added eval image and explanation

README.md

@@ -1,7 +1,7 @@
 ---
 title: "Evaluating NER Evaluation Metrics!"
 emoji: 🤗
-colorFrom: yellow
+colorFrom: blue
 colorTo: yellow
 sdk: streamlit
 sdk_version: "1.36.0"

app.py

@@ -5,7 +5,14 @@ import streamlit as st
 from annotated_text.util import get_annotated_html
 from streamlit_annotation_tools import text_labeler
 
-from constants import PREDICTION_ADDITION_INSTRUCTION
+from constants import (
+    APP_INTRO,
+    APP_TITLE,
+    EVAL_FUNCTION_INTRO,
+    EVAL_FUNCTION_PROPERTIES,
+    NER_TASK_EXPLAINER,
+    PREDICTION_ADDITION_INSTRUCTION,
+)
 from evaluation_metrics import EVALUATION_METRICS
 from predefined_example import EXAMPLES
 from span_dataclass_converters import (
@@ -28,15 +35,28 @@ def get_examples_attributes(selected_example):
 
 if __name__ == "__main__":
     st.set_page_config(layout="wide")
-    st.title("📈 NER Metrics Comparison ⚖️")
+    st.title(APP_TITLE)
 
-    st.write(
-        "Evaluation for the NER task requires a ground truth and a prediction that will be evaluated. The ground truth is shown below, add predictions in the next section to compare the evaluation metrics."
-    )
+    st.write(APP_INTRO)
     explanation_tab, comparision_tab = st.tabs(["📙 Explanation", "⚖️ Comparision"])
 
     with explanation_tab:
-        st.write("This is the place holder for explanation of all the metrics")
+        st.write(EVAL_FUNCTION_INTRO)
+        st.image("assets/eval_fnc_viz.png", caption="Evaluation Function Flow")
+        st.markdown(EVAL_FUNCTION_PROPERTIES)
+        st.markdown(NER_TASK_EXPLAINER)
+        st.subheader("Evaluation Metrics")
+        metric_names = "\n".join(
+            [
+                f"{index+1}. " + evaluation_metric.name
+                for index, evaluation_metric in enumerate(EVALUATION_METRICS)
+            ]
+        )
+        st.markdown(
+            "The different evaluation metrics we have for the NER task are\n"
+            "\n"
+            f"{metric_names}"
+        )
 
     with comparision_tab:
         # with st.container():
@@ -78,7 +98,7 @@ if __name__ == "__main__":
             st.subheader("Adding predictions")
             st.markdown(PREDICTION_ADDITION_INSTRUCTION)
             st.write(
-                "Note: Only the spans of the selected label name is shown at a given instance.",
+                "Note: Only the spans of the selected label name are shown at a given instance. Click on the label to see the corresponding spans. (or view the json below)",
             )
             labels = text_labeler(text, gt_labels)
             st.json(labels, expanded=False)

constants.py

@@ -1,3 +1,22 @@
+APP_TITLE = "📐 NER Metrics Comparison ⚖️"
+
+APP_INTRO = "The NER task is performed over a piece of text and involves recognition of entities belonging to a desired entity set and classifying them. The various metrics are explained in the explanation tab. Once you go through them, head to the comparision tab to test out some examples."
+
+
+### EXPLANATION TAB ###
+
+EVAL_FUNCTION_INTRO = "An evaluation function tells us how well a model is performing. The basic working of any evaluation function involves comparing the model's output with the ground truth to give a score of correctness."
+EVAL_FUNCTION_PROPERTIES = """
+Some basic properties of an evaluation function are -
+1. Give an output score equivalent to the upper bound when the prediction is completely correct(in some tasks, multiple variations of a predictions can be considered correct)
+2. Give an output score equivalent to the lower bound when the prediction is completely wrong.
+3. GIve an output score between upper and lower bound in other cases, corresponding to the degree of correctness.
+"""
+NER_TASK_EXPLAINER = """
+The output of the NER task can be represented in either token format or span format.
+"""
+### COMPARISION TAB ###
+
 PREDICTION_ADDITION_INSTRUCTION = """
 Add predictions to the list of predictions on which the evaluation metric will be caculated.
 - Select the entity type/label name and then highlight the span in the text below.

span_dataclass_converters.py

@@ -1,3 +1,12 @@
+"""
+There are 4 data formats for spans
+1. annotations - this is what we obtain from the text_annotator, the format can be seen in the predefined_examples, gt_labels
+2. higlight_spans - this is the format used by the highlighter to return the highlighted html text. This is a list of string/tuples("string", "label", color)
+3. ner_spans - this is the standard format used for representing ner_spans, it is a dict of {"start":int, "end":int, "label":str, "span_text":str}
+4. Token level output - this is delt with in the token_level_output file, this is either a list of tuples with [(token, label)] or just a list of [label, label]
+"""
+
+
 def get_ner_spans_from_annotations(annotated_labels):
     spans = []
     for entity_type, spans_list in annotated_labels.items():