The commands below use Anaconda to manage the project environment.
conda env create -f environment.yml
conda env update -f environment.yml
conda activate droids-env
# in case of error 'Your shell has not been properly configured ...', try 'source ...'
source activate droids-env
conda deactivatejupyter notebook &
# convert to HTML slides or use RISE for a dynamic slideshow
jupyter nbconvert hello.ipynb --to slides --post serveRISE turns a Jupyter notebook into a live reveal.js-based presentation - the code
cells can be executed in presentation mode. You can also edit the code.
To change the presentation theme:
- Go to Edit / Edit Notebook Metadata
- Append
"rise": {"theme": "sky"}to the JSON content - (!) Restart the notebook for the changes to take place
Check the documentation for other settings that can be customized.
The function cosine_similarity expects 2 arrays of shape (n_samples_X, n_features).
When using it to calculate a similarity of vectors A and B, you may get error:
ValueError: Expected 2D array, got 1D array instead:
array=[1. 1. 1.].
Reshape your data either using array.reshape(-1, 1) if your data has a single feature or array.reshape(1, -1) if it contains a single sample.
In that case, either use reshape as advised in the error message or convert each vector into
an artificial 2D array [A] and [B]:
from sklearn.metrics.pairwise import cosine_similarity
cosine_similarity([A], [B])