Docs: Add a Getting Started page.

Author: benjaminrose

docs/source/getting-started.rst

@@ -0,0 +1,25 @@
+Getting Started
+===============
+
+Written Mar 16, 2021
+
+Environment
+-----------
+
+Previous UNITY documentation use non-``conda`` environments. The scientific community is standardizing on ``conda`` and so should UNITY.
+
+Create a UNITY ``conda`` environment by use unity.yml (found in the root directory of the UNITY project): ``conda env create -f unity.yaml``.
+
+Activate the environnement, then add the UNITY cli, by running ``pip install .`` in repo's top level directory.
+
+
+Test that things are working
+----------------------------
+
+To ensure things are working correctly. Run ``simple_test_dataset.py`` (again found in ``docs/source/``) to create the ``test_simple_300_obs.pkl`` test data set. Run this through the UNITY cli with ``unity run --model=stan_code_simple_debug.txt --steps=100 --max_cores=1 ./path-to/test_simple_300_obs.pkl``.
+
+
+Your data
+---------
+
+I recommend adapting ``./simple_test_dataset.py`` to read in your SALT2 fits and redshifts. A more complete description of each parameter of the data pickle file can be found at ./docs_input_data.rst.

docs/source/simple_test_dataset.py

@@ -0,0 +1,76 @@
+"""simple_test_dataset.py -- Create a simple data set to test installation.
+
+Basic SALT/Tripp standardization.
+"""
+import pickle
+
+import numpy as np
+
+np.random.seed(13048293)
+
+DATA_NAME = "simple"  # default
+N_SNE = 300
+# Tripp Coefficients
+COFF = [
+    -0.14,  # alpha, note negative sign
+    3,  # beta
+]
+
+
+# TRUE VALUES
+c_true = np.random.randn(N_SNE) * 0.1
+x1_true = np.random.randn(N_SNE)
+mb_true = COFF[0] * x1_true + COFF[1] * c_true - 20
+
+
+# OBSERVATIONAL
+x1_obs = x1_true + np.random.randn(N_SNE) * 0.3
+c_obs = c_true + np.random.randn(N_SNE) * 0.04
+mb_obs = mb_true + np.random.randn(N_SNE) * 0.10
+
+# Don't use non-Gaussian parameters
+age_gaus_mean = np.zeros((0, N_SNE, 0))
+age_gaus_std = np.zeros((0, N_SNE, 0))
+age_gaus_A = np.zeros((0, N_SNE, 0))
+
+
+pickle.dump(
+    dict(  # general properties
+        n_sne=N_SNE,
+        n_props=3,
+        n_non_gaus_props=0,
+        n_sn_set=1,
+        sn_set_inds=[0] * N_SNE,
+        # redshifts
+        z_helio=[0.1] * N_SNE,
+        z_CMB=[0.1] * N_SNE,
+        # Gaussian defined properties
+        obs_mBx1c=[[mb_obs[i], x1_obs[i], c_obs[i]] for i in range(N_SNE)],
+        obs_mBx1c_cov=[
+            np.diag(
+                [
+                    0.05 ** 2,
+                    0.3 ** 2,
+                    0.04 ** 2,
+                ]
+            )
+        ]
+        * N_SNE,
+        # Non-Gaussian properties, skip
+        n_age_mix=0,
+        age_gaus_mean=age_gaus_mean,
+        age_gaus_std=age_gaus_std,
+        age_gaus_A=age_gaus_A,
+        # Other stuff that does not really need to change
+        do_fullDint=0,
+        outl_frac_prior_lnmean=-4.6,
+        outl_frac_prior_lnwidth=1,
+        lognormal_intr_prior=0,
+        allow_alpha_S_N=0,
+    ),
+    open(f"test_{DATA_NAME}_{N_SNE}_obs.pkl", "wb"),
+)
+pickle.dump(
+    {"x1": x1_true, "c": c_true, "mass": mass_true, "age": age_true, "mb": mb_true},
+    open(f"test_{DATA_NAME}_{N_SNE}_true.pkl", "wb"),
+)