Additional diamond compatibility #56
Conversation
I am suggesting a few changes that I think may be useful for installing eggnog-mapper via bioconda and using it in a galaxy installation. This allows the data dir to be independent of the installation directory. Also allows a path for the diamond db.
return value for get_db_present(level) should be a boolean
|
Also, this is a small data directory, hopefully something like what you and @jj-umn were talking about in #44. Using the capability that JJ added to specify the data directory at runtime, the single sequence in Although still big (~352MB), it is at least more downloadable for unit testing and Galaxy integration testing. It seems like further reduction in size would have to come from manually re-creating the hmmpressed files, which are the only truly big thing. Any comments/PRs on what else could be useful in a small database would be useful. Or suggestions about how to pare down the HMM. |
|
thanks for this. I need time to review all changes, but I'll let you know soon. |
This fixes an option clash with JJ's diamond database option (-d), which was also used for specific reference databases (e.g., homNOG). Also, I added try-catch exception handling to the diamond call, to expose diamond-specific errors when the call fails - I had some trouble figuring out where errors were.
This builds upon JJ's work to make fewer paths hard-coded.
I also updated the diamond version to the most recent (0.9.9)