[pdf] Switch PDF page rendering to an iterator format

[frgfm]

As suggested by @mara004 in #1000, the PDF rendering is using a list comprehension which holds page processing while the rendering isn't complete:

doctr/doctr/io/pdf.py

Line 45 in 78c9105

return [np.asarray(img) for img in pdf.render_topil(scale=scale, **kwargs)]

This could be fixed by modifying it to yield the page and limit RAM usage:

with pdfium.PdfDocument(file, password=password) as pdf:
    for img in pdf.render_topil(scale=scale, **kwargs): yield np.asarray(img)

[felixdittrich92]

Waiting until #1032 is available

[felixdittrich92]

Outdated by #1240