from io import StringIO

from pptx import Presentation
from unstructured.cleaners.core import clean
from unstructured.partition.pptx import partition_pptx

from ordered_multimap import OrderedMultiIndexMapWeakRef

def ppt_chunk(file_like, from_page=0, to_page=100000):
    weakDict = OrderedMultiIndexMapWeakRef()

    ppt = Presentation(file_like)
    total_pages = len(ppt.slides)

    ppt_parser = RAGFlowPptParser()
    metadata_main_title = ''
    
    for pn, slide in enumerate(ppt.slides):
        if pn < from_page:
            continue
        if pn >= to_page:
            break

        try:
            _ = slide.shapes[0]
        except IndexError:
            continue

        text_shapes = [shape for shape in slide.shapes if shape.has_text_frame]

        if len(text_shapes) == 1:
            metadata_main_title = text_shapes[0].text_frame.text
            continue
            
        slide_text = ppt_parser(slide)
        print(f'SLIDE TEXT: {slide_text}')
        weakDict.insert(slide_text, metadata_main_title)
    return weakDict

    raise NotImplementedError(
        "file type not supported yet(pptx)")