import re
import traceback
from dataclasses import dataclass
from typing import Callable
from pptx import Presentation as PPTXPre
from pptx.enum.shapes import MSO_SHAPE_TYPE
from pptx.oxml import parse_xml
from pptx.shapes.autoshape import Shape as PPTXAutoShape
from pptx.shapes.base import BaseShape
from pptx.shapes.connector import Connector as PPTXConnector
from pptx.shapes.group import GroupShape as PPTXGroupShape
from pptx.shapes.picture import Picture as PPTXPicture
from pptx.shapes.placeholder import PlaceholderPicture, SlidePlaceholder, TablePlaceholder
from pptx.shapes.freeform import FreeformBuilder
from pptx.slide import Slide as PPTXSlide
from pptx.text.text import _Paragraph, _Run
from rich import print
from utils.src.utils import (
IMAGE_EXTENSIONS,
Config,
apply_fill,
dict_to_object,
extract_fill,
get_font_pptcstyle,
get_font_style,
merge_dict,
object_to_dict,
parse_groupshape,
pexists,
pjoin,
runs_merge,
wmf_to_images,
)
INDENT = "\t"
# textframe: shape bounds font
# paragraph: space, alignment, level, font bullet
# run: font, hyperlink, text
@dataclass
class StyleArg:
"""
A class to represent style arguments for HTML conversion.
"""
paragraph_id: bool = True
element_id: bool = True
font_style: bool = True
area: bool = False
size: bool = False
geometry: bool = False
show_image: bool = True
@dataclass
class Closure:
closure: Callable
paragraph_id: int = -1
def apply(self, shape: BaseShape):
"""
Apply the closure to a shape.
Args:
shape (BaseShape): The shape to apply the closure to.
"""
self.closure(shape)
def __gt__(self, other):
if self.paragraph_id != other.paragraph_id:
return self.paragraph_id > other.paragraph_id
class Paragraph:
def __init__(self, paragraph: _Paragraph, idx: int):
run = runs_merge(paragraph)
self.idx = idx
self.real_idx = idx
self.bullet = paragraph.bullet
if run is None:
self.idx = -1
return
self.font = merge_dict(
object_to_dict(paragraph.font), [object_to_dict(run.font)]
)
self.text = re.sub(r"(_x000B_|\\x0b)", " ", paragraph.text)
def to_html(self, style_args: StyleArg):
if self.idx == -1:
raise ValueError(f"paragraph {self.idx} is not valid")
tag = "li" if self.bullet else "p"
id_str = f" id='{self.idx}'" if style_args.paragraph_id else ""
font_style = get_font_style(self.font)
style_str = (
f" style='{font_style}'" if style_args.font_style and font_style else ""
)
if self.bullet:
style_str += f" bullet-type='{self.bullet}'"
return f"<{tag}{id_str}{style_str}>{self.text}{tag}>"
def __repr__(self):
return f"Paragraph-{self.idx}: {self.text}"
class TextFrame:
def __init__(self, shape: BaseShape, level: int):
if not shape.has_text_frame:
self.is_textframe = False
return
self.paragraphs = [
Paragraph(paragraph, idx)
for idx, paragraph in enumerate(shape.text_frame.paragraphs)
]
para_offset = 0
for para in self.paragraphs:
if para.idx == -1:
para_offset += 1
else:
para.idx = para.idx - para_offset
if len(self.paragraphs) == 0:
self.is_textframe = False
return
self.level = level
self.text = shape.text
self.is_textframe = True
self.font = merge_dict(
object_to_dict(shape.text_frame.font),
[para.font for para in self.paragraphs if para.idx != -1],
)
def to_html(self, style_args: StyleArg):
"""
Convert the text frame to HTML.
Args:
style_args (StyleArg): The style arguments for HTML conversion.
Returns:
str: The HTML representation of the text frame.
"""
if not self.is_textframe:
return ""
repr_list = [
para.to_html(style_args) for para in self.paragraphs if para.idx != -1
]
return "\n".join([INDENT * self.level + repr for repr in repr_list])
def __repr__(self):
if not self.is_textframe:
return "TextFrame: null"
return f"TextFrame: {self.paragraphs}"
def __len__(self):
if not self.is_textframe:
return 0
return len(self.text)
def to_pptc(self, father_idx: int) -> str:
"""
Convert the text frame to PPTC format.
Args:
father_idx (int): The index of the parent shape.
Returns:
str: The PPTC representation of the text frame.
"""
if not self.is_textframe:
return ""
s = f"[Text id={father_idx}]"
for para in self.paragraphs:
if para.idx == -1:
continue
s += f"\n"
s += f"[Paragraph id={para.idx}]"
s += get_font_pptcstyle(para.font) + f"\n"
s += para.text + "\n"
return s
class ShapeElement:
def __init__(
self,
slide_idx: int,
shape_idx: int,
style: dict,
data: dict,
text_frame: TextFrame,
slide_area: float,
level: int,
):
self.slide_idx = slide_idx
self.shape_idx = shape_idx
self.style = style
self.data = data
self.text_frame = text_frame
self._closure_keys = ["clone", "replace", "delete", "style"]
self._closures: dict[str, list[Closure]] = {
key: [] for key in self._closure_keys
}
self.slide_area = slide_area
self.level = level
@classmethod
def from_shape(
cls,
slide_idx: int,
shape_idx: int,
shape: BaseShape,
config: Config,
slide_area: float,
level: int = 0,
):
"""
Create a ShapeElement from a BaseShape.
Args:
slide_idx (int): The index of the slide.
shape_idx (int): The index of the shape.
shape (BaseShape): The shape object.
config (Config): The configuration object.
slide_area (float): The area of the slide.
level (int): The indentation level.
Returns:
ShapeElement: The created ShapeElement.
"""
if shape_idx > 100 and isinstance(shape, PPTXGroupShape):
raise ValueError(f"nested group shapes are not allowed")
line = None
if "line" in dir(shape) and shape.line._ln is not None:
line = {
"fill": extract_fill(shape.line),
"width": shape.line.width,
"dash_style": shape.line.dash_style,
}
fill = extract_fill(shape)
style = {
"shape_bounds": {
"width": shape.width,
"height": shape.height,
"left": shape.left,
"top": shape.top,
},
"shape_type": str(shape.shape_type).split("(")[0].lower(),
"rotation": shape.rotation,
"fill": fill,
"line": line,
}
text_frame = TextFrame(shape, level + 1)
obj = SHAPECAST.get(shape.shape_type, UnsupportedShape).from_shape(
slide_idx,
shape_idx,
shape,
style,
text_frame,
config,
slide_area,
level,
)
obj.xml = shape._element.xml
# ? for debug, mask to enable pickling
# obj.shape = shape
return obj
def build(self, slide: PPTXSlide):
"""
Build the shape element in a slide.
Args:
slide (PPTXSlide): The slide to build the shape in.
Returns:
The built shape.
"""
return slide.shapes._shape_factory(
slide.shapes._spTree.insert_element_before(parse_xml(self.xml), "p:extLst")
)
def __repr__(self) -> str:
return f"{self.__class__.__name__}: shape {self.shape_idx} of slide {self.slide_idx}"
def to_html(self, style_args: StyleArg) -> str:
"""
Convert the shape element to HTML.
Args:
style_args (StyleArg): The style arguments for HTML conversion.
Returns:
str: The HTML representation of the shape element.
"""
return ""
@property
def closures(self):
"""
Get the closures associated with the shape element.
Returns:
list: A list of closures.
"""
closures = []
closures.extend(sorted(self._closures["clone"]))
closures.extend(self._closures["replace"] + self._closures["style"])
closures.extend(sorted(self._closures["delete"], reverse=True))
return closures
@property
def indent(self):
return "\t" * self.level
@property
def left(self):
return self.style["shape_bounds"]["left"].pt
@left.setter
def left(self, value):
self.style["shape_bounds"]["left"] = value
@property
def top(self):
return self.style["shape_bounds"]["top"].pt
@top.setter
def top(self, value):
self.style["shape_bounds"]["top"] = value
@property
def width(self):
return self.style["shape_bounds"]["width"].pt
@width.setter
def width(self, value):
self.style["shape_bounds"]["width"] = value
@property
def height(self):
return self.style["shape_bounds"]["height"].pt
@height.setter
def height(self, value):
self.style["shape_bounds"]["height"] = value
@property
def area(self):
"""
Get the area of the shape element.
Returns:
float: The area in square points.
"""
return self.width * self.height
@property
def pptc_text_info(self):
"""
Get the PPTC text information of the shape element.
Returns:
str: The PPTC text information.
"""
if isinstance(self, Picture):
return self.caption
return self.text_frame.to_pptc(self.shape_idx)
@property
def pptc_space_info(self):
"""
Get the PPTC space information of the shape element.
Returns:
str: The PPTC space information.
"""
return f"Visual Positions: left={self.left}pt, top={self.top}pt\n"
@property
def pptc_size_info(self):
"""
Get the PPTC size information of the shape element.
Returns:
str: The PPTC size information.
"""
return f"Size: height={self.height}pt, width={self.width}pt\n"
@property
def pptc_description(self):
"""
Get the PPTC description of the shape element.
Returns:
str: The PPTC description.
"""
return f"[{self.__class__.__name__} id={self.shape_idx}]\n"
def to_pptc(self):
"""
Convert the shape element to PPTC format.
Returns:
str: The PPTC representation of the shape element.
"""
s = ""
s += self.pptc_description
s += self.pptc_size_info
s += self.pptc_space_info
s += self.pptc_text_info
return s
def get_inline_style(self, style_args: StyleArg):
"""
Get the inline style for the shape element.
Args:
style_args (StyleArg): The style arguments for HTML conversion.
Returns:
str: The inline style string.
"""
id_str = f" id='{self.shape_idx}'" if style_args.element_id else ""
styles = []
if style_args.area:
styles.append(f"data-relative-area={self.area*100/self.slide_area:.2f}%;")
if style_args.size:
styles.append(f"width: {self.width}pt; height: {self.height}pt;")
if style_args.geometry:
styles.append(f"left: {self.left}pt; top: {self.top}pt;")
if style_args.font_style and self.text_frame.is_textframe:
font_style = get_font_style(self.text_frame.font)
if font_style:
styles.append(font_style)
if len(styles) != 0:
return id_str + " style='" + " ".join(styles) + "'"
return id_str
class UnsupportedShape(ShapeElement):
@classmethod
def from_shape(
cls,
slide_idx: int,
shape_idx: int,
shape: BaseShape,
*args,
**kwargs,
):
raise ValueError(f"unsupported shape {shape.shape_type}")
class TextBox(ShapeElement):
@classmethod
def from_shape(
cls,
slide_idx: int,
shape_idx: int,
shape: TextFrame,
style: dict,
text_frame: TextFrame,
config: Config,
slide_area: float,
level: int,
):
return cls(slide_idx, shape_idx, style, None, text_frame, slide_area, level)
def to_html(self, style_args: StyleArg) -> str:
return (
f"{self.indent}
\n"
)
class Picture(ShapeElement):
@classmethod
def from_shape(
cls,
slide_idx: int,
shape_idx: int,
shape: PPTXPicture,
style: dict,
text_frame: TextFrame,
config: Config,
slide_area: float,
level: int,
):
img_path = pjoin(
config.IMAGE_DIR,
f"{shape.image.sha1}.{shape.image.ext}",
)
if shape.image.ext == "wmf":
img_path = img_path.replace(".wmf", ".jpg")
if not pexists(img_path):
wmf_to_images(shape.image.blob, img_path)
elif shape.image.ext not in IMAGE_EXTENSIONS:
raise ValueError(f"unsupported image type {shape.image.ext}")
if not pexists(img_path):
with open(img_path, "wb") as f:
f.write(shape.image.blob)
style["img_style"] = {
"crop_bottom": shape.crop_bottom,
"crop_top": shape.crop_top,
"crop_left": shape.crop_left,
"crop_right": shape.crop_right,
}
picture = cls(
slide_idx,
shape_idx,
style,
[img_path, shape.name, ""],
text_frame,
slide_area,
level=level,
)
return picture
def build(self, slide: PPTXSlide):
shape = slide.shapes.add_picture(
self.img_path,
**self.style["shape_bounds"],
)
shape.name = self.data[1]
dict_to_object(self.style["img_style"], shape.image)
apply_fill(shape, self.style["fill"])
if self.style["line"] is not None:
apply_fill(shape.line, self.style["line"]["fill"])
dict_to_object(self.style["line"], shape.line, exclude=["fill"])
dict_to_object(self.style["shape_bounds"], shape)
if "rotation" in dir(shape):
shape.rotation = self.style["rotation"]
return shape
@property
def img_path(self):
return self.data[0]
@img_path.setter
def img_path(self, img_path: str):
self.data[0] = img_path
@property
def caption(self):
return self.data[2]
@caption.setter
def caption(self, caption: str):
self.data[2] = caption
def to_html(self, style_args: StyleArg) -> str:
if not style_args.show_image:
return ""
if not self.caption:
raise ValueError(
f"caption not found for picture {self.shape_idx} of slide {self.slide_idx}"
)
return (
self.indent
+ f""
)
class FreeformShape(ShapeElement):
"""
Represents a FREEFORM shape. Internally, a freeform is a shape with custom
geometry. This class demonstrates how you might capture that geometry and
rebuild it.
"""
@classmethod
def from_shape(
cls,
slide_idx: int,
shape_idx: int,
shape, # This will be the pptx.shapes.autoshape.Shape whose type is FREEFORM
style: dict,
text_frame: TextFrame,
config: Config,
slide_area: float,
level: int,
):
"""
Captures relevant geometry information (e.g., point sets) from the existing
freeform shape. In python-pptx, retrieving detailed vertex information from
the shape can be tricky, so you might rely on custom extension attributes, or
simply store bounding-box and minimal details. The 'data' dict is for your use.
For demonstration, we assume we have some way to extract or store
basic vertex info from `shape`. Not all details of a FREEFORM
can easily be retrieved from python-pptx, so you may have to
maintain that data separately as you create them.
"""
# Hypothetical way to store some geometry or bounding box:
data = {
"name": shape.name,
"width": shape.width,
"height": shape.height,
"left": shape.left,
"top": shape.top,
# Possibly store a list of points if you have them:
"points": [], # If you have a way to retrieve them
"closed": True, # Whether to close the path
}
return cls(
slide_idx,
shape_idx,
style,
data,
text_frame,
slide_area,
level=level,
)
def build(self, slide: PPTXSlide):
"""
Recreates the freeform geometry on the slide using a FreeformBuilder.
Note: If you loaded this shape from an existing slide, you may already
have it in place. Typically, build() is used when exporting back to PPTX
from your internal representation. If you don't need to reconstruct (because
the shape already exists), you can simply no-op here. This is an example of
how you'd do it if you needed to fully recreate geometry.
"""
# Retrieve your geometry data
left = self.data["left"]
top = self.data["top"]
width = self.data["width"]
height = self.data["height"]
points = self.data["points"]
closed = self.data["closed"]
# Start the builder at the first point (or top-left, etc.).
# This is completely up to you how you define your local coordinate system.
if points:
first_x, first_y = points[0]
else:
# Default to top-left if no data
first_x, first_y = (0, 0)
builder = slide.shapes.build_freeform(left + first_x, top + first_y)
# Now add line segments for the rest of the points (if any).
# This is an example usage:
if len(points) > 1:
# Skip the first point since we used it in build_freeform(...) above
builder.add_line_segments(points[1:], close=closed)
# Convert to a shape
shape = builder.convert_to_shape()
# Possibly apply local styling, rotation, lines, etc.
shape.name = self.data["name"]
apply_fill(shape, self.style.get("fill"))
if self.style.get("line") is not None:
apply_fill(shape.line, self.style["line"]["fill"])
dict_to_object(self.style["line"], shape.line, exclude=["fill"])
# If you have a bounding box in .style["shape_bounds"], apply it:
if "shape_bounds" in self.style:
dict_to_object(self.style["shape_bounds"], shape)
if "rotation" in self.style:
shape.rotation = self.style["rotation"]
return shape
def to_html(self, style_args: StyleArg) -> str:
"""
Returns HTML representation. This could be empty, or you could provide
something like an