derek-jones/pdf-2-csv

Extracts points in graphs to csv

pdf-2-csv's Issues

Identify pdf operations responsible for drawing a region on a page

Converting a graph from pdf to csv first requires that all the associated pdf operations be identified.

When viewing a pdf, using Mozilla's PDF.js, the user highlights a region on the page. All pdf operations associated with this region need to be identified and extracted from the pdf.

Identify the start/end of the x and y axis of graph

A graph will contain at least one x and y axis (assume one of each to start).

The most likely start/end of the x and y axis needs to be identified.

Identify the two longest lines at right-angles to each other, is one possible heuristic.

pdf operations use page relative coordinates. Knowing the page relative coordinates allows the page relative coordinates of graph points to be mapped to graph relative coordinates.

The numeric values of the x/y axis tick marks need to be identified

Extracting x/y axis relative coordinates of points in a graph requires knowing the start/end values and scale used for the x and y axis.

The start/end values and whether a linear or log scale is used, needs to be identified for each of the x and y axis.

Method used to denote a single point in the graph

The method used by particular a pdf generation tool to denote a single point (e.g., circle or cross) needs to be identified. Tools often embed an identifying string in the pdf they generate.

Some tools create their own 'point' character by, for instance, drawing vertical/horizontal lines. Things get complicated when the tool generates pdf operations that draws all the horizontal lines first, followed by all the vertical lines (or vice versa).

The following draws a circle on the page:
/F1 1 Tf 1 Tr 6.21 0 0 6.21 135.35 423.79 Tm (l) Tj 0 Tr