Text Positions

[meghanaviyyapu]

I have extracted rect values of text positions from a PDF. I want to insert text based on the positions extracted. Can you let me know if text matrix(Tm) values of a PDF can be obtained based on rect values? Is there any difference between rect values of text and text matrix values?

[JorjMcKie]

The text matrix Tm in PDF /Contents source is not directly available. Its use is also not intended in PyMuPDF - although you could of course read the /Contents as a stream and dig your way through to all information you want.

The bbox accompanying extracted text always already contains any computations required to deliver the correct value.

So you can use those bboxes (plus the span["origin"] - which is even more important) for inserting the span text.

[JorjMcKie]

In this case simple set the required text matrix to the identity matrix (1,0,0,1,0,0).
But confirm that that other package uses the same geometry as MuPDF: top-left = (0,0).
Otherwise use the matrix ~page.transformation_matrix to recover PDF geometry: bottom-left = (0,0).

[meghanaviyyapu]

My objective is to insert text in a new PDF based on the text extracted from the input PDF. To insert text at the same position as that of input PDF, can you let me know the values that will be required and if extracted values can be used directly without any transformation?

[meghanaviyyapu]

In this case simple set the required text matrix to the identity matrix (1,0,0,1,0,0). But confirm that that other package uses the same geometry as MuPDF: top-left = (0,0). Otherwise use the matrix ~page.transformation_matrix to recover PDF geometry: bottom-left = (0,0).

Can you let me know what (x,y) values I should provide while inserting text to insert at the exact position of the input PDF?

[JorjMcKie]

This is span["origin"]. As I said: potentially to be multiplied with the transformation matrix inverse, depending on the geometry priciple of that other package.

[meghanaviyyapu]

Ok, Thanks for clarifying.