From 6845e2d36317eb2b605db0ebe3649fe5da8aef5d Mon Sep 17 00:00:00 2001
From: Simon Pfreundschuh <s.pfreundschuh@gmail.com>
Date: Fri, 4 Oct 2024 16:31:01 -0600
Subject: [PATCH] Include observations from noaa20.

---
 gprof_nn/data/pretraining.py   | 3 ++-
 gprof_nn/data/training_data.py | 7 +++++--
 2 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/gprof_nn/data/pretraining.py b/gprof_nn/data/pretraining.py
index fa11afe..674cc4e 100644
--- a/gprof_nn/data/pretraining.py
+++ b/gprof_nn/data/pretraining.py
@@ -29,6 +29,7 @@
 from pansat.products.satellite.gpm import (
     l1c_gpm_gmi,
     l1c_npp_atms,
+    l1c_noaa20_atms,
     l1c_gcomw1_amsr2,
     merged_ir
 )
@@ -56,7 +57,7 @@
 # pansat products for each sensor.
 PRODUCTS = {
     "gmi": (l1c_gpm_gmi,),
-    "atms": (l1c_npp_atms,),
+    "atms": (l1c_npp_atms, l1c_noaa20_atms),
     "amsr2": (l1c_gcomw1_amsr2,)
 }
 
diff --git a/gprof_nn/data/training_data.py b/gprof_nn/data/training_data.py
index 7ce5405..7fdfa0d 100644
--- a/gprof_nn/data/training_data.py
+++ b/gprof_nn/data/training_data.py
@@ -1636,14 +1636,17 @@ def worker_init_fn(self, w_id: int):
 
 
     def __len__(self):
-        return len(self.input_files)
+        return len(self.input_files) // 100
 
     def __getitem__(self, ind: int):
 
+        ind_r = ind * 100 + self.rng.integers(0, 100)
+
         try:
-            data = xr.open_dataset(self.input_files[ind])
+            data = xr.open_dataset(self.input_files[ind_r])
         except Exception:
             return self[self.rng.integers(0, len(self))]
+        ind = ind_r
 
         n_chans_in = data.input_channels.size
         n_chans_out = data.target_channels.size