jupyanno
jupyanno copied to clipboard
chestrays
Make data labeling easy with Jupyter notebooks and Google Sheets!
Task file should contain - viewer specifications - backend for storage (CSV, gsheets, pastebin) - question type, meta-data to collect
Following on from #39 and connected to #34. We should setup tooling so that users can `pip install jupyanno` to use it locally to annotate their own datasets. Might also...
Is there a way to run it locally? If there is how, instructions would be nice.
 Right now the task is too biased and it would be better if sklearn.dummy got a worse score
We should consider splitting the "library" part of this repo from the "deployment" part. Will make reusing this for other annotation tasks easier.
it would make loading images / data much easier to replace open with smart_open (https://github.com/RaRe-Technologies/smart_open)
blank-out image after a certain amount of time (add countdown timer)
crashes when x or y has None/nulls in it ``` from matplotlib.patches import Rectangle for col_name in 'xy': for i in range(2): zooms_df['{}_{}'.format(col_name, i)] = zooms_df[col_name].map( lambda x: x[i]) zooms_df['d_{}'.format(col_name,...