This repo contains some Python scripts that I have been using to reverse-engineer figures (mainly from papers, or datasheets) to extract quantitatively their data. So far they only work for a selected number of figure types: line plots, scatter, bargraphs and heatmaps. Fortunately, these cover the vast majority of figures found around. The data is exported in a .csv file.
1 Their usage is fairly simple. From the console execute any file as python export_scatter.py -i input_filepath -o output_filepath
. And perhaps some of the optional parameters:
`-sz VALUE` to resize the image previews. Value can be any decimal number. Default = 1
`-th VALUE` to change the color similarity threshold that recognizes poits/lines. Value can be any number in [0,1]. default = 0.95
`-p VALUE` to specify the output precision in numbers of decimals. Value can be any integer >1. Default=2
2 The image will pop up and you need to specify the axis limits by click-and-dragging the bounding box. Press ENTER to continue.
3 Follow the questions asked in the terminal.
4 Enjoy your results in the csv file!