new api method for data image search

As mentioned, after doing some experiments for the KDnuggets article, I bundled some of the existing API methods into a new one, which will extract from a page figures that have x/y scaling information. The JSON output is well suited to elasticsearch or your favorite flavor of NoSQL eg., an extract of the response:

[{“input”:”tmp/quant-ph0002044-9.png”, “qc_output”:”/tmp/quant-ph0002044-9_qc_out_0.png”, “qc_input”:”/tmp/quant-ph0002044-9_qc_in_0.png”, “xlabel”:”BOB’S ERROR RATE”, “ylabel”:”EVE’S ERROR RATE”, “xmax”:”0.200000″, “xmin”:”0.000000″, “ymax”:”0.000000″, “ymin”:”0.400000″, “data”:[ {“x”:0.009246,”y”:0.159044}, {“x”:0.059028,”y”:0.355084}, {“x”:0.152082,”y”:0.327358}, {“x”:0.155371,”y”:0.126410}, {“x”:0.011028,”y”:-0.000748}, {“x”:0.174976,”y”:0.099922}

The API response also contains the input/output bounding box images used for the inversion of the data. With this meta data, a user can search for data types (based on x/y label fields), ranges and values, and receive a model of the data used to produce the plot.