Abstract: A computer implemented method configured to analyze contents of a page of a vector graphic format file includes dividing text content on the page into text blocks, pre-classifying each text block to be one of a raw body text block and a non-body text block, processing the raw body text blocks to form a plurality of body text blocks and combining the body text blocks into a text flow including the plurality of body text blocks. The method further includes defining a semantical role of each of the non-body text blocks, and combining the non-body text blocks among the body text blocks of the text flow in a geometrical order. Result data is provided, formatted as any one of a data structure and a data stream, the result data including text content of the page as geometrically arranged non-body text blocks and body text blocks.