Upload 4 files

pages/1_💾_Data_Panel.py

@@ -0,0 +1,57 @@
+# Read
+import streamlit as st
+import sys,os
+
+
+sys.path.append(f'{os.getcwd()}/utils')
+from utils.data_users import get_product_dev_page_layout,get_product_manager_page_layout,get_product_practitioner_page_layout
+
+
+print(os.getcwd())
+# st.write(st.session_state.user_group)
+USER_GROUPS = ["Developer", "Manager", "Practitioner"]
+
+st.set_page_config(layout="wide")
+
+if 'user_group' not in st.session_state:
+    index_tmp = 0
+else: 
+    index_tmp = USER_GROUPS.index(st.session_state['user_group'])
+
+#Sidebar for USER GROUPS
+st.sidebar.title("USER GROUPS")
+backend = st.sidebar.selectbox(
+            "Select User-Group ", USER_GROUPS, index=index_tmp
+        )
+
+st.session_state['user_group'] = backend
+
+# #     with st.sidebar:
+# st.sidebar.title("🎈Explore Data Panel")
+
+st.title("Data Panel for OCT Image Analysis")
+
+st.write(
+        """
+        ##
+        To gain a comprehensive understanding of the AI system, examining the data has a crucial role. The Data Panel adopts a data-centric approach, providing detailed information about the following aspects of the data:
+        """)
+
+
+list_test = """<ul>
+                <li>Data Source Information contains information related to the modality, format, domain, ethical considerations, including licensing and data version. </li>
+                <li>Exploratory Data Stats presents exploratory data analysis information covering train/validation/test data division, summary statistics, and sample visualization from each category. </li>
+                <li>Data Onboarding provides information about the data pre-processing and post-processing steps applied to the dataset before training, as well as any data augmentations that were used.</li>
+ </ul>"""
+st.markdown(list_test, unsafe_allow_html=True)
+
+
+if backend == "Developer":
+    get_product_dev_page_layout()
+
+
+if backend == "Manager":
+    get_product_manager_page_layout()
+
+if backend == "Practitioner":
+    get_product_practitioner_page_layout()

pages/2_📝_Model_Panel.py

@@ -0,0 +1,52 @@
+import streamlit as st
+import sys,os
+
+
+sys.path.append(f'{os.getcwd()}/utils')
+from utils.model_users import get_product_dev_page_layout,get_product_manager_page_layout,get_product_practitioner_page_layout
+
+# st.write(st.session_state.user_group)
+USER_GROUPS = ["Developer", "Manager", "Practitioner"]
+
+st.set_page_config(layout="wide")
+
+if 'user_group' not in st.session_state:
+    index_tmp = 0
+else: 
+    index_tmp = USER_GROUPS.index(st.session_state['user_group'])
+
+#Sidebar for USER GROUPS
+st.sidebar.title("USER GROUPS")
+backend = st.sidebar.selectbox(
+            "Select User-Group ", USER_GROUPS, index=index_tmp
+        )
+
+st.session_state['user_group'] = backend
+
+
+
+st.title("Model Panel for OCT Image Analysis")
+st.write(
+        """
+       The users can find following information regarding the AI model developed for the task: what is the purpose of the model, how it was developed and how it is used. Thus,answers 
+       to these questions areprovided via the tabs described below.
+        """)
+
+list_test = """<ul>
+                <li>Model Generic information contains a summary of the model’s details, including its main features, capabilities, and intended use. 
+                    It also highlights the model’s behaviour, such as the type of data inputs(image, feature etc) it can handle and the types of outputs it produces.</li>
+                <li>Model Development Information includes information on the hyperparameters, model development framework/library such as Tensorflow or PyTorch, and other technical details. It also includes a complete analysis of the model’s inference performance, such as the model size, 
+                        hardware-specific (GPU and CPU) inference time,and speed, as well as reproducibility check list.</li>
+                <li>Model Deployment Information provides information on how the model is used in production, including details on the inference speed and latency,  
+                        and how users can access and interact with the model in production.</li>
+                </ul>"""
+st.markdown(list_test, unsafe_allow_html=True)
+
+if backend == "Developer":
+    get_product_dev_page_layout()
+
+if backend == "Manager":
+    get_product_manager_page_layout()
+
+if backend == "Practitioner":
+    get_product_practitioner_page_layout()

pages/3_🎯_Performance_Evaluation_Panel.py

@@ -0,0 +1,42 @@
+import streamlit as st
+import sys,os
+
+sys.path.append(f'{os.getcwd()}/utils')
+
+from utils.eval_users import get_product_dev_page_layout
+
+# st.write(st.session_state.user_group)
+USER_GROUPS = ["Developer", "Manager", "Practitioner"]
+
+st.set_page_config(layout="wide")
+
+if 'user_group' not in st.session_state:
+    index_tmp = 0
+else: 
+    index_tmp = USER_GROUPS.index(st.session_state['user_group'])
+
+#Sidebar for USER GROUPS
+st.sidebar.title("USER GROUPS")
+backend = st.sidebar.selectbox(
+            "Select User-Group ", USER_GROUPS, index=index_tmp
+        )
+
+st.session_state['user_group'] = backend
+
+st.title("Performance Evaluation Panel for OCT Image Analysis")
+st.write(
+        """
+       This panel provides information on the evaluation of the AI model’s performance, including details on the metrics used and the results of the evaluation. Users can also find 
+       our notes regarding the encountered issues. The performance metric visualizations and samples of failure and success cases are given in this panel as well.""")
+
+list_test = """<ul>
+                    <li>Evaluation Metrics explains the details of the performance metrics and how each metric is calculated. </li>
+                    <li>Performance Summary includes visualizations of the performance metrics over the test set. Users can also find the characteristics of the evaluation data set</li>
+                    <li>Issues and Limitations provides examples of observed failure cases, along with visualizations and any possible observations behind the failure cases.</li>
+                </ul>"""
+st.markdown(list_test, unsafe_allow_html=True)
+
+
+get_product_dev_page_layout()
+# if backend == "Developer":
+#     get_product_dev_page_layout()

pages/4_🔬_Decision_Exploration_Panel.py

@@ -30,16 +30,19 @@ st.write(
        """)
 
 list_test = """<ul>
-                <li>Global Explanations: Trepresentative samples and task-specific borderline samples from the training data, such as representative
-                samples from each category in a classification task. These samples enable users
+                <li>Global Explanations: this tab illustrates both representative samples and task-specific borderline samples in the training data from each category in the classification task. These samples enable users
                 to understand which inputs contribute to the model’s decisions. On the other
-                hand, borderline cases can be used to highlight potential failure cases. Also, once can see manifold visualization
-                of both representative and borderline samples </li>
-                <li>Instance Level Explanations: Tthis panel is intended for model sensitivity analyis, decision correction, and decision highlighting using both local expla-
-                nation methods and presenting similar samples to the target sample from the
-                representative samples of the training set.</li>
+                hand, borderline cases can be used to highlight potential failure cases. Also, one can see a manifold visualization
+                of both representative and borderline samples here.</li>
+                <li>Instance Level Explanations: this panel is intended for model sensitivity analysis, decision correction, and decision highlighting using both local expla-
+                nation methods and presenting samples similar to the target sample from the
+                representative samples of the training set. [GradCAM](https://arxiv.org/abs/1610.02391) method is used as a post-hoc explanation technique to highlight aregion of interest.</li>
                     </ul>"""
 st.markdown(list_test, unsafe_allow_html=True)
 
-if backend == "Developer":
+# if backend == "Developer":
+#     get_product_dev_page_layout()
+if backend == "Manager":
+    get_product_dev_page_layout(user_type ="Manager")
+else:
     get_product_dev_page_layout()