doc-forge-reborn/backend/app/services/document_processor.py

import mammoth
from io import BytesIO
from bs4 import BeautifulSoup
from docx import Document
from docx.shared import Pt, Inches
from docx.enum.text import WD_ALIGN_PARAGRAPH
from app.services.file_storage import get_file_content

ALIGN_MAP = {
    WD_ALIGN_PARAGRAPH.CENTER: "center",
    WD_ALIGN_PARAGRAPH.RIGHT: "right",
    WD_ALIGN_PARAGRAPH.JUSTIFY: "justify",
}


async def docx_to_html(file_content: bytes) -> str:
    result = mammoth.convert_to_html(BytesIO(file_content))
    html = result.value

    try:
        doc = Document(BytesIO(file_content))
        soup = BeautifulSoup(html, "html.parser")
        paragraphs = doc.paragraphs
        html_blocks = soup.find_all(["p", "h1", "h2", "h3", "h4", "h5", "h6", "li"])

        for i, para in enumerate(paragraphs):
            if i >= len(html_blocks):
                break
            if para.alignment and para.alignment in ALIGN_MAP:
                css = ALIGN_MAP[para.alignment]
                existing = html_blocks[i].get("style", "")
                styles = f"text-align:{css}"
                if existing:
                    styles = existing.rstrip(";") + ";" + styles
                html_blocks[i]["style"] = styles

        html = str(soup)
    except Exception:
        pass

    return html


def html_to_docx_bytes(html_content: str) -> bytes:
    from html.parser import HTMLParser

    doc = Document()

    style = doc.styles["Normal"]
    font = style.font
    font.name = "Arial"
    font.size = Pt(11)

    class RichHTMLParser(HTMLParser):
        def __init__(self):
            super().__init__()
            self.paragraphs: list[dict] = []
            self.current = {"runs": [], "align": None}
            self.in_paragraph = False
            self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
            self.tag_stack: list[str] = []
            self.heading_level = 0

        def handle_starttag(self, tag, attrs):
            tag_lower = tag.lower()
            attrs_dict = dict(attrs)
            if tag_lower in ("p", "div", "li"):
                self.in_paragraph = True
                self.current = {"runs": [], "align": None}
                self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
                style = attrs_dict.get("style", "")
                if "text-align:center" in style:
                    self.current["align"] = "center"
                elif "text-align:right" in style:
                    self.current["align"] = "right"
                elif "text-align:justify" in style:
                    self.current["align"] = "justify"
            elif tag_lower in ("h1", "h2", "h3", "h4", "h5", "h6"):
                self.in_paragraph = True
                self.heading_level = int(tag_lower[1])
                self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
            elif tag_lower in ("strong", "b"):
                self.current_run["bold"] = True
            elif tag_lower in ("em", "i"):
                self.current_run["italic"] = True
            elif tag_lower == "u":
                self.current_run["underline"] = True
            elif tag_lower in ("br",):
                if self.in_paragraph:
                    self.current["runs"].append(dict(self.current_run))
                    self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
            self.tag_stack.append(tag_lower)

        def handle_endtag(self, tag):
            tag_lower = tag.lower()
            if tag_lower in ("p", "div", "li", "h1", "h2", "h3", "h4", "h5", "h6"):
                if self.current_run["text"].strip():
                    self.current["runs"].append(dict(self.current_run))
                if self.current["runs"]:
                    p = dict(self.current)
                    p["heading"] = self.heading_level
                    self.paragraphs.append(p)
                self.current = {"runs": []}
                self.current_run = {"text": "", "bold": False, "italic": False, "underline": False}
                self.in_paragraph = False
                self.heading_level = 0
            if self.tag_stack:
                self.tag_stack.pop()

        def handle_data(self, data):
            if self.in_paragraph:
                self.current_run["text"] += data

    parser = RichHTMLParser()
    parser.feed(html_content)

    for para_data in parser.paragraphs:
        heading = para_data.get("heading", 0)
        if heading > 0:
            p = doc.add_heading(level=min(heading, 9))
        else:
            p = doc.add_paragraph()

        align_val = para_data.get("align")
        if align_val == "center":
            p.alignment = WD_ALIGN_PARAGRAPH.CENTER
        elif align_val == "right":
            p.alignment = WD_ALIGN_PARAGRAPH.RIGHT
        elif align_val == "justify":
            p.alignment = WD_ALIGN_PARAGRAPH.JUSTIFY

        for run_data in para_data.get("runs", []):
            run = p.add_run(run_data["text"])
            if run_data.get("bold"):
                run.bold = True
            if run_data.get("italic"):
                run.italic = True
            if run_data.get("underline"):
                run.underline = True

    output = BytesIO()
    doc.save(output)
    return output.getvalue()


def docx_to_pdf_bytes(file_content: bytes) -> bytes:
    doc = Document(BytesIO(file_content))

    from io import BytesIO as Bio
    from reportlab.lib.pagesizes import A4
    from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
    from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer
    from reportlab.lib.enums import TA_LEFT

    buffer = Bio()
    pdf_doc = SimpleDocTemplate(buffer, pagesize=A4)

    styles = getSampleStyleSheet()
    story = []

    for para in doc.paragraphs:
        if para.text.strip():
            p = Paragraph(para.text, styles["Normal"])
            story.append(p)
            story.append(Spacer(1, 6))

    pdf_doc.build(story)
    return buffer.getvalue()