Update app.py

app.py

@@ -1,9 +1,9 @@
 from newspaper import Article
 from newspaper import Config
-import nltk
-nltk.download('punkt')
-from gradio.mix import Series
 import gradio as gr
+# Set up OpenAI API credentials
+import openai
+openai.api_key = os.getenv('api_token')
 def extract_article_text(url):
   USER_AGENT = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0) Gecko/20100101 Firefox/78.0'
   config = Config()
@@ -15,26 +15,46 @@ def extract_article_text(url):
   article.parse()
   text = article.text
   return text
-extractor = gr.Interface(extract_article_text, 'text', 'text')
-summarizer = gr.Interface.load("huggingface/facebook/bart-large-cnn")
 
-sample_url = [['https://www.technologyreview.com/2021/07/22/1029973/deepmind-alphafold-protein-folding-biology-disease-drugs-proteome/'],
-              ['https://www.technologyreview.com/2021/07/21/1029860/disability-rights-employment-discrimination-ai-hiring/'],
-              ['https://www.technologyreview.com/2021/07/09/1028140/ai-voice-actors-sound-human/']]
+def get_completion(prompt, model="gpt-3.5-turbo"):
+    messages = [{"role": "user", "content": prompt}]
+    response = openai.ChatCompletion.create(
+        model=model,
+        messages=messages,
+        temperature=0.5, # this is the degree of randomness of the model's output
+    )
+    return response.choices[0].message["content"]
 
-desc =  '''
-        Generate summary of an article using bart-large-cnn model by Facebook.
-        '''
+def prompt_summary(url,movie):
+  text = extract_article_text(url)
+  text = text[:4096]
+  prompt_sum = f"""
+  Summarize the text {text} as if a 8-year old kid understands. The summary should be atmost 200 words and should help the kid understand how the summary could help him solve a real-world problem. Here is the format:
+  1.Importance of the article:
+  2.Real world scenario:
+  3.Key takeaway:
+  """
 
-iface = Series(extractor, summarizer, 
-  inputs = gr.inputs.Textbox(
-      lines = 2,
-      label = 'URL'
-  ),
-  outputs = 'text',
-  title = 'Article Cortex: Transformers powered article summarization engine',
-  theme = 'huggingface',
-  description = desc,
-  examples=sample_url,allow_flagging="manual",flagging_options=["random summary", "useful summary"])
+  prompt_mov = f"""
+  Convert the technical article {text} into a short story involving the characters from the movie {movie}. The story should not exceed 200 words and should be written in a way that captures the essence of the article while also making it engaging and entertaining.
+  """
+  prompt_topic = f"""
+  Extract 5 key topics from the text {text}. The topics should clearly tell the user why it is important to read the article. Length of the topic should be limited to a single word
+  """
+  response_top = get_completion(prompt_topic)
+  response_mov = get_completion(prompt_mov)
+  response_sum = get_completion(prompt_sum)
+  return response_top, response_sum, response_mov
 
-iface.launch(debug=True)
+inputs = [
+    gr.inputs.Textbox(label="Article URL"),
+    gr.inputs.Textbox(label="Which is your favorite movie?")
+]
+
+outputs = [
+    gr.outputs.Textbox(label="Key topics"),
+    gr.outputs.Textbox(label="Summary without jargon"),
+    gr.outputs.Textbox(label="Summary as movie synopsis")
+]
+
+gr.Interface(prompt_summary, inputs, outputs, title="Article Cortex", description="Helps you understand any technical article as if it were a movie synopsis.").launch(debug=True)