updating branch (#55)

* template paper * template paper * Revert "template paper" This reverts commit 95d70f3. * Revert "template paper" This reverts commit dc59caa. * safer eval model * reverted change * added badge conda (#47) * Added conda information * clamped log_hazard to prevent torch.Inf * try fix CI builds (#53) * try fix builds * partial rollback * fix build dependency * make formatting mandatory & reformat kaplan meier test * add unreleased note * edited * adding sphinx-issues for Changelog links * reverted change --------- Co-authored-by: corolth1 <[email protected]> --------- Co-authored-by: melodiemonod <[email protected]> Co-authored-by: Peter Krusche (Novartis) <[email protected]>
Novartis · Sep 26, 2024 · 8ac3811 · 8ac3811
1 parent ba6ffb5
commit 8ac3811
Show file tree

Hide file tree

Showing 8 changed files with 44 additions and 18 deletions.
diff --git a/.github/workflows/build-release.yml b/.github/workflows/build-release.yml
@@ -27,7 +27,7 @@ jobs:
         run: python -m twine check dist/*
 
       - name: Upload artifact
-        uses: actions/upload-artifact@v2
+        uses: actions/upload-artifact@v4
         with:
           name: Python-package
           path: dist

diff --git a/.github/workflows/codeqc.yml b/.github/workflows/codeqc.yml
@@ -46,8 +46,8 @@ jobs:
       shell: bash -l {0}
       run: |
         conda activate torchsurv
-        ./dev/codeqc.sh
- 
+        ./dev/codeqc.sh check
+
     - name: Tests
       shell: bash -l {0}
       run: |

diff --git a/README.md b/README.md
@@ -2,11 +2,14 @@
 
 ![CodeQC](https://github.com/Novartis/torchsurv/actions/workflows/codeqc.yml/badge.svg?branch=main)
 ![Docs](https://github.com/Novartis/torchsurv/actions/workflows/docs.yml/badge.svg?branch=main)
-[![PyPI - Version](https://img.shields.io/pypi/v/torchsurv)](https://pypi.org/project/torchsurv/)
-[![arXiv](https://img.shields.io/badge/arXiv-2404.10761-f9f107.svg)](https://arxiv.org/abs/2404.10761)
-[![status](https://camo.githubusercontent.com/22fa65b2a659780cddfac609463c5fe719e3ea82a28eb7a61e24b7c4e40eb56d/68747470733a2f2f6a6f73732e7468656f6a2e6f72672f7061706572732f30326437343936646132623963633334663961366530346361626632323938642f7374617475732e737667)](https://joss.theoj.org/papers/02d7496da2b9cc34f9a6e04cabf2298d)
+[![PyPI - Version](https://img.shields.io/pypi/v/torchsurv?)](https://pypi.org/project/torchsurv/)
+[![Conda](https://img.shields.io/conda/v/conda-forge/torchsurv?label=conda)](https://anaconda.org/conda-forge/torchsurv)
+[![arXiv](https://img.shields.io/badge/arXiv-2404.10761-f9f107.svg?)](https://arxiv.org/abs/2404.10761)
 [![Documentation](https://img.shields.io/badge/GithubPage-Sphinx-blue)](https://opensource.nibr.com/torchsurv/)
-[![Downloads](https://static.pepy.tech/badge/torchsurv)](https://pepy.tech/project/torchsurv)
+[![PyPI Downloads](https://img.shields.io/pypi/dm/torchsurv.svg?label=PyPI%20downloads)](
+https://pypi.org/project/torchsurv/)
+[![Conda Downloads](https://img.shields.io/conda/dn/conda-forge/torchsurv.svg?label=Conda%20downloads)](
+https://anaconda.org/conda-forge/torchsurv)
 
 `TorchSurv` is a Python package that serves as a companion tool to perform deep survival modeling within the `PyTorch` environment. Unlike existing libraries that impose specific parametric forms on users, `TorchSurv` enables the use of custom `PyTorch`-based deep survival models.  With its lightweight design, minimal input requirements, full `PyTorch` backend, and freedom from restrictive survival model parameterizations, `TorchSurv` facilitates efficient survival model implementation, particularly beneficial for high-dimensional input data scenarios.
 
@@ -43,15 +46,23 @@ cindex.compare(cindexB)
 
 ## Installation and dependencies
 
-First, install the package:
 
+First, install the package using either [PyPI]([https://pypi.org/](https://pypi.org/project/torchsurv/)) or [Conda]([https://anaconda.org/anaconda/conda](https://anaconda.org/conda-forge/torchsurv))
+
+- Using conda (`recommended`)
+```bash
+conda install conda-forge::torchsurv
+```
+- Using PyPI
 ```bash
 pip install torchsurv
 ```
 
-or for local installation (from package root / clone of this git repository):
+- Using for local installation (`latest version`)
 
 ```bash
+git clone <repo>
+cd <repo>
 pip install -e .
 ```
 
@@ -237,4 +248,4 @@ If you use this project in academic work or publications, we appreciate citing i
       primaryClass={cs.LG},
       doi={https://doi.org/10.48550/arXiv.2404.10761}
 }
-```
+```
diff --git a/dev/environment.yml b/dev/environment.yml
@@ -3,7 +3,7 @@ channels:
   - conda-forge
   - pytorch
 dependencies:
-  - build=0.7.0
+  - python-build=1.2.2
   - pep517=0.13.0
   - numpy=1.26.4
   - pandas=2.2.0

diff --git a/docs/CHANGELOG.md b/docs/CHANGELOG.md
@@ -1,6 +1,16 @@
 Change log
 =========
 
+Version 0.1.3 (unreleased)
+--------------------------
+
+* Tutorial dataset error on momentum.ipynb #50
+* Fix issue #48 - log_hazard returns torch.Inf
+* Fix warning with Spearman correlation #41
+* Added in-depth statistical background to link AUC to C-index #39
+* Created Conda Forge version #47
+* Updated CICD builds #53
+
 Version 0.1.2
 -------------
 

diff --git a/src/torchsurv/loss/momentum.py b/src/torchsurv/loss/momentum.py
@@ -162,7 +162,7 @@ def forward(
                 self.memory_k.append(self.survtuple(*list(estimate)))
         return loss
 
-    @torch.no_grad()
+    @torch.no_grad()  # deactivates autograd
     def infer(self, inputs: torch.Tensor) -> torch.Tensor:
         """Evaluate data with target network
 
@@ -183,7 +183,7 @@ def infer(self, inputs: torch.Tensor) -> torch.Tensor:
                     [ 0.9771, -0.8513]])
 
         """
-        self.target.eval()  # Disable training tricks (augmentation, dropout, etc..)
+        self.target.eval()  # notify all your layers that you are in eval mode
         return self.target(inputs)
 
     def _bank_loss(self) -> torch.Tensor:

diff --git a/src/torchsurv/loss/weibull.py b/src/torchsurv/loss/weibull.py
@@ -226,6 +226,9 @@ def log_hazard(
         >>> for t in torch.tensor([100.0, 150.0]): log_hazard(log_params, time=t)  # Subject-specific log hazard at multiple new times
         tensor([ 1.1280, -0.0372, -3.9767,  1.0757])
         tensor([ 1.2330, -0.1062, -4.1680,  1.1999])
+        >>> log_params  *= 1e2  # Increase scale
+        >>> log_hazard(log_params, time, all_times = False)  # Check for Torch.Inf values
+        tensor([-1.0000e+10, -2.3197e+01, -6.8385e+01, -1.0000e+10])
     """
 
     log_scale, log_shape = _check_log_shape(log_params).unbind(1)
@@ -247,11 +250,13 @@ def log_hazard(
             f"Dimension mismatch: 'time' ({len(time)}) does not match the length of 'log_params' ({len(log_params)})."
         )
 
-    return (
+    return torch.clamp(
         log_shape
         - log_scale
         + torch.expm1(log_shape)
-        * (torch.log(torch.clip(time, 1e-100, torch.inf)) - log_scale)
+        * (torch.log(torch.clip(time, 1e-100, torch.inf)) - log_scale),
+        min=-TORCH_CLAMP_VALUE,
+        max=TORCH_CLAMP_VALUE,
     )
 
 

diff --git a/tests/test_kaplan_meier.py b/tests/test_kaplan_meier.py
@@ -208,9 +208,9 @@ def test_kaplan_meier_prediction_error_raised(self):
         for batch in batch_container.batches:
             (train_time, train_event, test_time, *_) = batch
 
-            train_event[
-                -1
-            ] = False  # if last event is censoring, the last KM is > 0 and it cannot predict beyond this time
+            train_event[-1] = (
+                False  # if last event is censoring, the last KM is > 0 and it cannot predict beyond this time
+            )
             km = KaplanMeierEstimator()
             km(train_event, train_time, censoring_dist=False)