From da06f12f533e01041803ab412f3130af3784aeb1 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Fri, 19 May 2023 21:27:43 -0300
Subject: [PATCH 01/26] begin adding files

---
 flood_forecast/interpretability.py | 13 +++++++++++++
 requirements.txt                   |  1 +
 2 files changed, 14 insertions(+)
 create mode 100644 flood_forecast/interpretability.py

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
new file mode 100644
index 000000000..ac20cd94e
--- /dev/null
+++ b/flood_forecast/interpretability.py
@@ -0,0 +1,13 @@
+from captum.attr import IntegratedGradients
+
+
+def run_attribution(model, test_loader, method):
+    """
+    """
+    ig = IntegratedGradients(model)
+    x, y = test_loader[0]
+    attributions, approximation_error = ig.attribute((input1, input2),
+                                                 baselines=(baseline1, baseline2),
+                                                 method='gausslegendre',
+                                                 return_convergence_delta=True)
+    pass
diff --git a/requirements.txt b/requirements.txt
index 292bc9c46..b25be39e7 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,6 @@
 shap==0.41.0
 scikit-learn>=1.0.1
+captum
 pandas
 torch
 tb-nightly

From c2d442e5d081f949111005d11712dc6858e498d5 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Fri, 19 May 2023 21:37:15 -0300
Subject: [PATCH 02/26] fixing code 2

---
 flood_forecast/interpretability.py | 10 ++++++----
 requirements.txt                   |  2 +-
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index ac20cd94e..fbbcf646b 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,13 +1,15 @@
 from captum.attr import IntegratedGradients
+from typing import Tuple
 
+attr_dict = {}
 
-def run_attribution(model, test_loader, method):
+def run_attribution(model, test_loader, method) -> Tuple:
     """
     """
     ig = IntegratedGradients(model)
     x, y = test_loader[0]
-    attributions, approximation_error = ig.attribute((input1, input2),
-                                                 baselines=(baseline1, baseline2),
+    attributions, approximation_error = ig.attribute(x,
+                                                 baselines=(baseline1),
                                                  method='gausslegendre',
                                                  return_convergence_delta=True)
-    pass
+    return attributions, approximation_error
diff --git a/requirements.txt b/requirements.txt
index b25be39e7..176359afe 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -7,7 +7,7 @@ tb-nightly
 seaborn
 future
 h5py
-wandb==0.15.1
+wandb==0.15.2
 google-cloud
 google-cloud-storage
 plotly~=5.14.1

From 0f3ddffa4f72eb4c506bcdb7579c699e93307012 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 13:48:53 -0300
Subject: [PATCH 03/26] adding fixes

---
 flood_forecast/interpretability.py            | 29 ++++++++++++-------
 flood_forecast/transformer_xl/cross_former.py |  2 +-
 2 files changed, 20 insertions(+), 11 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index fbbcf646b..96cc5f978 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,15 +1,24 @@
-from captum.attr import IntegratedGradients
-from typing import Tuple
+from captum.attr import IntegratedGradients, DeepLift, GradientShap, NoiseTunnel, FeatureAblation
+from typing import Tuple, Dict
 
-attr_dict = {}
+attr_dict = {"IntegratedGradients": IntegratedGradients, "DeepLift": DeepLift, "GradientSHAP": GradientShap,
+             "NoiseTunnel": NoiseTunnel, "FeatureAblation": FeatureAblation}
 
-def run_attribution(model, test_loader, method) -> Tuple:
-    """
+
+def run_attribution(model, test_loader, method, additional_params: Dict) -> Tuple:
+    """Function that creates attribution for a model based on Captum.
+
+    :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
+    :type model: _type_
+    :param test_loader: _description_
+    :type test_loader: _type_
+    :param method: _description_
+    :type method: _type_
+    :return: d
+    :rtype: Tuple
     """
-    ig = IntegratedGradients(model)
+
+    attribution_method = attr_dict[method](model)
     x, y = test_loader[0]
-    attributions, approximation_error = ig.attribute(x,
-                                                 baselines=(baseline1),
-                                                 method='gausslegendre',
-                                                 return_convergence_delta=True)
+    attributions, approximation_error = attribution_method.attribute(x.unsqueeze(0), **additional_params)
     return attributions, approximation_error
diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index c7aa42d47..bba0923fd 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -36,7 +36,7 @@ def __init__(
         :type win_size: int, optional
         :param factor: _description_, defaults to 10
         :type factor: int, optional
-        :param d_model: _description_, defaults to 512
+        :param d_model: _description_, sdefaults to 512
         :type d_model: int, optional
         :param d_ff: _description_, defaults to 1024
         :type d_ff: int, optional

From 726e07737865f3c27a4e4242a25158519dcbd69b Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 14:26:53 -0300
Subject: [PATCH 04/26] fixing captum 4 5

---
 .circleci/config.yml |  1 +
 tests/test_captum.py | 24 ++++++++++++++++++++++++
 2 files changed, 25 insertions(+)
 create mode 100644 tests/test_captum.py

diff --git a/.circleci/config.yml b/.circleci/config.yml
index 93fbda8c9..27a9c6ee9 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -65,6 +65,7 @@ jobs:
           name: Evaluator tests
           when: always
           command: |
+            coverage run -m unitest -v tests/test_captum.py
             coverage run -m unittest -v tests/test_deployment.py
             coverage run -m unittest -v tests/test_evaluation.py
             coverage run -m unittest -v tests/validation_loop_test.py
diff --git a/tests/test_captum.py b/tests/test_captum.py
new file mode 100644
index 000000000..7c390385c
--- /dev/null
+++ b/tests/test_captum.py
@@ -0,0 +1,24 @@
+import unittest
+import torch
+from flood_forecast.interpretability import run_attribution, create_attribution_plots
+from flood_forecast.basic.gru_vanilla import GRUVanilla
+from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
+
+
+class TestAttention(unittest.TestCase):
+    def setUp(self):
+        # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
+        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
+        self.test_data_loader = CSVDataLoader(
+            "tests/data/test_data.csv",
+            100,
+            20,
+            "precip",
+            ["precip", "cfs", "temp"]
+        )
+
+    def test_run_attribution(self):
+        """_summary_"""
+        attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
+        self.assertEqual(approx_error.shape, torch.Size([1, 20, 3]))
+        self.assertEqual(attributions.shape, torch.Size([1, 20, 3]))

From 127633f90c878b2faae29d1dfef27407a2870201 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 14:45:49 -0300
Subject: [PATCH 05/26] stfu linter

---
 flood_forecast/interpretability.py | 13 +++++++++++++
 tests/test_captum.py               |  5 +++++
 2 files changed, 18 insertions(+)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 96cc5f978..f7d6c015e 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -22,3 +22,16 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     x, y = test_loader[0]
     attributions, approximation_error = attribution_method.attribute(x.unsqueeze(0), **additional_params)
     return attributions, approximation_error
+
+
+def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
+    """_summary_
+
+    :param attributions: _description_
+    :type attributions: _type_
+    :param approximation_error: _description_
+    :type approximation_error: _type_
+    :param use_wandb: _description_, defaults to True
+    :type use_wandb: bool, optional
+    """
+    pass
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 7c390385c..8cd448e28 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -22,3 +22,8 @@ def test_run_attribution(self):
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
         self.assertEqual(approx_error.shape, torch.Size([1, 20, 3]))
         self.assertEqual(attributions.shape, torch.Size([1, 20, 3]))
+
+    def test_create_attribution_plots(self):
+        """_summary_"""
+        attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
+        create_attribution_plots(attributions, approx_error, use_wandb=False)
\ No newline at end of file

From 5e1aebd79c160c60780b7ac8ec3c1e4f07b08601 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 14:56:43 -0300
Subject: [PATCH 06/26] fixing code 4.5

---
 flood_forecast/interpretability.py | 2 +-
 tests/test_captum.py               | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index f7d6c015e..cb42391f2 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -10,7 +10,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
     :type model: _type_
-    :param test_loader: _description_
+    :param test_loader: Should be a FF CSVDataLoader or a related subclass.
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 8cd448e28..e235a4f41 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -26,4 +26,4 @@ def test_run_attribution(self):
     def test_create_attribution_plots(self):
         """_summary_"""
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
-        create_attribution_plots(attributions, approx_error, use_wandb=False)
\ No newline at end of file
+        create_attribution_plots(attributions, approx_error, use_wandb=False)

From d12e5c412707917fc8a80345e767ce8658a4103d Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 15:07:02 -0300
Subject: [PATCH 07/26] sleeplier than sleepy joe

---
 flood_forecast/transformer_xl/cross_former.py | 2 +-
 tests/test_captum.py                          | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index bba0923fd..adedfbc84 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -137,7 +137,7 @@ class SegMerging(nn.Module):
     we set win_size = 2 in our paper
     """
 
-    def __init__(self, d_model, win_size, norm_layer=nn.LayerNorm):
+    def __init__(self, d_model: int, win_size, norm_layer=nn.LayerNorm):
         super().__init__()
         self.d_model = d_model
         self.win_size = win_size
diff --git a/tests/test_captum.py b/tests/test_captum.py
index e235a4f41..617048305 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -5,7 +5,7 @@
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
-class TestAttention(unittest.TestCase):
+class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
         self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)

From c3d2707e4c2515eff82d669a3be4045c08f0bfde Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 15:07:14 -0300
Subject: [PATCH 08/26] r

---
 .circleci/config.yml                          | 2 +-
 flood_forecast/transformer_xl/cross_former.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/.circleci/config.yml b/.circleci/config.yml
index 27a9c6ee9..181c8ce85 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -65,7 +65,7 @@ jobs:
           name: Evaluator tests
           when: always
           command: |
-            coverage run -m unitest -v tests/test_captum.py
+            coverage run -m unittest -v tests/test_captum.py
             coverage run -m unittest -v tests/test_deployment.py
             coverage run -m unittest -v tests/test_evaluation.py
             coverage run -m unittest -v tests/validation_loop_test.py
diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index adedfbc84..f1b4f046f 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -144,7 +144,7 @@ def __init__(self, d_model: int, win_size, norm_layer=nn.LayerNorm):
         self.linear_trans = nn.Linear(win_size * d_model, d_model)
         self.norm = norm_layer(win_size * d_model)
 
-    def forward(self, x):
+    def forward(self, x: torch.Tensor):
         """
         x: B, ts_d, L, d_model
         """

From c8370a852fcd7e72c3254903304efaf84060cf90 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 15:07:46 -0300
Subject: [PATCH 09/26] Revert "r"

This reverts commit c3d2707e4c2515eff82d669a3be4045c08f0bfde.
---
 .circleci/config.yml                          | 2 +-
 flood_forecast/transformer_xl/cross_former.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/.circleci/config.yml b/.circleci/config.yml
index 181c8ce85..27a9c6ee9 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -65,7 +65,7 @@ jobs:
           name: Evaluator tests
           when: always
           command: |
-            coverage run -m unittest -v tests/test_captum.py
+            coverage run -m unitest -v tests/test_captum.py
             coverage run -m unittest -v tests/test_deployment.py
             coverage run -m unittest -v tests/test_evaluation.py
             coverage run -m unittest -v tests/validation_loop_test.py
diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index f1b4f046f..adedfbc84 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -144,7 +144,7 @@ def __init__(self, d_model: int, win_size, norm_layer=nn.LayerNorm):
         self.linear_trans = nn.Linear(win_size * d_model, d_model)
         self.norm = norm_layer(win_size * d_model)
 
-    def forward(self, x: torch.Tensor):
+    def forward(self, x):
         """
         x: B, ts_d, L, d_model
         """

From 770f20e208c4f7f81eb9b9f2d11d382aaa2a6cc5 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 15:39:56 -0300
Subject: [PATCH 10/26] refix config file 2

---
 .circleci/config.yml                          | 3 +--
 flood_forecast/transformer_xl/cross_former.py | 2 +-
 2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/.circleci/config.yml b/.circleci/config.yml
index 27a9c6ee9..a1bb37fe3 100644
--- a/.circleci/config.yml
+++ b/.circleci/config.yml
@@ -65,12 +65,11 @@ jobs:
           name: Evaluator tests
           when: always
           command: |
-            coverage run -m unitest -v tests/test_captum.py
+            coverage run -m unittest -v tests/test_captum.py
             coverage run -m unittest -v tests/test_deployment.py
             coverage run -m unittest -v tests/test_evaluation.py
             coverage run -m unittest -v tests/validation_loop_test.py
             coverage run -m unittest -v tests/test_handle_multi_crit.py
-
       - run:
           name: upload
           when: always
diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index adedfbc84..52c344283 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -40,7 +40,7 @@ def __init__(
         :type d_model: int, optional
         :param d_ff: _description_, defaults to 1024
         :type d_ff: int, optional
-        :param n_heads: _description_, defaults to 8
+        :param n_heads: The number of heads, defaults to 8
         :type n_heads: int, optional
         :param e_layers: _description_, defaults to 3
         :type e_layers: int, optional

From c4bd4b86b3952dd66fa90ef09c803f575c1d7b79 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 15:49:04 -0300
Subject: [PATCH 11/26] remove unused function

---
 flood_forecast/evaluator.py | 2 +-
 tests/test_captum.py        | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/flood_forecast/evaluator.py b/flood_forecast/evaluator.py
index 5f7e10a69..ad954c2de 100644
--- a/flood_forecast/evaluator.py
+++ b/flood_forecast/evaluator.py
@@ -175,7 +175,7 @@ def evaluate_model(
             idx += 1
             eval_log[target + "_" + evaluation_metric.__class__.__name__] = s
 
-    # Explain model behaviour using shap
+    # Explain the model behaviour using shap
     if "probabilistic" in inference_params:
         print("Probabilistic explainability currently not supported.")
     elif "n_targets" in model.params:
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 617048305..30ffb031e 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,6 +1,6 @@
 import unittest
 import torch
-from flood_forecast.interpretability import run_attribution, create_attribution_plots
+from flood_forecast.interpretability import run_attribution, make_attribution_plots
 from flood_forecast.basic.gru_vanilla import GRUVanilla
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
@@ -26,4 +26,4 @@ def test_run_attribution(self):
     def test_create_attribution_plots(self):
         """_summary_"""
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
-        create_attribution_plots(attributions, approx_error, use_wandb=False)
+        make_attribution_plots(attributions, approx_error, use_wandb=False)

From bd8061f6e9dbc58b2ae120efab3f85a8e7f427a0 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 17:34:39 -0300
Subject: [PATCH 12/26] fixing captum tests

---
 flood_forecast/interpretability.py | 9 ++++++---
 tests/test_captum.py               | 4 ++--
 2 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index cb42391f2..72b7accf6 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,5 +1,6 @@
 from captum.attr import IntegratedGradients, DeepLift, GradientShap, NoiseTunnel, FeatureAblation
 from typing import Tuple, Dict
+import numpy as np
 
 attr_dict = {"IntegratedGradients": IntegratedGradients, "DeepLift": DeepLift, "GradientSHAP": GradientShap,
              "NoiseTunnel": NoiseTunnel, "FeatureAblation": FeatureAblation}
@@ -14,7 +15,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
-    :return: d
+    :return: Returns a Tuple of attributions and approximation error. This data is used to create plots.
     :rtype: Tuple
     """
 
@@ -24,7 +25,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     return attributions, approximation_error
 
 
-def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
+def make_attribution_plots(x, attributions, approximation_error, feature_names, use_wandb: bool = True):
     """_summary_
 
     :param attributions: _description_
@@ -34,4 +35,6 @@ def make_attribution_plots(attributions, approximation_error, use_wandb: bool =
     :param use_wandb: _description_, defaults to True
     :type use_wandb: bool, optional
     """
-    pass
+    x_axis_data = np.arange(x.shape[1])
+    print(x_axis_data)
+    print("Hello world")
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 30ffb031e..692d2d98f 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,14 +1,14 @@
 import unittest
 import torch
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import GRUVanilla
+from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
+        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
         self.test_data_loader = CSVDataLoader(
             "tests/data/test_data.csv",
             100,

From ea3b2f1795b52785cb8b66dc4168bcccacbae511 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Wed, 24 May 2023 18:14:59 -0300
Subject: [PATCH 13/26] Revert "fixing captum tests"

This reverts commit bd8061f6e9dbc58b2ae120efab3f85a8e7f427a0.
---
 flood_forecast/interpretability.py | 9 +++------
 tests/test_captum.py               | 4 ++--
 2 files changed, 5 insertions(+), 8 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 72b7accf6..cb42391f2 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,6 +1,5 @@
 from captum.attr import IntegratedGradients, DeepLift, GradientShap, NoiseTunnel, FeatureAblation
 from typing import Tuple, Dict
-import numpy as np
 
 attr_dict = {"IntegratedGradients": IntegratedGradients, "DeepLift": DeepLift, "GradientSHAP": GradientShap,
              "NoiseTunnel": NoiseTunnel, "FeatureAblation": FeatureAblation}
@@ -15,7 +14,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
-    :return: Returns a Tuple of attributions and approximation error. This data is used to create plots.
+    :return: d
     :rtype: Tuple
     """
 
@@ -25,7 +24,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     return attributions, approximation_error
 
 
-def make_attribution_plots(x, attributions, approximation_error, feature_names, use_wandb: bool = True):
+def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
     """_summary_
 
     :param attributions: _description_
@@ -35,6 +34,4 @@ def make_attribution_plots(x, attributions, approximation_error, feature_names,
     :param use_wandb: _description_, defaults to True
     :type use_wandb: bool, optional
     """
-    x_axis_data = np.arange(x.shape[1])
-    print(x_axis_data)
-    print("Hello world")
+    pass
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 692d2d98f..30ffb031e 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,14 +1,14 @@
 import unittest
 import torch
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import VanillaGRU
+from flood_forecast.basic.gru_vanilla import GRUVanilla
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
+        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
         self.test_data_loader = CSVDataLoader(
             "tests/data/test_data.csv",
             100,

From bd6ba3bae858aa295bd6a39a6919b4ef543c505b Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Tue, 30 May 2023 03:47:00 -0300
Subject: [PATCH 14/26] fixing unit tests of captum 2

---
 flood_forecast/interpretability.py | 2 +-
 tests/test_captum.py               | 8 +++++---
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index cb42391f2..2b31283f3 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -25,7 +25,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
 
 def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
-    """_summary_
+    """TODO implemnt
 
     :param attributions: _description_
     :type attributions: _type_
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 30ffb031e..6f88c8f10 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,16 +1,18 @@
 import unittest
 import torch
+import os
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import GRUVanilla
+from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
+        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
+        self.test_data_path = self.test_data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
         self.test_data_loader = CSVDataLoader(
-            "tests/data/test_data.csv",
+            os.path.join(self.test_data_path, "keag_small.csv"),
             100,
             20,
             "precip",

From a13120b2e1d0aa71c10d8caf521b35812477d304 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Tue, 30 May 2023 03:47:44 -0300
Subject: [PATCH 15/26] Revert "fixing unit tests of captum 2"

This reverts commit bd6ba3bae858aa295bd6a39a6919b4ef543c505b.
---
 flood_forecast/interpretability.py | 2 +-
 tests/test_captum.py               | 8 +++-----
 2 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 2b31283f3..cb42391f2 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -25,7 +25,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
 
 def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
-    """TODO implemnt
+    """_summary_
 
     :param attributions: _description_
     :type attributions: _type_
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 6f88c8f10..30ffb031e 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,18 +1,16 @@
 import unittest
 import torch
-import os
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import VanillaGRU
+from flood_forecast.basic.gru_vanilla import GRUVanilla
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
-        self.test_data_path = self.test_data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
+        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
         self.test_data_loader = CSVDataLoader(
-            os.path.join(self.test_data_path, "keag_small.csv"),
+            "tests/data/test_data.csv",
             100,
             20,
             "precip",

From 03b902e4d2efc2f11b5ddd4c1ac892eb4610a1d7 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 17:14:15 -0300
Subject: [PATCH 16/26] add print debugging 1

---
 flood_forecast/interpretability.py | 3 ++-
 tests/test_captum.py               | 8 +++++---
 2 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index cb42391f2..b5f6c1de8 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -10,7 +10,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
     :type model: _type_
-    :param test_loader: Should be a FF CSVDataLoader or a related subclass.
+    :param test_loader: Should be a FF CSVDataLoader or a related sub-class.
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
@@ -20,6 +20,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     attribution_method = attr_dict[method](model)
     x, y = test_loader[0]
+    print(attribution_method.attribute(x.unsqueeze(0), **additional_params))
     attributions, approximation_error = attribution_method.attribute(x.unsqueeze(0), **additional_params)
     return attributions, approximation_error
 
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 30ffb031e..6f88c8f10 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,16 +1,18 @@
 import unittest
 import torch
+import os
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import GRUVanilla
+from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
+        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
+        self.test_data_path = self.test_data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
         self.test_data_loader = CSVDataLoader(
-            "tests/data/test_data.csv",
+            os.path.join(self.test_data_path, "keag_small.csv"),
             100,
             20,
             "precip",

From 48925f58f0fc259555b7e79671021b678d1d946f Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 17:14:54 -0300
Subject: [PATCH 17/26] Revert "add print debugging 1"

This reverts commit 03b902e4d2efc2f11b5ddd4c1ac892eb4610a1d7.
---
 flood_forecast/interpretability.py | 3 +--
 tests/test_captum.py               | 8 +++-----
 2 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index b5f6c1de8..cb42391f2 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -10,7 +10,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
     :type model: _type_
-    :param test_loader: Should be a FF CSVDataLoader or a related sub-class.
+    :param test_loader: Should be a FF CSVDataLoader or a related subclass.
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
@@ -20,7 +20,6 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     attribution_method = attr_dict[method](model)
     x, y = test_loader[0]
-    print(attribution_method.attribute(x.unsqueeze(0), **additional_params))
     attributions, approximation_error = attribution_method.attribute(x.unsqueeze(0), **additional_params)
     return attributions, approximation_error
 
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 6f88c8f10..30ffb031e 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,18 +1,16 @@
 import unittest
 import torch
-import os
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import VanillaGRU
+from flood_forecast.basic.gru_vanilla import GRUVanilla
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
-        self.test_data_path = self.test_data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
+        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
         self.test_data_loader = CSVDataLoader(
-            os.path.join(self.test_data_path, "keag_small.csv"),
+            "tests/data/test_data.csv",
             100,
             20,
             "precip",

From 25c5713ea3dff1c7d5688de852aca711621ebf98 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 17:46:01 -0300
Subject: [PATCH 18/26] fixing code 3

---
 flood_forecast/interpretability.py | 14 +++++++++++---
 tests/test_captum.py               |  3 ++-
 2 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index cb42391f2..44f812d38 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -7,20 +7,28 @@
 
 def run_attribution(model, test_loader, method, additional_params: Dict) -> Tuple:
     """Function that creates attribution for a model based on Captum.
-
     :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
     :type model: _type_
-    :param test_loader: Should be a FF CSVDataLoader or a related subclass.
+    :param test_loader: Should be a FF CSVDataLoader or a related sub-class.
     :type test_loader: _type_
     :param method: _description_
     :type method: _type_
     :return: d
     :rtype: Tuple
+
+    .. code-block:: python
+
+    ..
     """
 
     attribution_method = attr_dict[method](model)
     x, y = test_loader[0]
-    attributions, approximation_error = attribution_method.attribute(x.unsqueeze(0), **additional_params)
+    the_data = attribution_method.attribute(x.unsqueeze(0), **additional_params)
+    if isinstance(the_data, tuple):
+        attributions, approximation_error = the_data
+    else:
+        attributions = the_data
+        approximation_error = None
     return attributions, approximation_error
 
 
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 30ffb031e..82ee25e69 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -19,7 +19,8 @@ def setUp(self):
 
     def test_run_attribution(self):
         """_summary_"""
-        attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
+        attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients",
+                                                     {"return_convergence_delta": True})
         self.assertEqual(approx_error.shape, torch.Size([1, 20, 3]))
         self.assertEqual(attributions.shape, torch.Size([1, 20, 3]))
 

From 2f4ad13aa1033a4f7e24f26ad153bf713beb4360 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 18:12:38 -0300
Subject: [PATCH 19/26] adding code 2

---
 flood_forecast/transformer_xl/cross_former.py | 2 +-
 tests/test_captum.py                          | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/flood_forecast/transformer_xl/cross_former.py b/flood_forecast/transformer_xl/cross_former.py
index 52c344283..8f67b0ccd 100644
--- a/flood_forecast/transformer_xl/cross_former.py
+++ b/flood_forecast/transformer_xl/cross_former.py
@@ -24,7 +24,7 @@ def __init__(
         """Crossformer: Transformer Utilizing Cross-Dimension Dependency for Multivariate Time Series Forecasting.
         https://github.com/Thinklab-SJTU/Crossformer
 
-        :param n_time_series: The total number of time series
+        :param n_time_series: The total number of time series passed to the model
         :type n_time_series: int
         :param forecast_history: The length of the input sequence
         :type forecast_history: int
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 82ee25e69..a9cd7a4e7 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,7 +1,7 @@
 import unittest
 import torch
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import GRUVanilla
+from flood_forecast.basic.gru_vanilla import VanillaGRU as GRUVanilla
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 

From b3f66c2e370c8f103aa86f267dfec4caa0099fc8 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 18:22:32 -0300
Subject: [PATCH 20/26] fixing code

dwdgl,;msfdgbml,ksd
---
 flood_forecast/interpretability.py | 9 ++++++---
 tests/test_captum.py               | 8 +++++---
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 44f812d38..121aae4e9 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -6,9 +6,9 @@
 
 
 def run_attribution(model, test_loader, method, additional_params: Dict) -> Tuple:
-    """Function that creates attribution for a model based on Captum.
+    """Function that creates attribution for a model based on Captum library.
     :param model: The deep learning model to be used for attribution. This should be a PyTorch model.
-    :type model: _type_
+    :type model: torch.nn.Module
     :param test_loader: Should be a FF CSVDataLoader or a related sub-class.
     :type test_loader: _type_
     :param method: _description_
@@ -18,6 +18,9 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
     .. code-block:: python
 
+        from flood_forecast.interpretability import run_attribution
+        model = VanillaGRU(3, 128, 2, 1, 0.2)
+
     ..
     """
 
@@ -33,7 +36,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 
 
 def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
-    """_summary_
+    """Creates the attribution plots and logs them to wandb if use_wandb is True.
 
     :param attributions: _description_
     :type attributions: _type_
diff --git a/tests/test_captum.py b/tests/test_captum.py
index a9cd7a4e7..0cfe57b56 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -1,16 +1,18 @@
 import unittest
 import torch
+import os
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
-from flood_forecast.basic.gru_vanilla import VanillaGRU as GRUVanilla
+from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
 
 
 class TestCaptum(unittest.TestCase):
     def setUp(self):
         # n_time_series: int, hidden_dim: int, num_layers: int, n_target: int, dropout: float
-        self.test_model = GRUVanilla(3, 128, 2, 1, 0.2)
+        self.test_model = VanillaGRU(3, 128, 2, 1, 0.2)
+        self.test_data_path = self.test_data_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "test_data")
         self.test_data_loader = CSVDataLoader(
-            "tests/data/test_data.csv",
+            os.path.join(self.test_data_path, "keag_small.csv"),
             100,
             20,
             "precip",

From eaea42e45cffe00f6a3a9df3f74cfb95d10d1979 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Thu, 1 Jun 2023 20:27:04 -0300
Subject: [PATCH 21/26] r

---
 tests/multi_decoder_test.json | 2 +-
 tests/test_captum.py          | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/tests/multi_decoder_test.json b/tests/multi_decoder_test.json
index 65d01cc78..26588dc39 100644
--- a/tests/multi_decoder_test.json
+++ b/tests/multi_decoder_test.json
@@ -18,7 +18,7 @@
        "forecast_history":5,
        "forecast_length":1,
        "train_end": 100,
-       "valid_start":101,
+       "valid_start":102,
        "valid_end": 201,
        "test_start": 202,
        "test_end": 290,
diff --git a/tests/test_captum.py b/tests/test_captum.py
index 0cfe57b56..da25de9d5 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -23,7 +23,8 @@ def test_run_attribution(self):
         """_summary_"""
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients",
                                                      {"return_convergence_delta": True})
-        self.assertEqual(approx_error.shape, torch.Size([1, 20, 3]))
+        print(attributions.shape, approx_error.shape)
+        self.assertEqual(approx_error.shape[0], 1)
         self.assertEqual(attributions.shape, torch.Size([1, 20, 3]))
 
     def test_create_attribution_plots(self):

From 0fdf9b5364a3e569e3ac89bf5050f4c33217a868 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Tue, 6 Jun 2023 15:28:16 -0300
Subject: [PATCH 22/26] fixng tests and more 2

---
 flood_forecast/interpretability.py | 4 ++--
 tests/test_captum.py               | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 121aae4e9..129f5a91e 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -38,8 +38,8 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
 def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
     """Creates the attribution plots and logs them to wandb if use_wandb is True.
 
-    :param attributions: _description_
-    :type attributions: _type_
+    :param attributions: A tensor of the attributions should be of dimension (batch_size, , n_features).
+    :type attributions: torch.Tensor
     :param approximation_error: _description_
     :type approximation_error: _type_
     :param use_wandb: _description_, defaults to True
diff --git a/tests/test_captum.py b/tests/test_captum.py
index da25de9d5..7846be8b0 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -23,9 +23,9 @@ def test_run_attribution(self):
         """_summary_"""
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients",
                                                      {"return_convergence_delta": True})
-        print(attributions.shape, approx_error.shape)
         self.assertEqual(approx_error.shape[0], 1)
-        self.assertEqual(attributions.shape, torch.Size([1, 20, 3]))
+        self.assertIsInstance(attributions, torch.Tensor)
+        # self.assertEqual(attributions.shape[2], 3)
 
     def test_create_attribution_plots(self):
         """_summary_"""

From 5574ef78394564b298e753deef838d7aea0c47b1 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Fri, 9 Jun 2023 22:08:46 -0400
Subject: [PATCH 23/26] r

---
 tests/test_captum.py | 3 +++
 tests/test_dual.json | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/tests/test_captum.py b/tests/test_captum.py
index 7846be8b0..78e316283 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -4,6 +4,7 @@
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
 from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
+from captum.attr import IntegratedGradients
 
 
 class TestCaptum(unittest.TestCase):
@@ -29,5 +30,7 @@ def test_run_attribution(self):
 
     def test_create_attribution_plots(self):
         """_summary_"""
+        attr = IntegratedGradients(self.test_model)
+        attr.attribute(self.test_data_loader)
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
         make_attribution_plots(attributions, approx_error, use_wandb=False)
diff --git a/tests/test_dual.json b/tests/test_dual.json
index 850632927..a52b07afd 100644
--- a/tests/test_dual.json
+++ b/tests/test_dual.json
@@ -20,7 +20,7 @@
        "train_end": 100,
        "valid_start":101,
        "valid_end": 201,
-       "test_start": 202,
+       "test_start": 232,
        "test_end": 290,
        "target_col": ["cfs"],
        "relevant_cols": ["cfs", "precip", "temp"],

From b5e68c70bd2779516c58ee292fd60d3fd68a38e5 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Mon, 12 Jun 2023 01:38:57 -0400
Subject: [PATCH 24/26] Revert "r"

This reverts commit 5574ef78394564b298e753deef838d7aea0c47b1.
---
 tests/test_captum.py | 3 ---
 tests/test_dual.json | 2 +-
 2 files changed, 1 insertion(+), 4 deletions(-)

diff --git a/tests/test_captum.py b/tests/test_captum.py
index 78e316283..7846be8b0 100644
--- a/tests/test_captum.py
+++ b/tests/test_captum.py
@@ -4,7 +4,6 @@
 from flood_forecast.interpretability import run_attribution, make_attribution_plots
 from flood_forecast.basic.gru_vanilla import VanillaGRU
 from flood_forecast.preprocessing.pytorch_loaders import CSVDataLoader
-from captum.attr import IntegratedGradients
 
 
 class TestCaptum(unittest.TestCase):
@@ -30,7 +29,5 @@ def test_run_attribution(self):
 
     def test_create_attribution_plots(self):
         """_summary_"""
-        attr = IntegratedGradients(self.test_model)
-        attr.attribute(self.test_data_loader)
         attributions, approx_error = run_attribution(self.test_model, self.test_data_loader, "IntegratedGradients", {})
         make_attribution_plots(attributions, approx_error, use_wandb=False)
diff --git a/tests/test_dual.json b/tests/test_dual.json
index a52b07afd..850632927 100644
--- a/tests/test_dual.json
+++ b/tests/test_dual.json
@@ -20,7 +20,7 @@
        "train_end": 100,
        "valid_start":101,
        "valid_end": 201,
-       "test_start": 232,
+       "test_start": 202,
        "test_end": 290,
        "target_col": ["cfs"],
        "relevant_cols": ["cfs", "precip", "temp"],

From 7c2d1f01429739a8fd102e112921c5ac73a1965d Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Sun, 25 Jun 2023 04:21:51 -0400
Subject: [PATCH 25/26] add basic architecting

---
 flood_forecast/interpretability.py | 11 ++++++++++-
 tests/da_meta.json                 |  2 +-
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 129f5a91e..37b7f4f67 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,5 +1,6 @@
 from captum.attr import IntegratedGradients, DeepLift, GradientShap, NoiseTunnel, FeatureAblation
 from typing import Tuple, Dict
+import numpy as np
 
 attr_dict = {"IntegratedGradients": IntegratedGradients, "DeepLift": DeepLift, "GradientSHAP": GradientShap,
              "NoiseTunnel": NoiseTunnel, "FeatureAblation": FeatureAblation}
@@ -35,7 +36,7 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     return attributions, approximation_error
 
 
-def make_attribution_plots(attributions, approximation_error, use_wandb: bool = True):
+def make_attribution_plots(attributions, approximation_error, model, x, y, use_wandb: bool = True):
     """Creates the attribution plots and logs them to wandb if use_wandb is True.
 
     :param attributions: A tensor of the attributions should be of dimension (batch_size, , n_features).
@@ -45,4 +46,12 @@ def make_attribution_plots(attributions, approximation_error, use_wandb: bool =
     :param use_wandb: _description_, defaults to True
     :type use_wandb: bool, optional
     """
+    x_axis_data = np.arange(x.shape[1])
+    x_axis_data_labels = model.params["fea"]
+
+    ig_attr_test_sum = attributions.detach().numpy().sum(0)
+    ig_attr_test_norm_sum = ig_attr_test_sum / np.linalg.norm(ig_attr_test_sum, ord=1)
+
+    lin_weight = model.lin1.weight[0].detach().numpy()
+    y_axis_lin_weight = lin_weight / np.linalg.norm(lin_weight, ord=1)
     pass
diff --git a/tests/da_meta.json b/tests/da_meta.json
index d7b0cca62..5f7748685 100644
--- a/tests/da_meta.json
+++ b/tests/da_meta.json
@@ -35,7 +35,7 @@
        "forecast_history":5,
        "forecast_length":1,
        "train_end": 300,
-       "valid_start":302,
+       "valid_start":333,
        "valid_end": 404,
        "test_end": 500,
        "target_col": ["cfs"],

From a9053aeba678aff9ae0e27f94899b22e53f101e7 Mon Sep 17 00:00:00 2001
From: isaacmg <igodfried@isaac26.com>
Date: Sun, 25 Jun 2023 04:40:21 -0400
Subject: [PATCH 26/26] fixing code 3

---
 flood_forecast/interpretability.py | 16 +++++-----------
 tests/da_meta.json                 |  2 +-
 2 files changed, 6 insertions(+), 12 deletions(-)

diff --git a/flood_forecast/interpretability.py b/flood_forecast/interpretability.py
index 37b7f4f67..57ae68226 100644
--- a/flood_forecast/interpretability.py
+++ b/flood_forecast/interpretability.py
@@ -1,6 +1,5 @@
 from captum.attr import IntegratedGradients, DeepLift, GradientShap, NoiseTunnel, FeatureAblation
 from typing import Tuple, Dict
-import numpy as np
 
 attr_dict = {"IntegratedGradients": IntegratedGradients, "DeepLift": DeepLift, "GradientSHAP": GradientShap,
              "NoiseTunnel": NoiseTunnel, "FeatureAblation": FeatureAblation}
@@ -36,22 +35,17 @@ def run_attribution(model, test_loader, method, additional_params: Dict) -> Tupl
     return attributions, approximation_error
 
 
-def make_attribution_plots(attributions, approximation_error, model, x, y, use_wandb: bool = True):
+def make_attribution_plots(model, methods, use_wandb: bool = True):
     """Creates the attribution plots and logs them to wandb if use_wandb is True.
 
     :param attributions: A tensor of the attributions should be of dimension (batch_size, , n_features).
     :type attributions: torch.Tensor
     :param approximation_error: _description_
     :type approximation_error: _type_
-    :param use_wandb: _description_, defaults to True
+    :param use_wandb: _description_, defaults2 to True
     :type use_wandb: bool, optional
     """
-    x_axis_data = np.arange(x.shape[1])
-    x_axis_data_labels = model.params["fea"]
-
-    ig_attr_test_sum = attributions.detach().numpy().sum(0)
-    ig_attr_test_norm_sum = ig_attr_test_sum / np.linalg.norm(ig_attr_test_sum, ord=1)
-
-    lin_weight = model.lin1.weight[0].detach().numpy()
-    y_axis_lin_weight = lin_weight / np.linalg.norm(lin_weight, ord=1)
+    for method in methods:
+        attributions, approx = run_attribution(model.model, model.test_loader, methods, {})
+    # DO PLOTTING HERE
     pass
diff --git a/tests/da_meta.json b/tests/da_meta.json
index 5f7748685..d7b0cca62 100644
--- a/tests/da_meta.json
+++ b/tests/da_meta.json
@@ -35,7 +35,7 @@
        "forecast_history":5,
        "forecast_length":1,
        "train_end": 300,
-       "valid_start":333,
+       "valid_start":302,
        "valid_end": 404,
        "test_end": 500,
        "target_col": ["cfs"],