--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/tests/test_insertpdf.py	Mon Sep 15 11:37:51 2025 +0200
@@ -0,0 +1,335 @@
+"""
+* Join multiple PDFs into a new one.
+* Compare with stored earlier result:
+    - must have identical object definitions
+    - must have different trailers
+* Try inserting files in a loop.
+"""
+
+import io
+import os
+import re
+import pymupdf
+from pymupdf import mupdf
+
+scriptdir = os.path.abspath(os.path.dirname(__file__))
+resources = os.path.join(scriptdir, "resources")
+
+def approx_parse( text):
+    '''
+    Splits <text> into sequence of (text, number) pairs. Where sequence of
+    [0-9.] is not convertible to a number (e.g. '4.5.6'), <number> will be
+    None.
+    '''
+    ret = []
+    for m in re.finditer('([^0-9]+)([0-9.]*)', text):
+        text = m.group(1)
+        try:
+            number = float( m.group(2))
+        except Exception:
+            text += m.group(2)
+            number = None
+        ret.append( (text, number))
+    return ret
+
+def approx_compare( a, b, max_delta):
+    '''
+    Compares <a> and <b>, allowing numbers to differ by up to <delta>.
+    '''
+    aa = approx_parse( a)
+    bb = approx_parse( b)
+    if len(aa) != len(bb):
+        return 1
+    ret = 1
+    for (at, an), (bt, bn) in zip( aa, bb):
+        if at != bt:
+            break
+        if an is not None and bn is not None:
+            if abs( an - bn) >= max_delta:
+                print( f'diff={an-bn}: an={an} bn={bn}')
+                break
+        elif (an is None) != (bn is None):
+            break
+    else:
+        ret = 0
+    if ret:
+        print( f'Differ:\n    a={a!r}\n    b={b!r}')
+    return ret
+
+
+def test_insert():
+    all_text_original = []  # text on input pages
+    all_text_combined = []  # text on resulting output pages
+    # prepare input PDFs
+    doc1 = pymupdf.open()
+    for i in range(5):  # just arbitrary number of pages
+        text = f"doc 1, page {i}"  # the 'globally' unique text
+        page = doc1.new_page()
+        page.insert_text((100, 72), text)
+        all_text_original.append(text)
+
+    doc2 = pymupdf.open()
+    for i in range(4):
+        text = f"doc 2, page {i}"
+        page = doc2.new_page()
+        page.insert_text((100, 72), text)
+        all_text_original.append(text)
+
+    doc3 = pymupdf.open()
+    for i in range(3):
+        text = f"doc 3, page {i}"
+        page = doc3.new_page()
+        page.insert_text((100, 72), text)
+        all_text_original.append(text)
+
+    doc4 = pymupdf.open()
+    for i in range(6):
+        text = f"doc 4, page {i}"
+        page = doc4.new_page()
+        page.insert_text((100, 72), text)
+        all_text_original.append(text)
+
+    new_doc = pymupdf.open()  # make combined PDF of input files
+    new_doc.insert_pdf(doc1)
+    new_doc.insert_pdf(doc2)
+    new_doc.insert_pdf(doc3)
+    new_doc.insert_pdf(doc4)
+    # read text from all pages and store in list
+    for page in new_doc:
+        all_text_combined.append(page.get_text().replace("\n", ""))
+    # the lists must be equal
+    assert all_text_combined == all_text_original
+
+
+def test_issue1417_insertpdf_in_loop():
+    """Using a context manager instead of explicitly closing files"""
+    f = os.path.join(resources, "1.pdf")
+    big_doc = pymupdf.open()
+    fd1 = os.open( f, os.O_RDONLY)
+    os.close( fd1)
+    for n in range(0, 1025):
+        with pymupdf.open(f) as pdf:
+            big_doc.insert_pdf(pdf)
+        # Create a raw file descriptor. If the above pymupdf.open() context leaks
+        # a file descriptor, fd will be seen to increment.
+        fd2 = os.open( f, os.O_RDONLY)
+        assert fd2 == fd1
+        os.close( fd2)
+    big_doc.close()
+
+
+def _test_insert_adobe():
+    path = os.path.abspath( f'{__file__}/../../../PyMuPDF-performance/adobe.pdf')
+    if not os.path.exists(path):
+        print(f'Not running test_insert_adobe() because does not exist: {os.path.relpath(path)}')
+        return
+    a = pymupdf.Document()
+    b = pymupdf.Document(path)
+    a.insert_pdf(b)
+
+
+def _2861_2871_merge_pdf(content: bytes, coverpage: bytes):
+    with pymupdf.Document(stream=coverpage, filetype="pdf") as coverpage_pdf:
+        with pymupdf.Document(stream=content, filetype="pdf") as content_pdf:
+            coverpage_pdf.insert_pdf(content_pdf)
+            doc = coverpage_pdf.write()
+            return doc
+
+def test_2861():
+    path = os.path.abspath(f'{__file__}/../../tests/resources/test_2861.pdf')
+    with open(path, "rb") as content_pdf:
+        with open(path, "rb") as coverpage_pdf:
+            content = content_pdf.read()
+            coverpage = coverpage_pdf.read()
+            _2861_2871_merge_pdf(content, coverpage)
+
+def test_2871():
+    path = os.path.abspath(f'{__file__}/../../tests/resources/test_2871.pdf')
+    with open(path, "rb") as content_pdf:
+        with open(path, "rb") as coverpage_pdf:
+            content = content_pdf.read()
+            coverpage = coverpage_pdf.read()
+            _2861_2871_merge_pdf(content, coverpage)
+
+
+def test_3789():
+
+    file_path = os.path.abspath(f'{__file__}/../../tests/resources/test_3789.pdf')
+    result_path = os.path.abspath(f'{__file__}/../../tests/test_3789_out')
+    pages_per_split = 5
+
+    # Clean pdf
+    doc = pymupdf.open(file_path)
+    tmp = io.BytesIO()
+    tmp.write(doc.write(garbage=4, deflate=True))
+
+    source_doc = pymupdf.Document('pdf', tmp.getvalue())
+    tmp.close()
+
+    # Calculate the number of pages per split file and the number of split files
+    page_range = pages_per_split - 1
+    split_range = range(0, source_doc.page_count, pages_per_split)
+    num_splits = len(split_range)
+
+    # Loop through each split range and create a new PDF file
+    for i, start in enumerate(split_range):
+        output_doc = pymupdf.open()
+
+        # Determine the ending page for this split file
+        to_page = start + page_range if i < num_splits - 1 else -1
+        output_doc.insert_pdf(source_doc, from_page=start, to_page=to_page)
+
+        # Save the output document to a file and add the path to the list of split files
+        path = f'{result_path}_{i}.pdf'
+        output_doc.save(path, garbage=2)
+        print(f'Have saved to {path=}.')
+
+        # If this is the last split file, exit the loop
+        if to_page == -1:
+            break
+
+
+def test_widget_insert():
+    """Confirm copy of form fields / widgets."""
+    tar = pymupdf.open(os.path.join(resources, "merge-form1.pdf"))
+    pc0 = tar.page_count  # for later assertion
+    src = pymupdf.open(os.path.join(resources, "interfield-calculation.pdf"))
+    pc1 = src.page_count  # for later assertion
+
+    tarpdf = pymupdf._as_pdf_document(tar)
+    tar_field_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(tarpdf), "Root/AcroForm/Fields")
+    )
+    tar_co_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(tarpdf), "Root/AcroForm/CO")
+    )
+    srcpdf = pymupdf._as_pdf_document(src)
+    src_field_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(srcpdf), "Root/AcroForm/Fields")
+    )
+    src_co_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(srcpdf), "Root/AcroForm/CO")
+    )
+
+    tar.insert_pdf(src)
+    new_field_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(tarpdf), "Root/AcroForm/Fields")
+    )
+    new_co_count = mupdf.pdf_array_len(
+        mupdf.pdf_dict_getp(mupdf.pdf_trailer(tarpdf), "Root/AcroForm/CO")
+    )
+    assert tar.page_count == pc0 + pc1
+    assert new_field_count == tar_field_count + src_field_count
+    assert new_co_count == tar_co_count + src_co_count
+
+
+def names_and_kids(doc):
+    """Return a list of dictionaries with keys "name" and "kids".
+
+    "name" is the name of a root field in "Root/AcroForm/Fields", and
+    "kids" is the count of its immediate children.
+    """
+    rc = []
+    pdf = pymupdf._as_pdf_document(doc)
+    fields = mupdf.pdf_dict_getl(
+        mupdf.pdf_trailer(pdf),
+        pymupdf.PDF_NAME("Root"),
+        pymupdf.PDF_NAME("AcroForm"),
+        pymupdf.PDF_NAME("Fields"),
+    )
+    if not fields.pdf_is_array():
+        return rc
+    root_count = fields.pdf_array_len()
+    if not root_count:
+        return rc
+    for i in range(root_count):
+        field = fields.pdf_array_get(i)
+        kids = field.pdf_dict_get(pymupdf.PDF_NAME("Kids"))
+        kid_count = kids.pdf_array_len()
+        T = field.pdf_dict_get_text_string(pymupdf.PDF_NAME("T"))
+        field_dict = {"name": T, "kids": kid_count}
+        rc.append(field_dict)
+    return rc
+
+
+def test_merge_checks1():
+    """Merge Form PDFs making any duplicate names unique."""
+    merge_file1 = os.path.join(resources, "merge-form1.pdf")
+    merge_file2 = os.path.join(resources, "merge-form2.pdf")
+    tar = pymupdf.open(merge_file1)
+    rc0 = names_and_kids(tar)
+    src = pymupdf.open(merge_file2)
+    rc1 = names_and_kids(src)
+    tar.insert_pdf(src, join_duplicates=False)
+    rc2 = names_and_kids(tar)
+    assert len(rc2) == len(rc0) + len(rc1)
+
+
+def test_merge_checks2():
+    # Join / merge Form PDFs joining any duplicate names in the src PDF.
+    merge_file1 = os.path.join(resources, "merge-form1.pdf")
+    merge_file2 = os.path.join(resources, "merge-form2.pdf")
+    tar = pymupdf.open(merge_file1)
+    rc0 = names_and_kids(tar)  # list of root names and kid counts
+    names0 = [itm["name"] for itm in rc0]  # root names in target
+    kids0 = sum([itm["kids"] for itm in rc0])  # number of kids in target
+
+    src = pymupdf.open(merge_file2)
+    rc1 = names_and_kids(src)  # list of root namesand kids in source PDF
+    dup_count = 0  # counts duplicate names in source PDF
+    dup_kids = 0  # counts the expected kids after merge
+
+    for itm in rc1:  # walk root fields of source pdf
+        if itm["name"] not in names0:  # not a duplicate name
+            continue
+        # if target field has kids, add their count, else add 1
+        dup_kids0 = sum([i["kids"] for i in rc0 if i["name"] == itm["name"]])
+        dup_kids += dup_kids0 if dup_kids0 else 1
+        # if source field has kids add their count, else add 1
+        dup_kids += itm["kids"] if itm["kids"] else 1
+
+    names1 = [itm["name"] for itm in rc1]  # names in source
+
+    tar.insert_pdf(src, join_duplicates=True)  # join merging any duplicate names
+
+    rc2 = names_and_kids(tar)  # get names and kid counts in resulting PDF
+    names2 = [itm["name"] for itm in rc2]  # resulting names in target
+    kids2 = sum([itm["kids"] for itm in rc2])  # total resulting kid count
+
+    assert len(set(names0 + names1)) == len(names2)
+    assert kids2 == dup_kids
+
+
+test_4412_path = os.path.normpath(f'{__file__}/../../tests/resources/test_4412.pdf')
+
+def test_4412():
+    # This tests whether a page from a PDF containing widgets found in the wild
+    # can be inserted into a new document with default options (widget=True)
+    # and widget=False.
+    print()
+    for widget in True, False:
+        print(f'{widget=}', flush=1)
+        with pymupdf.open(test_4412_path) as doc, pymupdf.open() as new_doc:
+            buf = io.BytesIO()
+            new_doc.insert_pdf(doc, from_page=1, to_page=1)
+            new_doc.save(buf)
+            assert len(new_doc)==1
+
+
+def test_4571():
+    path = os.path.normpath(f'{__file__}/../../tests/resources/test_4571.pdf')
+    path_out = os.path.normpath(f'{__file__}/../../tests/resources/test_4571_out.pdf')
+    with pymupdf.open() as newdocument:
+        with pymupdf.open(path) as document:
+            newdocument.insert_pdf(document)
+        newdocument.save(path_out, garbage=4, clean=False)
+        print(f'Have saved to: {path_out=}')
+    with open(path_out, 'rb') as f:
+        content = f.read()
+    if pymupdf.mupdf_version_tuple >= (1, 26, 6):
+        # Correct.
+        assert b'<</Type/Pages/Count 6/Kids[4 0 R 6 0 R 12 0 R 13 0 R 14 0 R 15 0 R]>>' in content
+    else:
+        # Incorrect.
+        assert b'<</Type/Pages/Count 6/Kids[4 0 R 6 0 R 12 0 R 4 0 R 6 0 R 12 0 R]>>' in content
+