zhafen / cc

Natural language processing of scientific abstracts to extract quantitative trends.
1 stars 0 forks source link

Possible shape / index error in topography_metrics #13

Open nathimel opened 1 year ago

nathimel commented 1 year ago

In section 'Identify Publications to Analyze', the first cell gives

converged_kernel_size = data_k['converged_kernel_size'][:,-pm['convergence_degree']] (2, 31667, 10)

but should be (31667, 10)

We have the following print(data_k['converged_kernel_size'][0,:,-pm['convergence_degree']].shape) (31667,)

This may be a result of some error in the following steps that produce data consumed by topography_metrics:

populate_topography background_distribution (did not run Astrophysics Per Year, but doubt that's relevant) search_convergence (immediately prior to topography_metrics, for the relevant region)

zhafen commented 1 year ago

Thanks for opening this!