Parsing and Manipulating PDF Documents in Python with PyMuPDF and pdfminer.six#

1
import fitz # imports PyMuPDF
2

3
try:
4
    doc = fitz.open("example.pdf")
5
    print(f"Document opened successfully: {doc.name}")
6
    print(f"Number of pages: {doc.page_count}")
7
    print(f"Metadata: {doc.metadata}")
8
except fitz.FileDataError:
9
    print("Error: Cannot open the file or it is not a valid PDF.")
10
except Exception as e:
11
    print(f"An unexpected error occurred: {e}")
12
finally:
13
    # Documents should be closed when done
14
    if 'doc' in locals() and doc:
15
        doc.close()

1
doc = fitz.open("example.pdf")
2
if doc.page_count > 0:
3
    page = doc.load_page(0) # Load the first page (page index is 0-based)
4
    text = page.get_text()
5
    print("--- Text from page 1 ---")
6
    print(text[:500] + "..." if len(text) > 500 else text) # Print first 500 chars
7
else:
8
    print("Document is empty.")
9
doc.close()

1
doc = fitz.open("example.pdf")
2
if doc.page_count > 0:
3
    page = doc.load_page(0)
4
    # Get text blocks (paragraphs, etc.) with coordinates
5
    blocks = page.get_text("blocks")
6
    print("\n--- Text Blocks (Page 1) ---")
7
    for b in blocks:
8
        # b is a tuple: (x0, y0, x1, y1, text, block_no, block_type)
9
        # block_type 0: text, 1: image
10
        if b[6] == 0: # Check if it's a text block
11
            print(f"Block @ {b[0]:.0f},{b[1]:.0f} to {b[2]:.0f},{b[3]:.0f}:\n{b[4].strip()}\n")
12

13
    # Get words with coordinates
14
    words = page.get_text("words")
15
    print("\n--- Sample Words (Page 1) ---")
16
    # words are tuples: (x0, y0, x1, y1, word, block_no, line_no, word_no)
17
    for w in words[:10]: # Print first 10 words
18
         print(f"'{w[4]}' @ {w[0]:.0f},{w[1]:.0f}")
19
else:
20
     print("Document is empty.")
21
doc.close()

1
doc = fitz.open("example.pdf")
2
print(f"\n--- Extracting Images from {doc.name} ---")
3
image_list = doc.get_page_images(0) # Get images from the first page
4

5
if not image_list:
6
    print("No images found on page 1.")
7
else:
8
    print(f"Found {len(image_list)} images on page 1.")
9
    for img_index, img_info in enumerate(image_list):
10
        xref = img_info[0] # xref is the object ID
11
        pix = fitz.Pixmap(doc, xref) # Create a Pixmap object
12

13
        # Save the image
14
        output_filename = f"image_page1_{img_index+1}.png"
15
        if pix.n - pix.alpha < 4: # Check if grayscale or RGB
16
            pix.save(output_filename)
17
        else: # Handle CMYK or other color spaces by converting to RGB
18
            pix = fitz.Pixmap(fitz.csRGB, pix)
19
            pix.save(output_filename)
20

21
        print(f"Saved image {img_index+1} as {output_filename}")
22
        pix = None # Release memory
23

24
doc.close()

1
doc = fitz.open("example.pdf")
2
search_term = "document"
3
print(f"\n--- Searching for '{search_term}' in {doc.name} ---")
4

5
for page_num in range(doc.page_count):
6
    page = doc.load_page(page_num)
7
    text_instances = page.search_for(search_term)
8
    if text_instances:
9
        print(f"Found '{search_term}' on page {page_num + 1} at locations:")
10
        for inst in text_instances:
11
            # inst is a fitz.Rect object (x0, y0, x1, y1)
12
            print(f"  {inst}")
13

14
doc.close()

1
doc = fitz.open("example.pdf")
2
if doc.page_count > 0:
3
    page = doc.load_page(0)
4
    # Define point (x, y) where text will be added
5
    point = fitz.Point(50, 50)
6
    # Add text
7
    page.insert_text(point, "This is a sample text added by PyMuPDF.", fontsize=12, color=(0, 0, 1)) # Blue color
8
    print("\nAdded text to page 1.")
9

10
    # Save the modified document (create a new file)
11
    output_pdf = "example_modified_pymupdf.pdf"
12
    doc.save(output_pdf)
13
    print(f"Modified document saved as {output_pdf}")
14
else:
15
    print("Document is empty, cannot add text.")
16
doc.close()

1
from pdfminer.high_level import extract_text
2

3
try:
4
    text = extract_text("example.pdf")
5
    print("--- Text from example.pdf (basic) ---")
6
    print(text[:500] + "..." if len(text) > 500 else text) # Print first 500 chars
7
except Exception as e:
8
    print(f"Error extracting text with pdfminer.six: {e}")

1
from io import StringIO
2
from pdfminer.high_level import extract_text_to_fp
3
from pdfminer.layout import LAParams
4

5
output_string = StringIO()
6
# LAParams() contains parameters for layout analysis
7
# common_param = LAParams(all_texts=True, detect_vertical=True) # More detailed analysis
8
common_param = LAParams() # Default parameters often suffice
9

10
try:
11
    with open("example.pdf", 'rb') as in_file:
12
        # Redirect output to StringIO object
13
        extract_text_to_fp(in_file, output_string, laparams=common_param)
14

15
    text_with_layout = output_string.getvalue()
16
    print("\n--- Text from example.pdf (with layout) ---")
17
    print(text_with_layout[:500] + "..." if len(text_with_layout) > 500 else text_with_layout)
18

19
except Exception as e:
20
    print(f"Error extracting text with pdfminer.six layout: {e}")

1
from pdfminer.pdfparser import PDFParser
2
from pdfminer.pdfdocument import PDFDocument
3
from pdfminer.pdfpage import PDFPage
4
from pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter
5
from pdfminer.converter import PDFPageAggregator
6
from pdfminer.layout import LAParams, LTContainer, LTTextBox, LTTextLine, LTChar
7

8
def parse_layout(layout):
9
    """Recursively parse layout elements."""
10
    for lt_obj in layout:
11
        if isinstance(lt_obj, LTTextBox):
12
            print(f"TextBox: {lt_obj.bbox} Text: {lt_obj.get_text().strip()}")
13
            # Further parse lines within the text box
14
            parse_layout(lt_obj)
15
        elif isinstance(lt_obj, LTTextLine):
16
             print(f"  TextLine: {lt_obj.bbox} Text: {lt_obj.get_text().strip()}")
17
             # Further parse characters within the line
18
             parse_layout(lt_obj)
19
        elif isinstance(lt_obj, LTChar):
20
            # Access character details
21
            # print(f"    Char: {lt_obj.bbox} '{lt_obj.get_text()}' Font: {lt_obj.fontname}")
22
            pass # Often too verbose, comment out or process as needed
23
        elif isinstance(lt_obj, LTContainer):
24
            # Recursively process containers (pages, figures, text boxes, lines)
25
            parse_layout(lt_obj)
26

27
print("\n--- Detailed Layout Parsing (Page 1) ---")
28
try:
29
    with open("example.pdf", 'rb') as in_file:
30
        # Create a PDF parser object associated with the file object.
31
        parser = PDFParser(in_file)
32
        # Create a PDF document object that stores the document structure.
33
        document = PDFDocument(parser)
34
        # Check if the document allows text extraction.
35
        if not document.is_extractable:
36
            print("Document is not extractable.")
37
        else:
38
            # Create a PDF resource manager object that stores shared resources.
39
            rsrcmgr = PDFResourceManager()
40
            # Set parameters for layout analysis.
41
            laparams = LAParams()
42
            # Create a PDF page aggregator object.
43
            device = PDFPageAggregator(rsrcmgr, laparams=laparams)
44
            # Create a PDF interpreter object.
45
            interpreter = PDFPageInterpreter(rsrcmgr, device)
46

47
            # Process pages
48
            for i, page in enumerate(PDFPage.create_pages(document)):
49
                if i == 0: # Process only the first page for demonstration
50
                    interpreter.process_page(page)
51
                    # Receive the layout layout of the page.
52
                    layout = device.get_result()
53
                    # Parse the layout element by element
54
                    parse_layout(layout)
55
                    break # Stop after the first page
56

57
except Exception as e:
58
    print(f"Error during detailed layout parsing: {e}")