Query on filesize?

[ghost]

Is it possible to query on file size.
E.g. what if I would like to download only full tiles, which are over 750 MB?

[j08lue]

E.g. what if I would like to download only full tiles, which are over 750 MB?

For that I would use the footprint area. Something like this (not tested), in the case of Sentinel-2:

import sentinelsat
import shapely.wkt

api = sentinelsat.SentinelAPI(**credentials)
aoi_geom = shapely.wkt.loads(aoi)
products = api.query(area=aoi, **query_parameters)
gdf = api.to_geodataframe(products).to_crs({"init": "epsg:3857"})  # get metadata in some UTM crs
cond = gdf.intersection(aoi_geom).area > 0.5e10  # only take scenes covering more than half a tile
api.download_all(gdf[cond])

[j08lue]

But to answer your original question:

Is it possible to query on file size.

Have a look at the metadata that the query returns. Probably the file size is in there?

[ghost]

@j08lue hmm, like you implementation.
There was a comment here about OData example (https://sentinelsat.readthedocs.io/en/stable/api.html#odata-example)
Gonna try to fetch File Size from that.

[valgur]

You can filter by file size, but it is not implemented well on the server side. The values are strings like 1.5 GB and 653 MB and they are handled as strings when trying to filter by this field as well.

The closest thing to what you are trying to achieve would be

api.query(..., size="*MB")

The proper file size is only available from the OData API, unfortunately, and you will have to query it separately for each product.
See https://sentinelsat.readthedocs.io/en/stable/api.html#odata-example

Further filtering the returned product metadatas by the size field in Pandas, similar to what @j08lue suggested, would probably be the best approach.

[ghost]

@valgur Trying it, the OData way what you wrote.
But there are other problems as well:

products_df = api.to_dataframe(products)
## this one does not work, Pandas way
filtered_products = products_df[products_df['size'] >= 400]
## the string type have to converted a lot ('772.39 MB')
filesize = products_df['size'][0]

[valgur]

You can do

products_df['size'] = products_df['size'].str.split(' ').apply(lambda x: float(x[0]) * {'GB': 1e3, 'MB': 1, 'KB': 1e-3}[x[1]])
filtered_products = products_df.query('size >= 400')

to get only products >= 400 MB in size, for example.

[ghost]

@valgur Cool, best workaround :)