SHA-1 parameterization working directory name

[FredLoney]

Make a SHA-1 digest directory name for long parameterizations.

[satra]

this looks good - adding a couple of items for thought/implementation

• a test
• a config option (so that this doesn't happen by default - or does happen and people can turn it of
• make sure datasink doesn't really on the filepath
• why not reduce it to a single sha1 based on the entire path? this would render the working directory unreadable, but that would be ok in these circumstances.

[FredLoney]

Thanks, Satra. I will make the change a config option and add a test.

I doubt if datasink would rely on the path, since it is a manufactured name. I will test this.

Reducing the entire path loses the hierarchical directory structure. I am hesitant to do so, since:

1. I don't know if there are dependencies on this structure, or how to convincingly test this.
2. There is value in having a directory name that is not completely opaque.

[chrisgorgo]

It seems that some things got mixed up in this pull request. It's probably because you have not used branches separating different features/bugfixes. We would still love to see these improvements in master, but please separate them into branches. Thanks!

[FredLoney]

Hi Chris,

I thought you'd pull it into a branch. I will rebase and branch the change.

Fred

From: Chris Filo Gorgolewski <notifications@github.commailto:notifications@github.com>
Reply-To: nipy/nipype <reply@reply.github.commailto:reply@reply.github.com>
Date: Wednesday, May 1, 2013 11:28 AM
To: nipy/nipype <nipype@noreply.github.commailto:nipype@noreply.github.com>
Cc: Me <loneyf@ohsu.edumailto:loneyf@ohsu.edu>
Subject: Re: [nipype] SHA-1 parameterization working directory name (#535)

It seems that some things got mixed up in this pull request. It's probably because you have not used branches separating different features/bugfixes. We would still love to see these improvements in master, but please separate them into branches. Thanks!

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/535#issuecomment-17297542.

[chrisgorgo]

Hi,
Thanks for rebasing, but it seems that some ANTS and dcmstack related commits are still mixed up. Could you look into it and separate them into different PRs? It seems that the current thing is a bit of a mess. What would be the simplest way to clean it up (@satra correct me if I'm wrong) is to make sure your master branch is up to date, create a brand new branch and cherry pick only the commits related to hash directies into it. Then you can send a new PR on that branch.

[FredLoney]

Hi Chris,

My understanding is that the pull request was withdrawn pending isolation into a branch and adding a unit test. The FredLoney/nipype fork contains six branches merged into master which enable nipype to work in our environment. PR #535 is fixed in the sha-digest-dirname branch, which is rebased as of a week ago. The fork master should not be merged into nipy/nipype. I plan to submit a pull request specifying only the PR branch. Other branches implement different enhancements for which a PR has not yet been filed. These might result in separate pull requests.

Fred

From: Chris Filo Gorgolewski <notifications@github.commailto:notifications@github.com>
Reply-To: nipy/nipype <reply@reply.github.commailto:reply@reply.github.com>
Date: Saturday, June 1, 2013 5:19 PM
To: nipy/nipype <nipype@noreply.github.commailto:nipype@noreply.github.com>
Cc: Me <loneyf@ohsu.edumailto:loneyf@ohsu.edu>
Subject: Re: [nipype] SHA-1 parameterization working directory name (#535)

Hi,
Thanks for rebasing, but it seems that some ANTS and dcmstack related commits are still mixed up. Could you look into it and separate them into different PRs? It seems that the current thing is a bit of a mess. What would be the simplest way to clean it up (@satrahttps://github.com/satra correct me if I'm wrong) is to make sure your master branch is up to date, create a brand new branch and cherry pick only the commits related to hash directies into it. Then you can send a new PR on that branch.

—
Reply to this email directly or view it on GitHubhttps://github.com//pull/535#issuecomment-18799515.

[chrisgorgo]

Ok thanks - looking forward to your PRs!