Caching and reusing intermediate results between workflows¶
Within certain data analysis fields, there are certain intermediate results that reoccur in exactly the same way in many analysis. For example, in bioinformatics, reference genomes or annotations are downloaded, and read mapping indexes are built. Since such steps are independent of the actual data or measurements that are analyzed, but still computationally or timely expensive to conduct, it has been common practice to externalize their computation and assume the presence of the resulting files before execution of a workflow.
From version 5.8.0 on, Snakemake offers a way to keep those steps inside the actual analysis without requiring from redundant computations. By hashing all steps, parameters, software stacks (in terms of conda environments or containers), and raw input required up to a certain step in a blockchain, Snakemake is able to recognize before the computation whether a certain result is already available in a central cache on the same system. Note that this is explicitly intended for caching results between workflows! There is no need to use this feature to avoid redundant computations within a workflow. Snakemake does this already out of the box.
Such caching has to be explitly activated per rule, which can be done via the command line interface. For example,
$ export SNAKEMAKE_OUTPUT_CACHE=/mnt/snakemake-cache/ $ snakemake --cache download_reference create_index
would instruct Snakemake to cache and reuse the results of the rules
The environment variable definition that happens in the first line (defining the location of the cache) should of course be done only once and system wide in reality.
When Snakemake is executed without a shared filesystem (e.g., in the cloud, see Cloud Support), the environment variable has to point to a location compatible with the given remote provider (e.g. an S3 or Google Storage bucket).
In any case, the provided location should be shared between all workflows of your group, institute or computing environment, in order to benefit from the reuse of previously obtained intermediate results.
Note that only rules with just a single output file are eligible for caching.
Also note that the rules need to retrieve all their parameters via the
params directive (except input files).
It is not allowed to directly use
config or any global variable in the shell command or script, because these are not captured in the hash (otherwise, reuse would be unnecessarily limited).
Also note that Snakemake will store everything in the cache as readable and writeable for all users on the system (except in the remote case, where permissions are not enforced and depend on your storage configuration). Hence, caching is not intended for private data, just for steps that deal with publicly available resources.
Finally, be aware that the implementation has to be considered experimental until this note is removed.