Understand processing times

[dcsw2]

Goal: document processing times for different settings in the pipeline (esp. DeezyMatch vs. perfectmatch)

Use sample set of articles from The Sun for this test.

TASKS

• [x] confirm different settings to test (from 4 main versions) @fedenanni
• [x] process sample @npedrazzini @fedenanni
• [ ] report results to group here ca. Oct 12

Structure of "report":

• end to end time
• for a smaller sample, timings for each step

[dcsw2]

Kaspar's random sample from The Sun: randsample0002194.csv

[fedenanni]

The four versions of the pipelines are here as jupyter notebooks: https://github.com/Living-with-machines/toponym-resolution/tree/dev/experiments

[kmcdono2]

Now waiting for updates to including candidates

[kmcdono2]

@npedrazzini has done a nice analysis of processing times across the different T-Res methods: https://docs.google.com/spreadsheets/d/1ymjGPubsjq93VmCakBDxYOQW_TCXXxECG1OLG5n1YV8/edit#gid=175352557