Upload utils.py

utils.py

@@ -0,0 +1,223 @@
+# -*- coding: utf-8 -*-
+# Use ReportLab package to create PDF poster
+from reportlab.pdfbase import pdfmetrics
+from reportlab.lib.pagesizes import A4
+from reportlab.lib.styles import getSampleStyleSheet
+from reportlab.lib.units import cm
+from reportlab.platypus import (
+    SimpleDocTemplate,
+    Paragraph,
+    Spacer,
+    Table,
+    TableStyle,
+    Image,
+    Flowable,
+    ListFlowable, 
+    ListItem,
+)
+from reportlab.lib import colors
+from reportlab.pdfbase.cidfonts import UnicodeCIDFont
+
+import yt_dlp
+import cv2
+from PIL import Image as PILImage
+
+import os
+import tempfile
+import re
+import uuid
+
+import pymupdf
+
+'''
+# UnicodeCIDfont names
+$chs$ = Chinese Simplified (mainland): '$STSong-Light$'
+$cht$ = Chinese Traditional (Taiwan): '$MSung-Light$', '$MHei-Medium$'
+$kor$ = Korean: '$HYSMyeongJoStd-Medium$','$HYGothic-Medium$'
+$jpn$ = Japanese: '$HeiseiMin-W3$', '$HeiseiKakuGo-W5$'
+'''
+
+# Configuration
+# Register the Chinese font with Reportlab
+pdfmetrics.registerFont(UnicodeCIDFont('STSong-Light'))
+pdfmetrics.registerFont(UnicodeCIDFont('MSung-Light'))
+
+PAGE_SIZE = A4
+MARGIN = 1.0 * cm
+COLUMNS = 3  # Now using 3 columns
+STYLE = getSampleStyleSheet()
+style_body = STYLE["BodyText"]
+style_title = STYLE['Title']  
+style_title.alignment = 1   # center the title
+
+# Calculate available width for tables
+page_width = PAGE_SIZE[0] - 2*MARGIN
+col_width = page_width / COLUMNS
+img_width = col_width - 1*cm  # Leave some padding
+img_height = 5*cm
+#==========================================================================
+def create_poster(filename, images, lang, summary, url = None):
+    print("Output language is:", lang)       
+    #generate PDF file
+    doc = SimpleDocTemplate(filename, pagesize=PAGE_SIZE,
+                           leftMargin=MARGIN, rightMargin=MARGIN,
+                           topMargin=MARGIN, bottomMargin=MARGIN)
+                            
+    story = []
+    
+    # Define a style with the detected language font
+    if lang.lower() == 'chinese':
+        style_body.fontName = 'STSong-Light'
+        style_title.fontName = 'STSong-Light'
+    else:
+        style_body.fontName = 'Helvetica'
+        style_title.fontName = 'Helvetica-Bold'
+   
+    # Create table data for detected images
+    table_data = []
+    list_content = []
+    
+    with tempfile.TemporaryDirectory() as temp_dir: 
+        # Process output summary 
+        question = []
+        current_answer = []
+        answers_part = summary.strip().split("\n")
+        title_text = "Summary" # initialize title
+        title = Paragraph(f"<b>{title_text}</b>", style_title)
+
+        for line in answers_part: #.split("\n"):
+            if re.search("0.", line):  #title line
+                clean_line = line.replace("*", "")
+                title_text = clean_line.split("0.")
+                if len(title_text) > 1:
+                    title_text = title_text[1]
+                else:
+                    title_text = title_text[0]
+    
+                index = title_text.find(':')
+                if index != -1:
+                    title_text = title_text[index+1:]
+                title = Paragraph(f"<b>{title_text}</b>", style_title) 
+                
+            elif re.search(r'\d\.', line): 
+                # Start of a new question-answer section
+                list_content.append(Spacer(1, 0.3*cm))
+                if current_answer:
+                    list_item = Paragraph(f"<b>{current_answer}</b>", style_body),
+                    list_content.append(list_item)
+                    current_answer = []
+                    
+                line_content = line.replace("*", "").split(":")
+                question = str(line_content[0]) 
+                
+                question = Paragraph(f"<b>{question}</b>", style_body)
+                list_content.append(question)
+                 
+                if len(line_content) > 1: #handle same line answer
+                    list_item = Paragraph(f"<b>{line_content[1]}</b>", style_body)
+                    list_content.append(list_item)
+            elif line.strip() and question:
+                list_item = Paragraph(f"<b>{line}</b>", style_body)
+                list_content.append(list_item) 
+
+            #add last section
+            if current_answer:
+                list_item = Paragraph(f"<b>{line}</b>", style_body)
+                list_content.append(list_item)
+
+        # construct the full list
+        markdown_output = ListFlowable(list_content,                                 
+                                        bulletType='bullet',            
+                                        bulletColor='white', value='circle'                      
+                                       ),
+        output_list = markdown_output[0]  #workaround for converting tuple to list
+        
+        # Insert images into a table
+        row_cells = []
+        for id, image in enumerate(images):    
+            face_filename = f'{id}.jpg'
+            image_path =  os.path.join(temp_dir, face_filename)
+            cv2.imwrite(image_path, image)
+
+            # Load the image back into memory because Image object needs filepath input
+            pil_img = PILImage.open(image_path)
+                
+            # Create cell content
+            cell_content = [
+                Spacer(1, 0.3*cm),
+                Image(image_path, width=img_width, height=img_height),
+                Spacer(1, 0.3*cm),
+                ]
+            row_cells.append(cell_content)
+            
+        # Add row to table
+        table_data.append(row_cells)
+        
+        # Create table with styling
+        tbl = Table(table_data, 
+                   colWidths=[col_width]*COLUMNS,
+                   rowHeights=img_height+0.5*cm)
+    
+        tbl.setStyle(TableStyle([
+            ('ALIGN', (0,0), (-1,-1), 'CENTER'),
+            ('VALIGN', (0,0), (-1,-1), 'CENTER'),
+            ('PADDING', (0,0), (-1,-1), 10),
+            ('BOX', (0,0), (-1,-1), 0.5, colors.white),
+            ('INNERGRID', (0,0), (-1,-1), 0.5, colors.white),
+        ]))
+                   
+        # add flowables
+        story.append(title)
+        story.append(tbl)
+        story.append(Spacer(1,0.3*cm))
+        story.append(output_list)
+        story.append(Spacer(1,0.5*cm))
+
+        # Clickable video link
+        if url:
+            link_text = f'<link href="{url}"><font color="{colors.blue}"><u>View Original Video</u></font></link>'
+            link_paragraph = Paragraph(link_text, style_body)
+            story.append(link_paragraph)
+ 
+        # build a page
+        doc.build(story)
+
+    return filename
+
+def generate_unique_filename(extension):
+    return f"{uuid.uuid4()}{extension}"
+
+def generate_tmp_filename(basename, extension):
+    return f"{basename}{extension}"
+
+def pdf_to_jpg(pdf_path, output_path):
+    doc = pymupdf.open(pdf_path)
+    page = doc.load_page(0)
+    pix = page.get_pixmap()
+    pix.save(output_path, "JPEG")
+
+    image = PILImage.open(output_path)
+    
+    doc.close()
+
+    return image, output_path
+
+def download_youtube_video(url):
+    """Downloads a Youtube video using yt-dlp."""
+    basename = os.path.basename(url)
+    output_path = generate_tmp_filename(basename, ".mp4")
+    
+    ydl_opts = {
+        'outtmpl': output_path,  # Path where the video will be saved
+        'format': 'best',        # Download the best quality available
+        'cookiefile': 'cookies.txt',  # Path to your cookies file  #JW 20250115
+    }
+    
+    try:
+        with yt_dlp.YoutubeDL(ydl_opts) as ydl:          
+            ydl.download([url])
+        return output_path
+    except Exception as e:
+        print("load yt_dlp:", e)
+        return str(e)
+