deepSSF 0.2.1__tar.gz → 0.2.2__tar.gz

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

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: deepSSF
-Version: 0.2.1
+Version: 0.2.2
 Summary: Deep learning step selection functions for predicting animal movement.
 Project-URL: Homepage, https://swforrest.github.io/deepSSF/
 Project-URL: Repository, https://github.com/swforrest/deepssf
@@ -83,13 +83,18 @@ print(deepssf.__version__)
 If you are coming from R, think of a conda environment the way you think of
 an `renv` project library — it is a self-contained Python installation that
 keeps this project's packages separate from everything else on your computer.
-The steps below create one for deepssf and should take about five minutes.
+The steps below create one for deepSSF and should take about five minutes.
 
 ### 1. Install Miniconda (once, system-wide)
 
 Download and run the installer from the
 [official Miniconda page](https://docs.anaconda.com/miniconda/).
 
+If you click 'Download' towards the top right, the links to download 
+Miniconda are towards the bottom of the page - the links at the top are
+for the Anaconda Distribution, which has thousands of pacakges and is
+not necessary to get things up and running initially.
+
 - **Windows**: use the **Anaconda Prompt** for all subsequent commands, and
   choose an install path that contains **no spaces** (e.g. `C:\miniconda3`).
 - **macOS / Linux**: a normal terminal works fine.
@@ -101,7 +106,7 @@ Download and run the installer from the
 ### 2. Create the environment
 
 ```bash
-git clone https://github.com/swforrest/deepssf
+git clone https://github.com/swforrest/deepssf_package
 cd deepssf
 conda env create -f environment.yml
 ```

{deepssf-0.2.1 → deepssf-0.2.2}/README.md

@@ -44,13 +44,18 @@ print(deepssf.__version__)
 If you are coming from R, think of a conda environment the way you think of
 an `renv` project library — it is a self-contained Python installation that
 keeps this project's packages separate from everything else on your computer.
-The steps below create one for deepssf and should take about five minutes.
+The steps below create one for deepSSF and should take about five minutes.
 
 ### 1. Install Miniconda (once, system-wide)
 
 Download and run the installer from the
 [official Miniconda page](https://docs.anaconda.com/miniconda/).
 
+If you click 'Download' towards the top right, the links to download 
+Miniconda are towards the bottom of the page - the links at the top are
+for the Anaconda Distribution, which has thousands of pacakges and is
+not necessary to get things up and running initially.
+
 - **Windows**: use the **Anaconda Prompt** for all subsequent commands, and
   choose an install path that contains **no spaces** (e.g. `C:\miniconda3`).
 - **macOS / Linux**: a normal terminal works fine.
@@ -62,7 +67,7 @@ Download and run the installer from the
 ### 2. Create the environment
 
 ```bash
-git clone https://github.com/swforrest/deepssf
+git clone https://github.com/swforrest/deepssf_package
 cd deepssf
 conda env create -f environment.yml
 ```

{deepssf-0.2.1 → deepssf-0.2.2}/src/deepssf/__init__.py

@@ -7,7 +7,7 @@ expose the things you want people to rely on. Internal helpers stay private.
 
 # Single source of truth for the version. hatchling reads this string at build
 # time (see pyproject.toml [tool.hatch.version]). Bump it when you release.
-__version__ = "0.2.1"
+__version__ = "0.2.2"
 
 from deepssf.data import (
     MovementDataset,

{deepssf-0.2.1 → deepssf-0.2.2}/tests/test_deepssf.py

@@ -371,7 +371,7 @@ def test_make_simulation_inputs_hour_wraps():
 
 
 def test_simulate_next_step_returns_coords_and_tensors(small_params):
-    """simulate_next_step returns new coordinates and three log-prob tensors."""
+    """simulate_next_step returns new coordinates."""
     import math
 
     import rasterio.transform
@@ -396,15 +396,12 @@ def test_simulate_next_step_returns_coords_and_tensors(small_params):
     scalars = torch.zeros(1, 4)
     bearing = torch.zeros(1, 1)
 
-    new_x, new_y, hab, move, step, px, py = simulate_next_step(
+    new_x, new_y, px, py = simulate_next_step(
         model, rasters, scalars, bearing, window_size=W,
         x_loc=W * 25 / 2, y_loc=W * 25 / 2, transform=transform,
     )
     assert isinstance(new_x, float)
     assert isinstance(new_y, float)
-    assert hab.shape == (W, W)
-    assert move.shape == (W, W)
-    assert step.shape == (W, W)
     assert 0 <= px < W
     assert 0 <= py < W
 
@@ -676,11 +673,19 @@ def test_prepare_movement_df_has_dx_dy():
 
     assert "dx" in df.columns
     assert "dy" in df.columns
-    assert np.isfinite(df["dx"].values).all()
-    assert np.isfinite(df["dy"].values).all()
-    # dx = x2_ - x1_, dy = y2_ - y1_
-    np.testing.assert_allclose(df["dx"].values, df["x2_"].values - df["x1_"].values)
-    np.testing.assert_allclose(df["dy"].values, df["y2_"].values - df["y1_"].values)
+
+    dx = df["dx"].to_numpy(dtype=float)
+    dy = df["dy"].to_numpy(dtype=float)
+    x1 = df["x1_"].to_numpy(dtype=float)
+    x2 = df["x2_"].to_numpy(dtype=float)
+    y1 = df["y1_"].to_numpy(dtype=float)
+    y2 = df["y2_"].to_numpy(dtype=float)
+
+    assert np.isfinite(dx).all()
+    assert np.isfinite(dy).all()
+
+    np.testing.assert_allclose(dx, x2 - x1)
+    np.testing.assert_allclose(dy, y2 - y1)
 
 
 def test_filter_steps_by_window_removes_large_steps():
@@ -775,7 +780,7 @@ def test_fit_returns_loss_history(small_params):
     optimisers, schedulers = make_optimisers(model)
     history = fit(
         model,
-        n_conv_layers=3,       # number of conv layers (used only for snapshots)
+        image_trim_pixels=3,       # number of conv layers (used only for snapshots)
         window_size=H,         # spatial crop size (used only for snapshots)
         dl_train=_DL(),
         dl_val=_DL(),