If you want to work out strandedness within a pipeline the maybe how_are_we_stranded_here would work? I believe it does pretty much exactly what @benformatics and @swbarnes2 said above re aligning and figuring out
Read the method section of the publication associated with your SRA files where they should list the kit used for the library preparation. Then based on the protocol you should be able to surmise what strand orientation is expected.
Otherwise as @swbarnes2 already stated. It is much easier to align a small subset of the FASTQ and identify the strand bias that way.