pdf2txt_parser_converter / parser_sevenof9_v2_en.py
kalle07's picture
only commented out much more
3829982 verified
import os # OS module for interacting with the operating system (file management, etc.)
import sys # Provides access to system-specific parameters and functions
import tkinter as tk # GUI module for creating desktop applications
from tkinter import filedialog, messagebox # Additional tkinter components for file dialogs and message boxes
import subprocess # Module to run system commands
import threading # Threading module to run tasks concurrently
import tempfile # Module to create temporary files and directories
import shutil # Module for file operations like copy, move, and delete
import json # JSON module for working with JSON data
import logging # Logging module for tracking events and errors
import pdfplumber # Library for extracting text and tables from PDFs
from pdfplumber.utils import get_bbox_overlap, obj_to_bbox # Helper functions from pdfplumber for working with bounding boxes
from pdfplumber.utils.exceptions import PdfminerException # Exception related to PDF processing
from joblib import delayed, cpu_count, parallel_backend, Parallel # Joblib for parallel processing and optimization
import multiprocessing # Module for parallel processing using multiple CPU cores
from multiprocessing import Pool # Pool class for parallelizing tasks across multiple processes
# ========================
# Parser Configuration
# ========================
TEXT_EXTRACTION_SETTINGS = {
"x_tolerance": 1, # Horizontal tolerance for text extraction
"y_tolerance": 3, # Vertical tolerance for text extraction
"keep_blank_chars": False, # Option to retain blank characters in the extracted text
"use_text_flow": True # Option to use text flow for better structure
}
# Suppress stderr output on Windows platform to avoid cluttering the console
if sys.platform == "win32":
sys.stderr = open(os.devnull, 'w') # Redirect stderr to null
PARALLEL_THRESHOLD = 16 # Number of pages to use for deciding between serial or parallel processing
# Function to suppress PDFMiner logging, reducing verbosity
def suppress_pdfminer_logging():
for logger_name in [
"pdfminer", # Various pdfminer modules to suppress logging from
"pdfminer.pdfparser",
"pdfminer.pdfdocument",
"pdfminer.pdfpage",
"pdfminer.converter",
"pdfminer.layout",
"pdfminer.cmapdb",
"pdfminer.utils"
]:
logging.getLogger(logger_name).setLevel(logging.ERROR) # Set logging level to ERROR to suppress lower levels
# Function to clean up text by removing unwanted hyphenations and newlines
def clean_cell_text(text):
if not isinstance(text, str): # If text is not a string, return empty string
return ""
text = text.replace("-\n", "").replace("\n", " ") # Remove hyphenated line breaks and replace newlines with space
return " ".join(text.split()) # Split text into words and join with single spaces
# Function to safely clean and join row cell data
def safe_join(row):
return [clean_cell_text(str(cell)) if cell is not None else "" for cell in row] # Clean each cell in the row, or return empty if None
# Function to clamp bounding box coordinates within page boundaries
def clamp_bbox(bbox, page_width, page_height):
x0, top, x1, bottom = bbox # Extract bounding box coordinates
# Ensure each coordinate is within the page width and height limits
x0 = max(0, min(x0, page_width))
x1 = max(0, min(x1, page_width))
top = max(0, min(top, page_height))
bottom = max(0, min(bottom, page_height))
return (x0, top, x1, bottom) # Return the adjusted bounding box
# Function to process a single PDF page
def process_page(args):
suppress_pdfminer_logging() # Suppress unnecessary PDFMiner logging
try:
page_number, pdf_path, text_settings = args # Extract page number, PDF path, and text extraction settings
with pdfplumber.open(pdf_path) as pdf: # Open the PDF using pdfplumber
page = pdf.pages[page_number] # Get the specific page
output = f"Page {page_number + 1}\n" # Add page number to the output
width, height = page.width, page.height # Get page dimensions
filtered_page = page # Initialize filtered page
table_bboxes = [] # List to hold bounding boxes of tables
table_json_outputs = [] # List to hold JSON output of tables
# Iterate through all tables found on the page
for table in page.find_tables():
bbox = clamp_bbox(table.bbox, width, height) # Adjust the bounding box to fit within the page
table_bboxes.append(bbox) # Add the bounding box to the list
if not page.crop(bbox).chars: # Skip tables that have no characters
continue
# Filter out any elements that overlap with the table's bounding box
filtered_page = filtered_page.filter(
lambda obj: get_bbox_overlap(obj_to_bbox(obj), bbox) is None
)
# Extract the table data and structure it
table_data = table.extract()
if table_data and len(table_data) >= 1: # Ensure there is data in the table
headers = safe_join(table_data[0]) # Clean and join the headers
rows = [safe_join(row) for row in table_data[1:]] # Clean and join the table rows
json_table = [dict(zip(headers, row)) for row in rows] # Create a JSON object from headers and rows
table_json_outputs.append(json.dumps(json_table, indent=1, ensure_ascii=False)) # Convert table data to JSON
# Extract words outside the tables
words_outside_tables = [
word for word in page.extract_words(**text_settings) # Extract words from the page using the settings
if not any(
bbox[0] <= float(word['x0']) <= bbox[2] and
bbox[1] <= float(word['top']) <= bbox[3]
for bbox in table_bboxes # Ensure word is not inside any table bounding box
)
]
current_y = None # Track vertical position of words
line = [] # List to hold words for the current line
text_content = "" # Store the extracted text content
# Iterate through words and group them into lines
for word in words_outside_tables:
if current_y is None or abs(word['top'] - current_y) > 10: # Start a new line if Y position changes significantly
if line: # If there's a previous line, join and add it to text content
text_content += " ".join(line) + "\n"
line = [word['text']] # Start a new line with the current word
current_y = word['top'] # Update the current Y position
else:
line.append(word['text']) # Append the word to the current line
if line: # Add the last line to the text content
text_content += " ".join(line) + "\n"
output += text_content.strip() + "\n" # Add the final text content for the page
# Add table JSON outputs to the page output
for idx, table in enumerate(table_json_outputs, start=1):
output += f'"table {idx}":\n{table}\n'
return page_number, output # Return the processed page number and output content
except Exception as e:
return args[0], f"[ERROR] Page {args[0]+1} ({args[1]}): {str(e)}" # Return an error message if an exception occurs
# Function to process the entire PDF document
def process_pdf(pdf_path):
suppress_pdfminer_logging() # Suppress unnecessary logging
try:
if not os.path.exists(pdf_path): # Check if the file exists
return f"[ERROR] File not found: {pdf_path}" # Return error message if file does not exist
print(f"[INFO] Starting processing: {pdf_path}") # Log the start of processing
try:
with pdfplumber.open(pdf_path) as pdf: # Open the PDF using pdfplumber
num_pages = len(pdf.pages) # Get the number of pages in the PDF
except PdfminerException as e:
return f"[ERROR] Cannot open PDF: {pdf_path}{str(e)}" # Return error if the PDF cannot be opened
except Exception as e:
return f"[ERROR] General error opening PDF: {pdf_path}{str(e)}" # Return general error if any exception occurs
pages = [(i, pdf_path, TEXT_EXTRACTION_SETTINGS) for i in range(num_pages)] # Prepare the pages for processing
try:
results = run_serial(pages) if num_pages <= PARALLEL_THRESHOLD else run_parallel(pages) # Run serial or parallel processing
except (EOFError, BrokenPipeError, KeyboardInterrupt):
return "[INFO] Processing was interrupted." # Handle interruptions during processing
sorted_results = sorted(results, key=lambda x: x[0]) # Sort results by page number
final_output = "\n".join(text for _, text in sorted_results) # Combine all page results into a single string
base_name = os.path.splitext(os.path.basename(pdf_path))[0] # Get the base name of the PDF file
output_dir = os.path.dirname(pdf_path) # Get the directory of the PDF file
output_path = os.path.join(output_dir, f"{base_name}.txt") # Generate the output file path
with open(output_path, "w", encoding="utf-8", errors="ignore") as f: # Open the output file for writing
f.write(final_output) # Write the final output to the file
print(f"[INFO] Processing complete: {output_path}") # Log the successful processing completion
except (EOFError, BrokenPipeError, KeyboardInterrupt):
return "[INFO] Processing interrupted by user." # Handle user interruptions
except Exception as e:
return f"[ERROR] Unexpected error with '{pdf_path}': {str(e)}" # Handle unexpected errors during processing
# Function to run the PDF processing serially (one page at a time)
def run_serial(pages):
return [process_page(args) for args in pages] # Process each page in sequence
# Function to run the PDF processing in parallel (across multiple cores)
def run_parallel(pages):
available_cores = max(1, cpu_count() - 2) # Calculate the number of available CPU cores, leaving 2 for system processes
num_cores = min(available_cores, len(pages)) # Limit the number of cores based on the number of pages
print(f"Starting parallel processing with {num_cores} cores...") # Log the number of cores used
with Pool(processes=num_cores) as pool: # Create a pool of processes
return pool.map(process_page, pages) # Distribute the page processing across the available cores
# Main function to process a list of PDFs
def process_pdfs_main():
suppress_pdfminer_logging() # Suppress unnecessary logging
pdf_files = sys.argv[1:] # Get PDF file paths from command-line arguments
if not pdf_files: # Check if any PDFs are provided
print("No PDF files provided.") # Log message if no PDFs are provided
return
small_pdfs = [] # List to store small PDFs (less than the parallel threshold)
large_pdfs = [] # List to store large PDFs (greater than the parallel threshold)
# Categorize PDFs into small and large based on the number of pages
for path in pdf_files:
if not os.path.exists(path): # Check if the file exists
print(f"File not found: {path}") # Log error if file does not exist
continue
try:
with pdfplumber.open(path) as pdf: # Open the PDF
if len(pdf.pages) <= PARALLEL_THRESHOLD: # If the PDF has fewer pages than the threshold
small_pdfs.append(path) # Add to small PDFs list
else:
large_pdfs.append(path) # Add to large PDFs list
except PdfminerException:
print(f"[ERROR] Password-protected PDF skipped: {path}") # Log if the PDF is password-protected
except Exception as e:
print(f"[ERROR] Error opening {path}: {str(e)}") # Log any other errors when opening the PDF
# Process small PDFs in parallel (if there are any)
if small_pdfs:
available_cores = max(1, cpu_count() - 2) # Determine the number of available cores
num_cores = min(available_cores, len(small_pdfs)) # Use the lesser of available cores or small PDFs count
print(f"\n[Phase 1] Starting parallel processing of small PDFs with {num_cores} cores...") # Log processing start
results = Parallel(n_jobs=num_cores)( # Run parallel processing for small PDFs
delayed(process_pdf)(path) for path in small_pdfs
)
for r in results:
print(r) # Print the results for each small PDF
# Process large PDFs one by one (in serial)
for path in large_pdfs:
print(f"\n[Phase 2] Processing large PDF: {os.path.basename(path)}") # Log processing of large PDF
print(process_pdf(path)) # Process the large PDF
# GUI
class FileManager:
def __init__(self, master):
# Initialize the main window and title
self.master = master
self.master.title("Parser-Sevenof9")
# Internal list to track selected PDF files
self.files = []
self.last_selected_index = None # Stores the last clicked index for shift-selection
# Label for file list
self.label = tk.Label(master, text="Selected PDF files:")
self.label.pack(pady=5)
# Frame to contain the listbox and its scrollbar
listbox_frame = tk.Frame(master)
listbox_frame.pack(pady=5)
# Scrollbar for the listbox
scrollbar_listbox = tk.Scrollbar(listbox_frame)
self.listbox = tk.Listbox(
listbox_frame, selectmode=tk.MULTIPLE, width=80, height=6,
yscrollcommand=scrollbar_listbox.set
)
scrollbar_listbox.config(command=self.listbox.yview)
# Pack listbox and scrollbar side by side
self.listbox.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
scrollbar_listbox.pack(side=tk.RIGHT, fill=tk.Y)
# Bind selection and click events for the listbox
self.listbox.bind("<<ListboxSelect>>", self.show_text_file)
self.listbox.bind("<Button-1>", self.on_listbox_click)
self.listbox.bind("<Shift-Button-1>", self.on_listbox_shift_click)
# Create a context menu for right-click actions
self.context_menu = tk.Menu(master, tearoff=0)
self.context_menu.add_command(label="Remove selected", command=self.remove_file)
self.listbox.bind("<Button-3>", self.show_context_menu)
# Frame for action buttons (Add/Remove)
self.frame = tk.Frame(master)
self.frame.pack(pady=10)
# Action buttons
tk.Button(self.frame, text="Add Folder", command=self.add_folder).pack(side=tk.LEFT, padx=5)
tk.Button(self.frame, text="Select Files", command=self.add_file).pack(side=tk.LEFT, padx=5)
tk.Button(self.frame, text="Remove Selected", command=self.remove_file).pack(side=tk.LEFT, padx=5)
tk.Button(self.frame, text="Remove All", command=self.remove_all).pack(side=tk.LEFT, padx=5)
tk.Button(master, text="Stop", command=self.stop_parser).pack(pady=5)
# Placeholder for the parser process (used in threading)
self.parser_process = None
# Start button for parsing process
tk.Button(master, text="Start Parser", command=self.start_parser).pack(pady=10)
# Text frame to display the contents of the selected .txt file
text_frame = tk.Frame(master)
text_frame.pack(padx=10, pady=5)
scrollbar_text = tk.Scrollbar(text_frame)
self.text_widget = tk.Text(
text_frame, height=15, width=100, wrap=tk.WORD,
yscrollcommand=scrollbar_text.set
)
scrollbar_text.config(command=self.text_widget.yview)
# Pack text viewer and scrollbar
self.text_widget.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
scrollbar_text.pack(side=tk.RIGHT, fill=tk.Y)
# Label for progress section
tk.Label(master, text="Progress:").pack()
# Frame for progress output
progress_frame = tk.Frame(master)
progress_frame.pack(padx=10, pady=5)
scrollbar_progress = tk.Scrollbar(progress_frame)
self.progress_text = tk.Text(
progress_frame, height=8, width=100, state=tk.DISABLED,
yscrollcommand=scrollbar_progress.set
)
scrollbar_progress.config(command=self.progress_text.yview)
self.progress_text.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
scrollbar_progress.pack(side=tk.RIGHT, fill=tk.Y)
def on_listbox_click(self, event):
# Handle single left-click selection; clear previous selection
index = self.listbox.nearest(event.y)
self.listbox.selection_clear(0, tk.END)
self.listbox.selection_set(index)
self.last_selected_index = index
self.show_text_file(None)
return "break" # Prevent default event propagation
def on_listbox_shift_click(self, event):
# Handle shift-click for range selection
index = self.listbox.nearest(event.y)
if self.last_selected_index is None:
self.last_selected_index = index
start, end = sorted((self.last_selected_index, index))
self.listbox.selection_clear(0, tk.END)
for i in range(start, end + 1):
self.listbox.selection_set(i)
return "break"
def show_context_menu(self, event):
# Show right-click context menu if any item is selected
if self.listbox.curselection():
self.context_menu.tk_popup(event.x_root, event.y_root)
def add_folder(self):
# Add all PDFs from a selected folder
folder = filedialog.askdirectory(title="Select Folder")
if not folder:
return
for root, _, files in os.walk(folder):
for file in files:
if file.lower().endswith(".pdf"):
path = os.path.join(root, file)
if path not in self.files:
self.files.append(path)
self.listbox.insert(tk.END, path)
def add_file(self):
# Add selected individual PDF files
paths = filedialog.askopenfilenames(title="Select PDF Files", filetypes=[("PDF Files", "*.pdf")])
for path in paths:
if path not in self.files:
self.files.append(path)
self.listbox.insert(tk.END, path)
def remove_file(self):
# Remove selected files from list and internal storage
selection = self.listbox.curselection()
if not selection:
messagebox.showwarning("Notice", "Please select an entry to remove.")
return
for index in reversed(selection): # Reverse to avoid index shifting
self.listbox.delete(index)
del self.files[index]
self.text_widget.delete(1.0, tk.END)
def remove_all(self):
# Remove all files from the list
self.listbox.delete(0, tk.END)
self.files.clear()
self.text_widget.delete(1.0, tk.END)
def start_parser(self):
# Validate input and launch parser in separate thread
if not self.files:
messagebox.showinfo("No Files", "Please select at least one file.")
return
self.progress_text.config(state=tk.NORMAL)
self.progress_text.delete(1.0, tk.END)
self.progress_text.insert(tk.END, "Starting parser...\n")
self.progress_text.config(state=tk.DISABLED)
# Launch parsing in background to avoid UI freeze
thread = threading.Thread(target=self.run_parser)
thread.start()
def stop_parser(self):
# Terminate running parser process if active
if self.parser_process and self.parser_process.poll() is None:
self.parser_process.terminate()
self.append_progress_text("Parser process was stopped.\n")
else:
self.append_progress_text("No active parser process to stop.\n")
def run_parser(self):
# Internal method to run the external parser script
try:
self.parser_process = subprocess.Popen(
[sys.executable, __file__] + self.files,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
text=True,
encoding='utf-8',
errors='ignore',
bufsize=4096
)
for line in self.parser_process.stdout:
self.append_progress_text(line)
self.parser_process.stdout.close()
self.parser_process.wait()
if self.parser_process.returncode == 0:
self.append_progress_text("\nParser finished successfully.\n")
self.show_messagebox_threadsafe("Parser Done", "The parser was executed successfully.")
else:
self.append_progress_text("\nError while running the parser.\n")
self.show_messagebox_threadsafe("Error", "Error while running the parser.")
except Exception as e:
self.append_progress_text(f"Error: {e}\n")
self.show_messagebox_threadsafe("Error", f"Error during execution:\n{e}")
finally:
self.parser_process = None
def append_progress_text(self, text):
# Thread-safe method to append text to the progress view
self.progress_text.after(0, lambda: self._insert_text(text))
def _insert_text(self, text):
# Append text and scroll to bottom
self.progress_text.config(state=tk.NORMAL)
self.progress_text.insert(tk.END, text)
self.progress_text.see(tk.END)
self.progress_text.config(state=tk.DISABLED)
def show_messagebox_threadsafe(self, title, message):
# Display a messagebox from a background thread
self.master.after(0, lambda: messagebox.showinfo(title, message))
def show_text_file(self, event):
# Load and show the content of the corresponding .txt file (if available)
selection = self.listbox.curselection()
if not selection:
return
index = selection[0]
path = self.files[index]
txt_path = os.path.splitext(path)[0] + ".txt"
self.text_widget.delete(1.0, tk.END)
if os.path.exists(txt_path):
try:
with open(txt_path, "r", encoding="utf-8", errors="ignore") as f:
self.text_widget.insert(tk.END, f.read())
except Exception as e:
self.text_widget.insert(tk.END, f"Error loading text file:\n{e}")
else:
self.text_widget.insert(tk.END, "[No corresponding .txt file found]")
# MAIN
if __name__ == "__main__":
multiprocessing.freeze_support() # Required for Windows compatibility with multiprocessing
if len(sys.argv) > 1:
# If called with file arguments, execute parsing logic (e.g., from subprocess)
process_pdfs_main()
else:
# Otherwise, launch the GUI application
root = tk.Tk()
app = FileManager(root)
root.mainloop()