mirror of
https://github.com/Unstructured-IO/unstructured.git
synced 2025-11-04 12:03:15 +00:00
e73cf09977
* page breaks for pptx * added page breaks for image/pdf * tests for images with page breaks * page breaks for html documents * linting, linting, linting * changelog and bump version * update docs * fix typo * refactor reusable code to common.py * add type back in
147 lines
4.9 KiB
Python
147 lines
4.9 KiB
Python
import os
|
|
import pathlib
|
|
import pytest
|
|
|
|
import pptx
|
|
|
|
from unstructured.partition.pptx import partition_pptx
|
|
from unstructured.documents.elements import ListItem, NarrativeText, PageBreak, Text, Title
|
|
|
|
DIRECTORY = pathlib.Path(__file__).parent.resolve()
|
|
EXAMPLE_DOCS_DIRECTORY = os.path.join(DIRECTORY, "..", "..", "example-docs")
|
|
|
|
EXPECTED_PPTX_OUTPUT = [
|
|
Title(text="Adding a Bullet Slide"),
|
|
ListItem(text="Find the bullet slide layout"),
|
|
ListItem(text="Use _TextFrame.text for first bullet"),
|
|
ListItem(text="Use _TextFrame.add_paragraph() for subsequent bullets"),
|
|
NarrativeText(text="Here is a lot of text!"),
|
|
NarrativeText(text="Here is some text in a text box!"),
|
|
]
|
|
|
|
|
|
def test_partition_pptx_from_filename():
|
|
filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "fake-power-point.pptx")
|
|
elements = partition_pptx(filename=filename)
|
|
assert elements == EXPECTED_PPTX_OUTPUT
|
|
|
|
|
|
def test_partition_pptx_from_file():
|
|
filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "fake-power-point.pptx")
|
|
with open(filename, "rb") as f:
|
|
elements = partition_pptx(file=f)
|
|
assert elements == EXPECTED_PPTX_OUTPUT
|
|
|
|
|
|
def test_partition_pptx_raises_with_both_specified():
|
|
filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "fake-power-point.pptx")
|
|
with open(filename, "rb") as f:
|
|
with pytest.raises(ValueError):
|
|
partition_pptx(filename=filename, file=f)
|
|
|
|
|
|
def test_partition_pptx_raises_with_neither():
|
|
with pytest.raises(ValueError):
|
|
partition_pptx()
|
|
|
|
|
|
def test_partition_pptx_adds_page_breaks(tmpdir):
|
|
filename = os.path.join(tmpdir, "test-page-breaks.pptx")
|
|
|
|
presentation = pptx.Presentation()
|
|
blank_slide_layout = presentation.slide_layouts[6]
|
|
|
|
slide = presentation.slides.add_slide(blank_slide_layout)
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is the first slide."
|
|
|
|
slide = presentation.slides.add_slide(blank_slide_layout)
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is the second slide."
|
|
|
|
presentation.save(filename)
|
|
|
|
elements = partition_pptx(filename=filename)
|
|
|
|
assert elements == [
|
|
NarrativeText(text="This is the first slide."),
|
|
PageBreak(),
|
|
NarrativeText(text="This is the second slide."),
|
|
]
|
|
|
|
|
|
def test_partition_pptx_page_breaks_toggle_off(tmpdir):
|
|
filename = os.path.join(tmpdir, "test-page-breaks.pptx")
|
|
|
|
presentation = pptx.Presentation()
|
|
blank_slide_layout = presentation.slide_layouts[6]
|
|
|
|
slide = presentation.slides.add_slide(blank_slide_layout)
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is the first slide."
|
|
|
|
slide = presentation.slides.add_slide(blank_slide_layout)
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is the second slide."
|
|
|
|
presentation.save(filename)
|
|
|
|
elements = partition_pptx(filename=filename, include_page_breaks=False)
|
|
|
|
assert elements == [
|
|
NarrativeText(text="This is the first slide."),
|
|
NarrativeText(text="This is the second slide."),
|
|
]
|
|
|
|
|
|
def test_partition_pptx_orders_elements(tmpdir):
|
|
filename = os.path.join(tmpdir, "test-ordering.pptx")
|
|
|
|
presentation = pptx.Presentation()
|
|
blank_slide_layout = presentation.slide_layouts[6]
|
|
slide = presentation.slides.add_slide(blank_slide_layout)
|
|
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is lower and should come second"
|
|
|
|
left = top = width = height = pptx.util.Inches(1)
|
|
left = top = pptx.util.Inches(-10)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is off the page and shouldn't appear"
|
|
|
|
left = top = width = height = pptx.util.Inches(2)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = ""
|
|
|
|
left = top = width = height = pptx.util.Inches(1)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "This is higher and should come first"
|
|
|
|
top = width = height = pptx.util.Inches(1)
|
|
left = pptx.util.Inches(0.5)
|
|
txBox = slide.shapes.add_textbox(left, top, width, height)
|
|
tf = txBox.text_frame
|
|
tf.text = "-------------TOP-------------"
|
|
|
|
presentation.save(filename)
|
|
|
|
elements = partition_pptx(filename=filename)
|
|
assert elements == [
|
|
Text("-------------TOP-------------"),
|
|
NarrativeText("This is higher and should come first"),
|
|
NarrativeText("This is lower and should come second"),
|
|
]
|