Merge pull request #24 from Vincent-Maladiere/add_seer_preprocessing

Add SEER preprocessing
soda-inria · Dec 16, 2023 · bac9af1 · bac9af1
2 parents cd545b8 + 7ff9d74
commit bac9af1
Show file tree

Hide file tree

Showing 12 changed files with 561 additions and 12 deletions.
diff --git a/.gitignore b/.gitignore
@@ -161,5 +161,8 @@ cython_debug/
 #.idea/
 
 .vscode
-
 doc/generated/
+
+
+# This dataset should not be redistributed, because users have to sign an agreement.
+hazardous/data/seer_cancer_cardio_raw_data.txt
diff --git a/doc/_static/seer_extract.png b/doc/_static/seer_extract.png
diff --git a/doc/_static/seer_home.png b/doc/_static/seer_home.png
diff --git a/doc/api.rst b/doc/api.rst
@@ -40,6 +40,7 @@ Datasets
     :nosignatures:
 
     data.make_synthetic_competing_weibull
+    data.load_seer
 
 
 Inverse Probability Censoring Weight

diff --git a/doc/downloading_seer.rst b/doc/downloading_seer.rst
@@ -0,0 +1,46 @@
+
+.. _downloading_seer:
+
+How to get the SEER dataset
+===========================
+
+.. currentmodule:: hazardous
+
+SEER is a reference dataset for cancer statistics in the US, used in competitive risk analysis.
+
+Below is a quick guide to obtain it.
+Note that you will need a **Windows computer** (or emulator) to download the files.
+
+1. Head to the `SEER data access webpage <https://seerdataaccess.cancer.gov/seer-data-access>`_, enter your email address and click on "Research Data Requests".
+2. Fill the form.
+3. Confirm your email address and wait a few days before the confirmation.
+4. After receiving the confirmation email, go to the `SEER*Stat Software webpage <https://seer.cancer.gov/seerstat/software/>`_ to download the software on Windows.
+5. Open it and sign in with your SEER credendials received by email.
+
+.. image:: _static/seer_home.png
+  :width: 300
+
+
+.. raw:: html
+
+   <br></br>
+
+
+6. Use seerstat to open the `data/seer.sl` file.
+
+.. image:: _static/seer_extract.png
+   :width: 400
+
+
+.. raw:: html
+
+   <br></br>
+
+
+You should get a txt file. This file can be loaded using :func:`hazardous.data.load_seer`.
+
+
+.. raw:: html
+
+   <br></br>
+
diff --git a/doc/index.rst b/doc/index.rst
@@ -40,3 +40,4 @@ of this library at this time.
 
    api
    auto_examples/index
+   downloading_seer
diff --git a/hazardous/data/__init__.py b/hazardous/data/__init__.py
@@ -1,3 +1,4 @@
 from ._competing_weibull import make_synthetic_competing_weibull
+from ._seer import load_seer
 
-__all__ = ["make_synthetic_competing_weibull"]
+__all__ = ["make_synthetic_competing_weibull", "load_seer"]