dfm-python 0.6.13__tar.gz → 0.7.0__tar.gz

{dfm_python-0.6.13 → dfm_python-0.7.0}/.gitignore

@@ -215,4 +215,5 @@ Nowcasting/
 agent/
 
 *.db
-*.pkl
+*.pkl
+*.pt

dfm_python-0.7.0/CHANGELOG.md

@@ -0,0 +1,11 @@
+# Changelog
+
+## 0.7.0
+
+- The Attention Factor Model (`dfm_python.models.afm`: `AFM`, attention and PCA
+  factors, LongConv filter, Avellaneda-Lee baseline) now ships in the
+  distribution; the 0.6.14 wheel on PyPI did not include it.
+- README added and used as the package long description.
+- Metadata: project URLs point to `github.com/minkeymouse/dfm-python`.
+- `build` and `twine` moved from runtime dependencies to the `dev` extra.
+- `__version__` now matches the release; the sdist no longer bundles the example CSVs or the lock file.

dfm_python-0.7.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Minkey Chang
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

dfm_python-0.7.0/PKG-INFO

@@ -0,0 +1,186 @@
+Metadata-Version: 2.5
+Name: dfm-python
+Version: 0.7.0
+Summary: Dynamic Factor Model (DFM) estimation and forecasting in Python
+Project-URL: Homepage, https://github.com/minkeymouse/dfm-python
+Project-URL: Documentation, https://github.com/minkeymouse/dfm-python#readme
+Project-URL: Repository, https://github.com/minkeymouse/dfm-python
+Project-URL: Issues, https://github.com/minkeymouse/dfm-python/issues
+Author: Minkey Chang
+License: MIT
+License-File: LICENSE
+Keywords: dfm,dynamic-factor-model,econometrics,forecasting,time-series
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+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
+Classifier: Topic :: Scientific/Engineering
+Requires-Python: >=3.10
+Requires-Dist: einops>=0.8.1
+Requires-Dist: hydra-core>=1.3.2
+Requires-Dist: matplotlib>=3.5.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: optuna>=4.7.0
+Requires-Dist: pandas>=2.0.0
+Requires-Dist: pyarrow>=22.0.0
+Requires-Dist: pykalman>=0.11.0
+Requires-Dist: pytorch-lightning>=2.0.0
+Requires-Dist: scikit-learn>=1.7.2
+Requires-Dist: scipy>=1.10.0
+Requires-Dist: sktime>=0.40.1
+Requires-Dist: torch>=2.0.0
+Provides-Extra: all
+Requires-Dist: build>=1.4.0; extra == 'all'
+Requires-Dist: hydra-core>=1.3.0; extra == 'all'
+Requires-Dist: omegaconf>=2.3.0; extra == 'all'
+Requires-Dist: pytest-cov>=4.0.0; extra == 'all'
+Requires-Dist: pytest>=7.0.0; extra == 'all'
+Requires-Dist: scikit-learn>=1.0.0; extra == 'all'
+Requires-Dist: sktime>=0.27.0; extra == 'all'
+Requires-Dist: sqlalchemy>=2.0.0; extra == 'all'
+Requires-Dist: twine>=6.2.0; extra == 'all'
+Provides-Extra: db
+Requires-Dist: sqlalchemy>=2.0.0; extra == 'db'
+Provides-Extra: deep
+Provides-Extra: dev
+Requires-Dist: build>=1.4.0; extra == 'dev'
+Requires-Dist: pytest-cov>=4.0.0; extra == 'dev'
+Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: scikit-learn>=1.0.0; extra == 'dev'
+Requires-Dist: twine>=6.2.0; extra == 'dev'
+Provides-Extra: forecasting
+Requires-Dist: sktime[forecasting]>=0.40.1; extra == 'forecasting'
+Provides-Extra: hydra
+Requires-Dist: hydra-core>=1.3.0; extra == 'hydra'
+Requires-Dist: omegaconf>=2.3.0; extra == 'hydra'
+Provides-Extra: transform
+Requires-Dist: sktime>=0.27.0; extra == 'transform'
+Description-Content-Type: text/markdown
+
+# dfm-python
+
+Dynamic factor models in Python.
+
+- **DFM**: linear, mixed-frequency Dynamic Factor Model estimated by EM with
+  Kalman filtering and smoothing.
+- **DDFM**: Deep Dynamic Factor Model with a nonlinear (PyTorch) encoder.
+- **iVDFM**: Identifiable Variational Dynamic Factor Model.
+- **AFM**: the Attention Factor Model of Epstein, Wang, Choi and Pelger (2025),
+  a cross-sectional conditional factor model whose residuals are traded
+  portfolios, with PCA and Avellaneda-Lee benchmarks.
+
+Mixed frequencies are handled with a clock: latent factors evolve at one
+frequency (for example monthly) and lower-frequency series load on them through
+tent-kernel aggregation in the observation equation. Series faster than the
+clock are not supported.
+
+## Install
+
+```bash
+pip install dfm-python        # or: uv add dfm-python
+```
+
+Requires Python 3.10 or later; PyTorch is a required dependency.
+
+## DFM
+
+Pass preprocessed data; the model standardizes it internally.
+
+```python
+import numpy as np
+import pandas as pd
+from dfm_python import DFM, DFMDataset
+from dfm_python.config import DFMConfig
+
+rng = np.random.default_rng(0)
+cols = ["s0", "s1", "s2", "s3"]
+df = pd.DataFrame(rng.standard_normal((120, 4)), columns=cols)
+df["date"] = pd.date_range("2010-01-31", periods=120, freq="ME")
+
+config = DFMConfig(
+    blocks={"global": {"num_factors": 1, "series": cols}},
+    frequency={c: "m" for c in cols},
+    clock="m",
+    max_iter=50,
+)
+dataset = DFMDataset(config=config, data=df, time_index="date")
+model = DFM(dataset=dataset, config=config)
+model.fit()                                     # EM + Kalman smoother
+X_forecast, Z_forecast = model.predict(horizon=6)
+print(X_forecast.shape, Z_forecast.shape)       # (6, 4) series, (6, state dim)
+```
+
+`DDFM` (`from dfm_python import DDFM, DDFMDataset`) and `iVDFM`
+(`from dfm_python.models.ivdfm import iVDFM`) follow the same
+dataset-then-model pattern. Hydra YAML configurations for each model are in
+[`config/`](https://github.com/minkeymouse/dfm-python/tree/main/config) and
+end-to-end scripts in
+[`tutorial/`](https://github.com/minkeymouse/dfm-python/tree/main/tutorial).
+
+## AFM
+
+AFM takes a returns panel `(T, N)` and, for learned attention factors, firm
+characteristics `(T, N, M)`. It is trained on an after-cost Sharpe ratio plus an
+explained-variance term (`lambda_var`).
+
+```python
+import numpy as np
+import torch
+from dfm_python import AFM, AFMDataset
+from dfm_python.models.afm.factors import residuals
+
+rng = np.random.default_rng(0)
+T, N, M = 120, 20, 8                        # dates, assets, characteristics
+returns = 0.01 * rng.standard_normal((T, N))
+characteristics = rng.standard_normal((T, N, M))
+
+ds = AFMDataset(returns=returns, characteristics=characteristics)
+model = AFM(dataset=ds, n_factors=3, hist_len=20, max_epochs=20, seed=0)
+model.fit()
+
+out = model.predict()
+out["residuals"].shape                      # (1, T, N) residual returns
+out["weights"].shape                        # (1, T - hist_len, N) portfolio weights
+out["book_returns"].shape                   # (1, T - hist_len) after-cost returns
+result = model.get_result()
+print(result.sharpe, result.explained_variance)
+
+# Residuals from the fitted factor block directly
+R, X = ds.get_tensors(model.device)
+with torch.no_grad():
+    omega_f, beta_t = model.factor_block(X)  # (T, K, N), (T, N, K)
+    eps = residuals(omega_f, beta_t, R)      # (T, N)
+
+# Benchmarks: PCA factors need no characteristics
+pca = AFM(dataset=AFMDataset(returns=returns), n_factors=3, hist_len=20,
+          factor_model="pca", trading="ou")  # or trading="longconv"
+pca.fit()
+```
+
+## Layout
+
+```
+src/dfm_python/
+  models/      dfm/, ddfm/, ivdfm/, afm/ model classes
+  dataset/     DFMDataset, DDFMDataset, iVDFMDataset, AFMDataset
+  config/      model configs (DFMConfig, ..., AFMConfig), YAML adapter, results
+  ssm/         state-space form and Kalman filtering
+  functional/  EM algorithm
+  numeric/, layer/, utils/, logger/   shared numerics, network layers, helpers
+src/test/      pytest suite
+config/        example Hydra YAML configurations
+tutorial/      end-to-end scripts per model and dataset
+```
+
+Tests, from a clone: `uv sync --group dev && uv run pytest src/test -q`.
+
+The paper repository [minkeymouse/nfe](https://github.com/minkeymouse/nfe) uses
+AFM as a baseline.
+
+## License
+
+MIT. See [LICENSE](https://github.com/minkeymouse/dfm-python/blob/main/LICENSE).

dfm_python-0.7.0/README.md

@@ -0,0 +1,123 @@
+# dfm-python
+
+Dynamic factor models in Python.
+
+- **DFM**: linear, mixed-frequency Dynamic Factor Model estimated by EM with
+  Kalman filtering and smoothing.
+- **DDFM**: Deep Dynamic Factor Model with a nonlinear (PyTorch) encoder.
+- **iVDFM**: Identifiable Variational Dynamic Factor Model.
+- **AFM**: the Attention Factor Model of Epstein, Wang, Choi and Pelger (2025),
+  a cross-sectional conditional factor model whose residuals are traded
+  portfolios, with PCA and Avellaneda-Lee benchmarks.
+
+Mixed frequencies are handled with a clock: latent factors evolve at one
+frequency (for example monthly) and lower-frequency series load on them through
+tent-kernel aggregation in the observation equation. Series faster than the
+clock are not supported.
+
+## Install
+
+```bash
+pip install dfm-python        # or: uv add dfm-python
+```
+
+Requires Python 3.10 or later; PyTorch is a required dependency.
+
+## DFM
+
+Pass preprocessed data; the model standardizes it internally.
+
+```python
+import numpy as np
+import pandas as pd
+from dfm_python import DFM, DFMDataset
+from dfm_python.config import DFMConfig
+
+rng = np.random.default_rng(0)
+cols = ["s0", "s1", "s2", "s3"]
+df = pd.DataFrame(rng.standard_normal((120, 4)), columns=cols)
+df["date"] = pd.date_range("2010-01-31", periods=120, freq="ME")
+
+config = DFMConfig(
+    blocks={"global": {"num_factors": 1, "series": cols}},
+    frequency={c: "m" for c in cols},
+    clock="m",
+    max_iter=50,
+)
+dataset = DFMDataset(config=config, data=df, time_index="date")
+model = DFM(dataset=dataset, config=config)
+model.fit()                                     # EM + Kalman smoother
+X_forecast, Z_forecast = model.predict(horizon=6)
+print(X_forecast.shape, Z_forecast.shape)       # (6, 4) series, (6, state dim)
+```
+
+`DDFM` (`from dfm_python import DDFM, DDFMDataset`) and `iVDFM`
+(`from dfm_python.models.ivdfm import iVDFM`) follow the same
+dataset-then-model pattern. Hydra YAML configurations for each model are in
+[`config/`](https://github.com/minkeymouse/dfm-python/tree/main/config) and
+end-to-end scripts in
+[`tutorial/`](https://github.com/minkeymouse/dfm-python/tree/main/tutorial).
+
+## AFM
+
+AFM takes a returns panel `(T, N)` and, for learned attention factors, firm
+characteristics `(T, N, M)`. It is trained on an after-cost Sharpe ratio plus an
+explained-variance term (`lambda_var`).
+
+```python
+import numpy as np
+import torch
+from dfm_python import AFM, AFMDataset
+from dfm_python.models.afm.factors import residuals
+
+rng = np.random.default_rng(0)
+T, N, M = 120, 20, 8                        # dates, assets, characteristics
+returns = 0.01 * rng.standard_normal((T, N))
+characteristics = rng.standard_normal((T, N, M))
+
+ds = AFMDataset(returns=returns, characteristics=characteristics)
+model = AFM(dataset=ds, n_factors=3, hist_len=20, max_epochs=20, seed=0)
+model.fit()
+
+out = model.predict()
+out["residuals"].shape                      # (1, T, N) residual returns
+out["weights"].shape                        # (1, T - hist_len, N) portfolio weights
+out["book_returns"].shape                   # (1, T - hist_len) after-cost returns
+result = model.get_result()
+print(result.sharpe, result.explained_variance)
+
+# Residuals from the fitted factor block directly
+R, X = ds.get_tensors(model.device)
+with torch.no_grad():
+    omega_f, beta_t = model.factor_block(X)  # (T, K, N), (T, N, K)
+    eps = residuals(omega_f, beta_t, R)      # (T, N)
+
+# Benchmarks: PCA factors need no characteristics
+pca = AFM(dataset=AFMDataset(returns=returns), n_factors=3, hist_len=20,
+          factor_model="pca", trading="ou")  # or trading="longconv"
+pca.fit()
+```
+
+## Layout
+
+```
+src/dfm_python/
+  models/      dfm/, ddfm/, ivdfm/, afm/ model classes
+  dataset/     DFMDataset, DDFMDataset, iVDFMDataset, AFMDataset
+  config/      model configs (DFMConfig, ..., AFMConfig), YAML adapter, results
+  ssm/         state-space form and Kalman filtering
+  functional/  EM algorithm
+  numeric/, layer/, utils/, logger/   shared numerics, network layers, helpers
+src/test/      pytest suite
+config/        example Hydra YAML configurations
+tutorial/      end-to-end scripts per model and dataset
+```
+
+Tests, from a clone: `uv sync --group dev && uv run pytest src/test -q`.
+
+The paper repository [minkeymouse/nfe](https://github.com/minkeymouse/nfe) uses
+AFM as a baseline.
+
+## License
+
+MIT. See [LICENSE](https://github.com/minkeymouse/dfm-python/blob/main/LICENSE).

dfm_python-0.7.0/config/afm_finance.yaml

@@ -0,0 +1,27 @@
+# Attention Factor Model (Epstein et al., 2025) — finance / statistical arbitrage.
+# Flat AFMConfig keys; loaded via AFMConfig.from_dict(OmegaConf.to_container(cfg)).
+clock: d
+scaler: null
+
+# Factor / trading structure
+num_factors: 5
+factor_model: attention      # 'attention' (learned) | 'pca' (benchmark)
+trading: longconv            # 'longconv' (learned) | 'ou' (Avellaneda-Lee)
+
+# Architecture
+embed_dim: 32                # characteristic embedding dimension d
+hist_len: 20                 # residual-history length s for the trading filter
+n_kernels: 32                # number of long convolutions
+ridge: 0.01                  # ridge penalty for the closed-form loadings
+
+# Objective and costs (paper values: 5bps turnover, 1bp shorting)
+lambda_var: 0.1
+turnover_cost: 0.0005
+short_cost: 0.0001
+risk_free: 0.0
+
+# Training
+learning_rate: 0.001
+max_epochs: 50
+patience: 10
+seed: 0

{dfm_python-0.6.13 → dfm_python-0.7.0}/config/ddfm_exchange.yaml

@@ -9,16 +9,20 @@ encoder_layers: [16, 4]  # Hidden layers: [16] -> latent: 4 (matching original s
 activation: relu  # Matching original link='relu'
 use_batch_norm: true  # Matching original batch_norm=True
 
-# Training Parameters (matching original TensorFlow DDFM)
-learning_rate: 0.005  # Matching original learning_rate=0.005
-n_mc_samples: 10  # Number of MC samples per MCMC iteration (matching original epochs=10)
-window_size: 100  # Time-step batch size for training (matching original batch_size=100)
+# Training Parameters
+# NOTE: for fair comparison with iVDFM exchange tutorial/config, we align the
+# effective window length and training budget (see ivdfm_exchange.yaml).
+learning_rate: 0.005
+n_mc_samples: 10
+window_size: 200  # Align with iVDFM sequence_length=200
 
-# MCMC/Iterative Parameters (matching original TensorFlow DDFM)
-max_epoch: 200  # Maximum MCMC iterations (matching original max_iter=200)
-tolerance: 0.0005  # Convergence tolerance (matching original tolerance=0.0005)
-disp: 10  # Display frequency for training progress (matching original disp=10)
-seed: 3  # Random seed (matching original seed=3)
+# MCMC/Iterative Parameters
+# NOTE: DDFM uses max_epoch as an iteration budget; iVDFM uses max_epochs.
+# We set this to 10 here to make the tutorials comparable in runtime budget.
+max_epoch: 10
+tolerance: 0.0005
+disp: 10
+seed: 3
 
 # Interpolation Parameters
 interpolation_method: linear  # Interpolation method for missing values
@@ -28,16 +32,3 @@ interpolation_limit_direction: both  # Direction to fill ('forward', 'backward',
 # Data Handling (used by Dataset, not Model)
 clock: d  # Daily clock frequency (all series are daily)
 scaler: null  # No scaler - data should be pre-scaled before passing to Dataset
-
-# Frequency mapping (all series are daily)
-# If not specified, all series use clock frequency (d)
-# Format: {"series_name": "frequency_code"}
-# frequency:
-#   AUD_USD: d
-#   GBP_USD: d
-#   CAD_USD: d
-#   CHF_USD: d
-#   CNY_USD: d
-#   JPY_USD: d
-#   SGD_USD: d
-#   NZD_USD: d

{dfm_python-0.6.13 → dfm_python-0.7.0}/config/ddfm_finance.yaml

@@ -3,17 +3,24 @@
 # Clock: Monthly (m) - all series are monthly frequency
 
 # Neural Network Architecture
-encoder_layers: [64, 32, 2]  # Hidden layers: [64, 32] -> latent: 2 factors
+# Match iVDFM depth/width more closely: 3 hidden layers of size ~200
+encoder_layers: [64, 32, 5]  # Hidden layers: [200, 200] -> latent: 5 factors (match iVDFM capacity)
 activation: relu  # Activation function: 'tanh', 'relu', 'sigmoid'
 use_batch_norm: true  # Use batch normalization
+decoder_type: mlp  # Use non-linear (MLP) decoder instead of linear
+
+# Target/covariate split
+# For DDFM finance: covariates are ALL columns except the target.
+target_col: market_forward_excess_returns
+covariates: all_except_target
 
 # Training Parameters
 learning_rate: 0.001  # Learning rate for optimizer
-n_mc_samples: 1  # Number of MC samples per MCMC iteration (reduced for faster execution)
-window_size: 100  # Time-step batch size for training
+n_mc_samples: 10  # Number of MC samples per MCMC iteration
+window_size: 200  # Align with iVDFM sequence_length=200
 
 # MCMC/Iterative Parameters
-max_epoch: 200  # Maximum MCMC iterations
+max_epoch: 10  # Align with iVDFM max_epochs=10
 tolerance: 0.0005  # Convergence tolerance
 disp: 10  # Display frequency for training progress
 seed: null  # Random seed (None uses default)
@@ -27,10 +34,4 @@ interpolation_limit_direction: both  # Direction to fill ('forward', 'backward',
 clock: m  # Monthly clock frequency
 scaler: standard  # Scaler for target series: 'standard', 'robust', 'minmax', 'maxabs', 'quantile', or null
 
-# Frequency mapping (optional)
-# If not specified, all series use clock frequency
-# Format: {"series_name": "frequency_code"}
-# frequency:
-#   gdp: q
-#   unemployment: m
-#   interest_rate: m
+

dfm_python-0.7.0/config/ivdfm_exchange.yaml

@@ -0,0 +1,39 @@
+# iVDFM Configuration for Exchange Rate Dataset
+# Data: 8 exchange rate series (AUD_USD, GBP_USD, CAD_USD, CHF_USD, CNY_USD, JPY_USD, SGD_USD, NZD_USD)
+# Clock: Daily (d) - all series are daily frequency
+# Date range: 1990-01-01 to 2010-10-10
+
+# ===== Model Structure =====
+# Latent dimension
+num_factors: 2
+sequence_length: 200  # Sequence length for sliding windows
+context: null  # Column names (DataFrame) or indices (array) for context variables. If None, time-based context is generated.
+context_dim: 1  # Dimension of context. Used when context is None (generates time-based context) or when context is multivariate.
+
+# ===== Neural Network Architecture =====
+encoder_hidden_dim: 200  # Encoder hidden dimension
+encoder_n_layers: 3  # Number of encoder layers
+decoder_hidden_dim: 200  # Decoder hidden dimension
+decoder_n_layers: 3  # Number of decoder layers
+prior_hidden_dim: 100  # Prior network hidden dimension
+prior_n_layers: 2  # Number of prior network layers
+activation: lrelu  # Activation function: 'relu', 'lrelu', 'tanh', 'sigmoid'
+slope: 0.1  # Leaky ReLU slope
+
+# ===== Dynamics and Distribution Parameters =====
+factor_order: 1  # AR order for factors (p in AR(p))
+innovation_distribution: laplace  # Innovation distribution: 'laplace', 'gaussian', 'student_t'
+decoder_var: 0.01  # Decoder variance (observation noise)
+
+# ===== Training Parameters =====
+learning_rate: 0.001  # Learning rate for optimizer
+optimizer: Adam  # Optimizer type: 'Adam', 'AdamW', 'SGD'
+batch_size: 256  # Batch size for training (larger to reduce #batches/epoch)
+max_epochs: 10  # Maximum training epochs
+tolerance: 1.0e-5  # Convergence tolerance
+seed: null  # Random seed (None uses default)
+
+# ===== Base Model Config (inherited) =====
+frequency: null  # iVDFM doesn't use mixed-frequency structure
+clock: d  # Daily clock frequency (all series are daily)
+scaler: null  # Scaler for targets: 'standard', 'robust', 'minmax', 'maxabs', 'quantile', or null. Applied to targets only (not context/auxiliary variables)

dfm_python-0.7.0/config/ivdfm_finance.yaml

@@ -0,0 +1,38 @@
+# iVDFM Configuration for Finance Dataset
+# Data: Finance data with market_forward_excess_returns as target variable
+# Clock: Monthly (m) - all series are monthly frequency
+
+# ===== Model Structure =====
+# Latent dimension
+num_factors: 5
+sequence_length: 200  # Sequence length for sliding windows
+context: null  # Column names (DataFrame) or indices (array) for context variables. If None, time-based context is generated.
+context_dim: 1  # Dimension of context. Used when context is None (generates time-based context) or when context is multivariate.
+
+# ===== Neural Network Architecture =====
+encoder_hidden_dim: 50  # Encoder hidden dimension
+encoder_n_layers: 2  # Number of encoder layers
+decoder_hidden_dim: 50  # Decoder hidden dimension
+decoder_n_layers: 2  # Number of decoder layers
+prior_hidden_dim: 100  # Prior network hidden dimension
+prior_n_layers: 2  # Number of prior network layers
+activation: lrelu  # Activation function: 'relu', 'lrelu', 'tanh', 'sigmoid'
+slope: 0.1  # Leaky ReLU slope
+
+# ===== Dynamics and Distribution Parameters =====
+factor_order: 2  # AR order for factors (p in AR(p))
+innovation_distribution: laplace  # Innovation distribution: 'laplace', 'gaussian', 'student_t'
+decoder_var: 0.01  # Decoder variance (observation noise)
+
+# ===== Training Parameters =====
+learning_rate: 0.001  # Learning rate for optimizer
+optimizer: Adam  # Optimizer type: 'Adam', 'AdamW', 'SGD'
+batch_size: 128  # Batch size for training (larger to reduce #batches/epoch)
+max_epochs: 10  # Maximum training epochs
+tolerance: 1.0e-5  # Convergence tolerance
+seed: null  # Random seed (None uses default)
+
+# ===== Base Model Config (inherited) =====
+frequency: null  # iVDFM doesn't use mixed-frequency structure
+clock: m  # Monthly clock frequency
+scaler: standard  # Scaler for targets: 'standard', 'robust', 'minmax', 'maxabs', 'quantile', or null. Applied to targets only (not context/auxiliary variables)

dfm_python-0.7.0/config/ivdfm_macro.yaml

@@ -0,0 +1,40 @@
+# iVDFM Configuration for Macroeconomic Datasets
+# Based on Korean economic indicators (KO* series)
+# Data: 101 series, mixed weekly/monthly frequency (1985-04-05 to 2025-11-21)
+# Clock: Weekly (w) - base frequency for latent factors
+# Target: KOEQUIPTE (Investment, Equipment, Estimation, SA)
+
+# ===== Model Structure =====
+# Latent dimension
+num_factors: 5
+sequence_length: 200  # Sequence length for sliding windows
+context: null  # Column names (DataFrame) or indices (array) for context variables. If None, time-based context is generated.
+context_dim: 1  # Dimension of context. Used when context is None (generates time-based context) or when context is multivariate.
+
+# ===== Neural Network Architecture =====
+encoder_hidden_dim: 200  # Encoder hidden dimension
+encoder_n_layers: 3  # Number of encoder layers
+decoder_hidden_dim: 200  # Decoder hidden dimension
+decoder_n_layers: 3  # Number of decoder layers
+prior_hidden_dim: 100  # Prior network hidden dimension
+prior_n_layers: 2  # Number of prior network layers
+activation: lrelu  # Activation function: 'relu', 'lrelu', 'tanh', 'sigmoid'
+slope: 0.1  # Leaky ReLU slope
+
+# ===== Dynamics and Distribution Parameters =====
+factor_order: 3  # AR order for factors (p in AR(p))
+innovation_distribution: laplace  # Innovation distribution: 'laplace', 'gaussian', 'student_t'
+decoder_var: 0.01  # Decoder variance (observation noise)
+
+# ===== Training Parameters =====
+learning_rate: 0.001  # Learning rate for optimizer
+optimizer: Adam  # Optimizer type: 'Adam', 'AdamW', 'SGD'
+batch_size: 256  # Batch size for training (larger to reduce #batches/epoch)
+max_epochs: 10  # Maximum training epochs
+tolerance: 1.0e-5  # Convergence tolerance
+seed: null  # Random seed (None uses default)
+
+# ===== Base Model Config (inherited) =====
+frequency: null  # iVDFM doesn't use mixed-frequency structure
+clock: w  # Weekly clock frequency (base frequency for latent factors)
+scaler: standard  # Scaler for targets: 'standard', 'robust', 'minmax', 'maxabs', 'quantile', or null. Applied to targets only (not context/auxiliary variables)

{dfm_python-0.6.13 → dfm_python-0.7.0}/pyproject.toml

@@ -1,7 +1,8 @@
 [project]
 name = "dfm-python"
-version = "0.6.13"
+version = "0.7.0"
 description = "Dynamic Factor Model (DFM) estimation and forecasting in Python"
+readme = "README.md"
 requires-python = ">=3.10"
 license = {text = "MIT"}
 authors = [
@@ -16,14 +17,15 @@ classifiers = [
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
     "Topic :: Scientific/Engineering",
 ]
 dependencies = [
-    "build>=1.4.0",
     "einops>=0.8.1",
     "hydra-core>=1.3.2",
     "matplotlib>=3.5.0",
     "numpy>=1.24.0",
+    "optuna>=4.7.0",
     "pandas>=2.0.0",
     "pyarrow>=22.0.0",
     "pykalman>=0.11.0",
@@ -32,7 +34,6 @@ dependencies = [
     "scipy>=1.10.0",
     "sktime>=0.40.1",
     "torch>=2.0.0",
-    "twine>=6.2.0",
 ]
 
 [project.optional-dependencies]
@@ -58,6 +59,8 @@ dev = [
     "pytest>=7.0.0",
     "pytest-cov>=4.0.0",
     "scikit-learn>=1.0.0",
+    "build>=1.4.0",
+    "twine>=6.2.0",
 ]
 all = [
     "dfm-python[hydra,db,dev,deep,transform]",
@@ -83,7 +86,8 @@ packages = [
 exclude = [
     "test_venv",
     "test_venv/**",
-    "LICENSE",
+    "data/**",
+    "uv.lock",
 ]
 
 [dependency-groups]
@@ -92,7 +96,7 @@ dev = [
 ]
 
 [project.urls]
-Homepage = "https://github.com/yourusername/dfm-python"
-Documentation = "https://github.com/yourusername/dfm-python#readme"
-Repository = "https://github.com/yourusername/dfm-python"
-Issues = "https://github.com/yourusername/dfm-python/issues"
+Homepage = "https://github.com/minkeymouse/dfm-python"
+Documentation = "https://github.com/minkeymouse/dfm-python#readme"
+Repository = "https://github.com/minkeymouse/dfm-python"
+Issues = "https://github.com/minkeymouse/dfm-python/issues"