Output of drawings

[meghanaviyyapu]

I have extracted drawings using page.get_drawings(). Can you let me know how to extract the values that are present in the content stream without any transformation?

For example in the content stream the rect values are 100 700 100 100 whereas the rect value obtained using get_drawings had values
100 42 200 142.

Can you let me know how to obtain the same rect values present in the content stream?

[JorjMcKie]

In (Py-) MuPDF, coordinates count from top to bottom: top-left has coordinates (0, 0).
In PDF, page bottom-left has (0, 0).
So you need the ptm = page.transformation_matrix to translate between these coordinate systems. Given a PDF point (x, y), then Point(x, y) * ptm gives the point in MuPDF coordinates. The other way round, a point in MuPDF coordinates (as contained in .get_drawings()) has ccordinates Point(x, y) * ~ptm in PDF.

[JorjMcKie]

But keep in mind, that values seen in the content stream may be subject to multiple matrix transformations ("cm") coded before, like this

1 2 3 4 5 6 cm
...
6 5 4 3 2 1 cm
...
100 700 100 100 re
...

So MuPDF sequentially walks through all those cm instructions, multiplies these matrices one by one and finally with the ptm and then the rectangle coordinates to give the MuPDF values.

[meghanaviyyapu]

Is there difference in drawings creation when page.getcdrawings() is used instead of page.get_drawings()?

[JorjMcKie]

Except for a great performance difference, the C-version only delivers tuples, no fitz.Point, fitz.Rect, fitz.Quad.

[meghanaviyyapu]

Can you let me know if the blue drawing behind some text in the PDF can be extracted using PyMuPDF?
sample_pdf.pdf

[JorjMcKie]

These are vector graphics, so you can extract them using page.get_drawings().
However, the fill color is shaded: this is not currently supported.

[meghanaviyyapu]

Thanks. Is there any way to get the shaded part using PyMuPDF?

[JorjMcKie]

Thanks. Is there any way to get the shaded part using PyMuPDF?

Sorry, not currently.