Update app.py

app.py

@@ -4,14 +4,25 @@ import gradio as gr
 
 markdownn = '''
 # Named Entity Recognition Pipeline with DeepSparse
+Named Entity Recognition is the task of extracting and locating named entities in a sentence. The entities include, people's names, location, organizations, etc. 
+![Named Entity Recognition Pipeline with DeepSparse](https://huggingface.co/spaces/neuralmagic/nlp-ner/resolve/main/ner.png)
+
+## What is DeepSparse?
 DeepSparse is sparsity-aware inference runtime offering GPU-class performance on CPUs and APIs to integrate ML into your application. DeepSparse provides sparsified pipelines for computer vision and NLP. 
-The pipelines are similar to Hugging Face pipelines but are faster because they have been pruned and quantized. SparseML Named Entity Recognition Pipelines integrate with Hugging Face’s Transformers library to enable the sparsification of a large set of transformers models. Here is sample code for a token classification pipeline: 
+The pipelines are similar to Hugging Face pipelines but are faster because they have been pruned and quantized. SparseML Named Entity Recognition Pipelines integrate with Hugging Face’s Transformers library to enable the sparsification of a large set of transformers models. 
+### Inference
+Here is sample code for a token classification pipeline: 
 ```python
 from deepsparse import Pipeline
 pipeline = Pipeline.create(task="ner", model_path="zoo:nlp/token_classification/distilbert-none/pytorch/huggingface/conll2003/pruned80_quant-none-vnni")
 inference = pipeline(text)
 print(inference)
 ```
+## Use case example 
+The Named Entity Recognition Pipeline can process text before storing the information in a database. 
+
+For example, you may want to process texta and store the entities in different columns depending on the entity type. 
+
 '''
 task = "ner"
 dense_qa_pipeline = Pipeline.create(
@@ -23,36 +34,39 @@ sparse_qa_pipeline = Pipeline.create(
         task=task,
         model_path="zoo:nlp/token_classification/bert-base/pytorch/huggingface/conll2003/12layer_pruned80_quant-none-vnni",
     )
+
+def map_ner(inference):
+  entities = []
+  for item in dict(inference)['predictions'][0]:
+    dictionary = dict(item)
+    entity = dictionary['entity']
+    if entity == "LABEL_0":
+        value = "O"
+    elif entity == "LABEL_1":
+        value = "B-PER"
+    elif entity == "LABEL_2":
+        value = "I-PER"
+    elif entity == "LABEL_3":
+        value = "-ORG"
+    elif entity == "LABEL_4":
+        value = "I-ORG"
+    elif entity == "LABEL_5":
+        value = "B-LOC"
+    elif entity == "LABEL_6":
+        value = "I-LOC"
+    elif entity == "LABEL_7":
+        value = "B-MISC"
+    else:
+        value = "I-MISC"
+    dictionary['entity'] = value
+    entities.append(dictionary)
+    return entities
+      
 def run_pipeline(text):
     dense_start = time.perf_counter()
-    dense_entities = []
-    sparse_entities = []
     
     dense_output = dense_qa_pipeline(text)
-    for item in dict(dense_output)['predictions'][0]:
-        dictionary = dict(item)
-        entity = dictionary['entity']
-        
-        if entity == "LABEL_0":
-            value = "O"
-        elif entity == "LABEL_1":
-            value = "B-PER"
-        elif entity == "LABEL_2":
-            value = "I-PER"
-        elif entity == "LABEL_3":
-            value = "-ORG"
-        elif entity == "LABEL_4":
-            value = "I-ORG"
-        elif entity == "LABEL_5":
-            value = "B-LOC"
-        elif entity == "LABEL_6":
-            value = "I-LOC"
-        elif entity == "LABEL_7":
-            value = "B-MISC"
-        else:
-            value = "I-MISC"
-        dictionary['entity'] = value
-        dense_entities.append(dictionary)
+    dense_entities = map_ner(dense_output)
         
     dense_output = {"text": text, "entities": dense_entities}
         
@@ -62,30 +76,7 @@ def run_pipeline(text):
     sparse_start = time.perf_counter()
     
     sparse_output = sparse_qa_pipeline(text)
-    for item in dict(sparse_output)['predictions'][0]:
-        sparse_dictionary = dict(item)
-        entity = sparse_dictionary['entity']
-        
-        if entity == "LABEL_0":
-            value = "O"
-        elif entity == "LABEL_1":
-            value = "B-PER"
-        elif entity == "LABEL_2":
-            value = "I-PER"
-        elif entity == "LABEL_3":
-            value = "-ORG"
-        elif entity == "LABEL_4":
-            value = "I-ORG"
-        elif entity == "LABEL_5":
-            value = "B-LOC"
-        elif entity == "LABEL_6":
-            value = "I-LOC"
-        elif entity == "LABEL_7":
-            value = "B-MISC"
-        else:
-            value = "I-MISC"
-        sparse_dictionary['entity'] = value
-        sparse_entities.append(sparse_dictionary)
+    sparse_entities = map_ner(sparse_output)
 
     sparse_output = {"text": text, "entities": sparse_entities}