Update app.py

app.py

@@ -269,7 +269,35 @@ from transformers import pipeline, AutoProcessor, AutoModel
 # =======================================
 # 
 # ======================================= 
-def sentence_to_audio(summary_txt):
+def sentence_to_audio(fileobj):
+
+
+    from transformers import pipeline
+
+    # text mining from pdf
+    text_per_page = read_pdf(fileobj.name)
+    text_per_page.keys()
+    page_1 = text_per_page['Page_0']
+
+
+    # picking up the abstract from the first page content
+    flag=False
+    abstract_sect=""
+    
+    for i in range(len(page_1)):
+        if page_1[0][i].strip()=="Abstract":
+            flag=True
+        if page_1[0][i].strip()=="1 Introduction":
+            flag = False
+        if flag:
+            # abstract_sect contains the Abstract section content
+            abstract_sect+=page_1[0][i]
+
+    # abstract summarization
+    summarizer = pipeline("summarization", model="knkarthick/MEETING_SUMMARY")
+    summary=(summarizer(abstract_sect))
+    summary_text=summary[0].get("summary_text")
+    
     # Sentence 2 Speech
     processor = AutoProcessor.from_pretrained("suno/bark-small")
     model = AutoModel.from_pretrained("suno/bark-small")
@@ -282,42 +310,18 @@ def sentence_to_audio(summary_txt):
     return sampling_rate, speech_values.cpu().numpy().squeeze()
 
 
-#text_per_page = read_pdf(pdf_path)
-#text_per_page.keys()
-#page_1 = text_per_page['Page_0']
-
 # ============================================================================================
 
-# picking up the abstract from the first page content
-#flag=False
-#abstract_sect=""
-
-#for i in range(len(page_1)):
-#  if page_1[0][i].strip()=="Abstract":
-#    flag=True
-#  if page_1[0][i].strip()=="1 Introduction":
-#    flag = False
-#  if flag:
-#    # abstract_sect contains the Abstract section content
-#    abstract_sect+=page_1[0][i]
-
-
-#from transformers import pipeline
-#
-#summarizer = pipeline("summarization", model="knkarthick/MEETING_SUMMARY")
-#summary=(summarizer(abstract_sect))
-#summary_text=summary[0].get("summary_text")
-#print(summary_text)
 
 
 # ===========================================================    
 
-summary_txt="It is dangerous to think of machine learning as a free-to-use toolkit, as it is common to incur ongoing maintenance costs in real-world ML systems"
+#summary_txt="It is dangerous to think of machine learning as a free-to-use toolkit, as it is common to incur ongoing maintenance costs in real-world ML systems"
 
 sentence_to_audio(summary_txt)
 
 pdf_path=os.path.join(os.path.abspath(""), "hidden-technical-debt-in-machine-learning-systems-Paper.pdf")
 pdf_path2=os.path.join(os.path.abspath(""), "1812_05944.pdf")
 
-demo = gr.Interface(fn=sentence_to_audio, inputs="file", outputs="audio",examples=[pdf_path,pdf_path2])
+demo = gr.Interface(fn=sentence_to_audio, inputs="file", outputs=["audio","text"],examples=[pdf_path,pdf_path2])
 demo.launch(share=True)