ no_implicit_optional = true
89
+ warn_redundant_casts = true
90
+ warn_unused_ignores = true
91
+
92
+ [tool.ruff]
93
+ line-length = 120
94
+ target-version = "py39"
95
+
96
+ [tool.ruff.lint]
97
+ select = [
98
+ "E", # pycodestyle errors
99
+ "W", # pycodestyle warnings
100
+ "F", # pyflakes
101
+ "I", # isort
102
+ "N", # pep8-naming
103
+ "UP", # pyupgrade
104
+ ]
105
+ ignore = [
106
+ "E501", # line too long (handled by formatter)
107
+ "N802", "N803", "N806", "N815", # DA notation: M, Af, Cdd, Nq, ... (see Notation in CLAUDE.md)
108
+ ]