It seems like the newer DwC-A files stores taxonomy keys in a string format. The current command ami-dataset fetch-images results in an empty dataframe if the input keys are in int or float. There are two possible fixes:
- Let the keys remain in str format or do not convert it into
int.
|
keys_list = [int(x) for x in keys_list] |
- Set the data type of the key columns to
int
|
occ_df = dwca.pd_read("occurrence.txt", parse_dates=True, on_bad_lines="skip") |
It seems like the newer DwC-A files stores taxonomy keys in a string format. The current command
ami-dataset fetch-imagesresults in an empty dataframe if the input keys are inintorfloat. There are two possible fixes:int.ami-ml/src/dataset_tools/fetch_images.py
Line 76 in 884044d
intami-ml/src/dataset_tools/utils.py
Line 23 in 884044d