Update app.py

app.py

@@ -1,4 +1,283 @@
-#test
+#temp
+# https://huggingface.co/spaces/Mishmosh/MichelleAssessment3
+
+# Install Rust 
+RUN curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
+
+#RUN python -m pip install --upgrade pip
+python -m pip install --upgrade pip
+
+#pip install --upgrade pip
+RUN pip install --no-cache-dir -r requirements.txt
+RUN pip install --use-feature=in-tree-build tokenizers
+
+#!pip install PyPDF2
+#!pip install sentencepiece
+#!pip install pdfminer.six
+#!pip install pdfplumber
+#!pip install pdf2image
+#!pip install Pillow
+#!pip install pytesseract
+# @title
+#!apt-get install poppler-utils
+#!apt install tesseract-ocr
+#!apt install libtesseract-dev
+import PyPDF2
+from pdfminer.high_level import extract_pages, extract_text
+from pdfminer.layout import LTTextContainer, LTChar, LTRect, LTFigure
+import pdfplumber
+from PIL import Image
+from pdf2image import convert_from_path
+import pytesseract
+import os
+def text_extraction(element):
+    # Extracting the text from the in-line text element
+    line_text = element.get_text()
+
+    # Find the formats of the text
+    # Initialize the list with all the formats that appeared in the line of text
+    line_formats = []
+    for text_line in element:
+        if isinstance(text_line, LTTextContainer):
+            # Iterating through each character in the line of text
+            for character in text_line:
+                if isinstance(character, LTChar):
+                    # Append the font name of the character
+                    line_formats.append(character.fontname)
+                    # Append the font size of the character
+                    line_formats.append(character.size)
+    # Find the unique font sizes and names in the line
+    format_per_line = list(set(line_formats))
+
+    # Return a tuple with the text in each line along with its format
+    return (line_text, format_per_line)
+# @title
+# Create a function to crop the image elements from PDFs
+def crop_image(element, pageObj):
+    # Get the coordinates to crop the image from the PDF
+    [image_left, image_top, image_right, image_bottom] = [element.x0,element.y0,element.x1,element.y1]
+    # Crop the page using coordinates (left, bottom, right, top)
+    pageObj.mediabox.lower_left = (image_left, image_bottom)
+    pageObj.mediabox.upper_right = (image_right, image_top)
+    # Save the cropped page to a new PDF
+    cropped_pdf_writer = PyPDF2.PdfWriter()
+    cropped_pdf_writer.add_page(pageObj)
+    # Save the cropped PDF to a new file
+    with open('cropped_image.pdf', 'wb') as cropped_pdf_file:
+        cropped_pdf_writer.write(cropped_pdf_file)
+
+# Create a function to convert the PDF to images
+def convert_to_images(input_file,):
+    images = convert_from_path(input_file)
+    image = images[0]
+    output_file = "PDF_image.png"
+    image.save(output_file, "PNG")
+
+# Create a function to read text from images
+def image_to_text(image_path):
+    # Read the image
+    img = Image.open(image_path)
+    # Extract the text from the image
+    text = pytesseract.image_to_string(img)
+    return text
+# @title
+# Extracting tables from the page
+
+def extract_table(pdf_path, page_num, table_num):
+    # Open the pdf file
+    pdf = pdfplumber.open(pdf_path)
+    # Find the examined page
+    table_page = pdf.pages[page_num]
+    # Extract the appropriate table
+    table = table_page.extract_tables()[table_num]
+    return table
+
+# Convert table into the appropriate format
+def table_converter(table):
+    table_string = ''
+    # Iterate through each row of the table
+    for row_num in range(len(table)):
+        row = table[row_num]
+        # Remove the line breaker from the wrapped texts
+        cleaned_row = [item.replace('\n', ' ') if item is not None and '\n' in item else 'None' if item is None else item for item in row]
+        # Convert the table into a string
+        table_string+=('|'+'|'.join(cleaned_row)+'|'+'\n')
+    # Removing the last line break
+    table_string = table_string[:-1]
+    return table_string
+# @title
+def read_pdf(pdf_path):
+  # create a PDF file object
+  pdfFileObj = open(pdf_path, 'rb')
+  # create a PDF reader object
+  #pdfReaded = PyPDF2.PdfReader(pdfFileObj) #coded out as suggested by chatgpt
+  pdfReaded = PyPDF2.PdfFileReader(pdfFileObj)
+
+  # Create the dictionary to extract text from each image
+  text_per_page = {}
+  # We extract the pages from the PDF
+  for pagenum, page in enumerate(extract_pages(pdf_path)):
+      print("Elaborating Page_" +str(pagenum))
+      # Initialize the variables needed for the text extraction from the page
+      pageObj = pdfReaded.pages[pagenum]
+      page_text = []
+      line_format = []
+      text_from_images = []
+      text_from_tables = []
+      page_content = []
+      # Initialize the number of the examined tables
+      table_num = 0
+      first_element= True
+      table_extraction_flag= False
+      # Open the pdf file
+      pdf = pdfplumber.open(pdf_path)
+      # Find the examined page
+      page_tables = pdf.pages[pagenum]
+      # Find the number of tables on the page
+      tables = page_tables.find_tables()
+
+
+      # Find all the elements
+      page_elements = [(element.y1, element) for element in page._objs]
+      # Sort all the elements as they appear in the page
+      page_elements.sort(key=lambda a: a[0], reverse=True)
+
+      # Find the elements that composed a page
+      for i,component in enumerate(page_elements):
+          # Extract the position of the top side of the element in the PDF
+          pos= component[0]
+          # Extract the element of the page layout
+          element = component[1]
+
+          # Check if the element is a text element
+          if isinstance(element, LTTextContainer):
+              # Check if the text appeared in a table
+              if table_extraction_flag == False:
+                  # Use the function to extract the text and format for each text element
+                  (line_text, format_per_line) = text_extraction(element)
+                  # Append the text of each line to the page text
+                  page_text.append(line_text)
+                  # Append the format for each line containing text
+                  line_format.append(format_per_line)
+                  page_content.append(line_text)
+              else:
+                  # Omit the text that appeared in a table
+                  pass
+
+          # Check the elements for images
+          if isinstance(element, LTFigure):
+              # Crop the image from the PDF
+              crop_image(element, pageObj)
+              # Convert the cropped pdf to an image
+              convert_to_images('cropped_image.pdf')
+              # Extract the text from the image
+              image_text = image_to_text('PDF_image.png')
+              text_from_images.append(image_text)
+              page_content.append(image_text)
+              # Add a placeholder in the text and format lists
+              page_text.append('image')
+              line_format.append('image')
+
+          # Check the elements for tables
+          if isinstance(element, LTRect):
+              # If the first rectangular element
+              if first_element == True and (table_num+1) <= len(tables):
+                  # Find the bounding box of the table
+                  lower_side = page.bbox[3] - tables[table_num].bbox[3]
+                  upper_side = element.y1
+                  # Extract the information from the table
+                  table = extract_table(pdf_path, pagenum, table_num)
+                  # Convert the table information in structured string format
+                  table_string = table_converter(table)
+                  # Append the table string into a list
+                  text_from_tables.append(table_string)
+                  page_content.append(table_string)
+                  # Set the flag as True to avoid the content again
+                  table_extraction_flag = True
+                  # Make it another element
+                  first_element = False
+                  # Add a placeholder in the text and format lists
+                  page_text.append('table')
+                  line_format.append('table')
+
+                  # Check if we already extracted the tables from the page
+                  if element.y0 >= lower_side and element.y1 <= upper_side:
+                      pass
+                  elif not isinstance(page_elements[i+1][1], LTRect):
+                      table_extraction_flag = False
+                      first_element = True
+                      table_num+=1
+
+
+      # Create the key of the dictionary
+      dctkey = 'Page_'+str(pagenum)
+      # Add the list of list as the value of the page key
+      text_per_page[dctkey]= [page_text, line_format, text_from_images,text_from_tables, page_content]
+
+  # Closing the pdf file object
+  pdfFileObj.close()
+
+  # Deleting the additional files created
+  #os.remove('cropped_image.pdf')
+  #os.remove('PDF_image.png')
+  return text_per_page
+
+#google drive
+#from google.colab import drive
+#drive.mount('/content/drive')
+#read PDF
+
+pdf_path = 'test.pdf' #article 11
+#pdf_path = 'https://huggingface.co/spaces/Mishmosh/MichelleAssessment3/blob/main/test.pdf' #article 11
+
+text_per_page = read_pdf(pdf_path)
+
+# This section finds the abstract. My plan was to find the end of the abstract by identifying the same font size as the text 'abstract', but it was too late
+#to try this here since the formatting of the text has already been removed.
+#  Instead I extracted just one paragraph. If an abstract is more than 1 paragraph this will not extract the entire abstract
+abstract_from_pdf='' # define empty variable that will hold the text from the abstract
+found_abstract=False # has the abstract been found
+for key in text_per_page.keys(): # go through keys in dictionary
+  current_item=text_per_page[key] #current key
+  for paragraphs in current_item: #go through each item
+    for index,paragraph in enumerate(paragraphs): #go through each line
+      if 'Abstract\n' == paragraph: #does line match paragraph
+        found_abstract=True #word abstract has been found
+        abstract_from_pdf=paragraphs[index+1] #get next paragraph
+    if found_abstract: #if abstract found
+      break
+print(abstract_from_pdf)
+
+from transformers import pipeline
+summarizer = pipeline("summarization", model="ainize/bart-base-cnn")
+#summarizer = pipeline("summarization", model="linydub/bart-large-samsum") # various models were tried and the best one was selected
+#summarizer = pipeline("summarization", model="slauw87/bart_summarisation")
+#summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+#summarizer = pipeline("summarization", model="google/pegasus-cnn_dailymail")
+#print(summarizer(abstract_from_pdf, max_length=50, min_length=5, do_sample=False))
+summarized_text=(summarizer(abstract_from_pdf))
+print(summarized_text)
+#summary_of_abstract=str(summarizer)
+#type(summary_of_abstract)
+#print(summary_of_abstract)
+
+# the aim of this section of code is to get a summary of just one sentence by summarizing the summary all while the summary is longer than one sentence.
+# unfortunately, I tried many many models and none of them actually summarize the text to as short as one sentence.
+#I had searched for ways to fine tune the summarization model to specify that the summarization should be done in just one sentence but did not find a way to implement it
+from transformers import pipeline
+summarized_text_list_list=summarized_text_list['summary_text']
+summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+#print(summarizer)
+number_of_sentences=summarized_text_list_list.count('.')
+print(number_of_sentences)
+while(number_of_sentences)>1:
+    print(number_of_sentences)
+    summarized_text_list_list=summarizer(summarized_text_list_list)[0]['summary_text']
+    number_of_sentences-=1
+    print(summarized_text_list_list)
+    print(number_of_sentences)
+
+
 #text to speech
 #!pip install git+https://github.com/huggingface/transformers.git
 #!pip install datasets sentencepiece
@@ -10,7 +289,7 @@ from transformers import pipeline
 from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech
 processor = SpeechT5Processor.from_pretrained("microsoft/speecht5_tts")
 model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts")
-text = "The future belongs to those who believe in the beauty of their dreams."
+#text = "The future belongs to those who believe in the beauty of their dreams."
 #text = (summarized_text_list_list)
 
 inputs = processor(text=summarized_text_list_list, return_tensors="pt")