PDFMarkup / app.py
gschleusner's picture
Back to temp PDf storage
961d741
import fitz # PyMuPDF
import gradio as gr
import zipfile
import io
import os
from datetime import datetime
import tempfile
# Helper function to find color areas
def find_color_areas(page, target_color, tolerance=30):
pix = page.get_pixmap()
width, height = pix.width, pix.height
visited = [[False for _ in range(width)] for _ in range(height)]
rectangles = []
def flood_fill(x, y):
stack = [(x, y)]
rects = []
while stack:
cx, cy = stack.pop()
if visited[cy][cx]:
continue
visited[cy][cx] = True
pixel_color = pix.pixel(cx, cy)
r, g, b = pixel_color[:3]
if (abs(r - target_color[0]) <= tolerance and
abs(g - target_color[1]) <= tolerance and
abs(b - target_color[2]) <= tolerance):
rects.append(fitz.Rect(cx, cy, cx + 1, cy + 1))
if cx > 0: stack.append((cx - 1, cy))
if cx < width - 1: stack.append((cx + 1, cy))
if cy > 0: stack.append((cx, cy - 1))
if cy < height - 1: stack.append((cx, cy + 1))
if rects:
bbox = fitz.Rect(min([r.x0 for r in rects]),
min([r.y0 for r in rects]),
max([r.x1 for r in rects]),
max([r.y1 for r in rects]))
return bbox
return None
for y in range(height):
for x in range(width):
if not visited[y][x]:
bbox = flood_fill(x, y)
if bbox:
rectangles.append(bbox)
return rectangles
def merge_overlapping_rectangles(rectangles):
merged_rects = []
while rectangles:
rect = rectangles.pop(0)
to_merge = [rect]
for other in rectangles[:]:
if rect.intersects(other):
to_merge.append(other)
rectangles.remove(other)
merged_rect = fitz.Rect(
min([r.x0 for r in to_merge]),
min([r.y0 for r in to_merge]),
max([r.x1 for r in to_merge]),
max([r.y1 for r in to_merge])
)
merged_rects.append(merged_rect)
return merged_rects
def markup_color_regions(doc, color_comment_pairs, tolerance=30, progress=None):
for page_num in range(len(doc)):
page = doc[page_num]
progress((page_num + 1) / len(doc)) # Update the progress as each page is processed
for color_comment_pair in color_comment_pairs:
target_color = color_comment_pair['color']
comment = color_comment_pair['comment']
stroke_color = color_comment_pair['stroke_color']
rectangles = find_color_areas(page, target_color, tolerance)
if rectangles:
merged_rectangles = merge_overlapping_rectangles(rectangles)
for bbox in merged_rectangles:
annot = page.add_rect_annot(bbox)
annot.set_colors(stroke=stroke_color)
annot.set_border(width=2)
annot.set_info({"title": "Markup", "content": comment})
annot.update()
def process_pdf_files(input_pdfs, selected_color_comment_indices, tolerance, custom_color, custom_comment, custom_stroke_color, progress=gr.Progress()):
color_comment_pairs = [
{
"color": (235, 128, 138),
"comment": "Structural Slab greater than architectural slab",
"stroke_color": (1, 0, 0)
},
{
"color": (128, 253, 128),
"comment": "Arch Slab greater than Structure",
"stroke_color": (0, 1, 0)
}
]
# Add custom color-comment pair if provided
if custom_color and custom_comment and custom_stroke_color:
custom_color_tuple = tuple(map(int, custom_color.split(','))) # Convert color to tuple
custom_stroke_tuple = tuple(map(int, custom_stroke_color.split(','))) # Convert stroke to tuple
color_comment_pairs.append({
"color": custom_color_tuple,
"comment": custom_comment,
"stroke_color": custom_stroke_tuple
})
# Get the selected color-comment pairs
selected_color_comment_pairs = [color_comment_pairs[i] for i in selected_color_comment_indices]
# List to keep track of all modified PDFs (in-memory)
modified_pdfs = []
# Process each input PDF file
for i, pdf_file in enumerate(input_pdfs):
with open(pdf_file.name, "rb") as file_stream:
doc = fitz.open(stream=file_stream.read(), filetype="pdf")
markup_color_regions(doc, selected_color_comment_pairs, tolerance, progress=lambda p: progress((i + p) / len(input_pdfs)))
# Save the modified PDF in memory
pdf_in_memory = io.BytesIO()
doc.save(pdf_in_memory)
doc.close()
# Move pointer to the beginning of the BytesIO object
pdf_in_memory.seek(0)
modified_pdfs.append((os.path.basename(pdf_file.name), pdf_in_memory))
# Create an in-memory ZIP file containing all modified PDFs
zip_in_memory = io.BytesIO()
with zipfile.ZipFile(zip_in_memory, 'w') as zipf:
for pdf_name, pdf_bytes in modified_pdfs:
zipf.writestr(pdf_name, pdf_bytes.read())
# Move pointer to the beginning of the ZIP file BytesIO object
zip_in_memory.seek(0)
# Generate the current date string
current_date = datetime.now().strftime("%Y-%m-%d")
# Create a named temporary file to store the ZIP
temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=f"_CoordinationPDFS_{current_date}.zip")
with open(temp_file.name, 'wb') as tmp_file:
tmp_file.write(zip_in_memory.read())
return temp_file.name # Return the path to the temp file for Gradio to handle the download
# Define the Gradio interface
interface = gr.Interface(
fn=process_pdf_files,
inputs=[
gr.Files(label="Input PDF Files", file_types=[".pdf"]),
gr.CheckboxGroup( # Use CheckboxGroup for multi-select
label="Select Color-Comment Pairs",
choices=["Structural Slab vs Arch Slab", "Arch Slab vs Structural Slab", "Custom Option"],
type="index"
),
gr.Slider(label="Tolerance", minimum=0, maximum=100, step=1, value=30),
gr.Textbox(label="Custom Color (R,G,B)", placeholder="Enter custom color in RGB format, e.g., 255,0,0"),
gr.Textbox(label="Custom Comment", placeholder="Enter custom comment for this color"),
gr.Textbox(label="Custom Stroke Color (R,G,B)", placeholder="Enter stroke color in RGB format, e.g., 0,0,255")
],
outputs=gr.File(label="Download ZIP File"), # Use gr.File to trigger the download of the ZIP file
title="PDF Color Region Markup"
)
# Launch the Gradio app
interface.launch()