In the new version of mlprimitives we upgrade tensorflow, keras, and pandas. After upgrading pandas to its newer version >=1. Some functions have been deprecated.pd.Series now needs to be converted to numpy before treating it as such. For example in the following snippet, errors is a numpy array which is then converted into pandas.Series but is treated still as a numpy array:
above = pd.Series(errors > epsilon) index_above = np.argwhere(above)
https://github.com/signals-dev/Orion/blob/master/orion/primitives/timeseries_anomalies.py#L221.
To fix this we could go two routes:
pandas.Series to numpy using to_numpy() before treating it as such.above = pd.Series(errors > epsilon).to_numpy() index_above = np.argwhere(above) # or above = pd.Series(errors > epsilon) index_above = np.argwhere(above.values)
numpy from the start (might require more changes).Similarly, this needs to be addressed in all primitives that rely on pandas.Series.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4