Closed amberthomas closed 5 years ago
Thanks for making a note, I am going to tag this as an enhancement to look for clip keyword and then match from the metadata. Should have a resolution in the next release
Hi @amberthomas Seems the fix was easy by simple stripping out _metadata. Now, this tool should be able to handle both clipped and unclipped files from the metadata. Upgrade to ppipe v0.4.5 for applying the fix
I think it's similar to the last issue "metadata file produced from files downloaded with the clip option of the old planet API aren't correct to use"
When I batch download PSOrthotiles the id names in the metadata file will be something like "1539866_0466621_2018-07-02_1105_BGRN_Analytic_metadata" which means that it won't upload properly when using selupload. I fixed my problem by just clipping the "_metadata" off in cli_metadata.py near line 88.
writer.writerow([filename.split('.')[0].replace('_metadata',''),epoch,productType,orbit,provider,instrument,satellite_id,tile_id,bands,epsg_code,resampling_kernel.....
This is super specific to my case though. I know Planet uses annoyingly different naming conventions across the board, so this would be tedious to find all these cases.
Would it worth doing something like:
metadata_sfx = filename.split('.')[0].split('_')[-1]
and checking against a list of common suffixes, then replacing with the empty string if it is in the list?I could also see queries being done against the metadata dictionary (near line 15 of metadata_ingest.py ) to accommodate the fact that filename is likely just a substring of a metadata key. But you read from the metadata csv a few times in that file so that might not be enough.