CLI: generalize 'map lookup' to lookup many identifiers at once
ClosedPublic
Actions

Authored by zack on Nov 30 2019, 2:50 PM.

Details

Reviewers

seirl

Group Reviewers

Reviewers

Maniphest Tasks

T2112: make "swh graph map lookup" accept lists of identifiers

Commits

rDGRPH71ce98054b4e: CLI: generalize 'map lookup' to lookup many identifiers at once

Summary

Multiple identifiers (of either kind) can be passed either directly on the CLI
or via stdin. In the latter case logical lines in stidn will be preserved in
stdout.

Closes T2112

Diff Detail

Repository

rDGRPH Compressed graph representation

Lint

Automatic diff as part of commit; lint not applicable.

Unit

Automatic diff as part of commit; unit tests not applicable.

Event Timeline

Updating D2379: CLI: generalize 'map lookup' to lookup many identifiers at once

zack edited the summary of this revision. (Show Details)Nov 30 2019, 2:52 PM

zack added a task: T2112: make "swh graph map lookup" accept lists of identifiers.

Build is green
See https://jenkins.softwareheritage.org/job/DGRPH/job/tox/112/ for more details.

Harbormaster completed remote builds in B9457: Diff 8384.Nov 30 2019, 2:52 PM

Build is green
See https://jenkins.softwareheritage.org/job/DGRPH/job/tox/113/ for more details.

Harbormaster completed remote builds in B9458: Diff 8385.Nov 30 2019, 2:54 PM

seirl requested changes to this revision.Dec 4 2019, 2:32 PM

seirl added inline comments.

swh/graph/cli.py
213	Isn't there an overhead to the `mmap` call here? Couldn't the mappings be created in the closure instead maybe?
229	`for line in sys.stdin:`
232	You're going to get a trailing space on all lines here, it might be significant at some point. It should be possible to `str.join` instead?

This revision now requires changes to proceed.Dec 4 2019, 2:32 PM

zack marked 3 inline comments as done.Dec 6 2019, 12:08 PM

zack added inline comments.

swh/graph/cli.py
229	(WTH was I thinking?)
232	Done. Note that, as a consequence of this change, logical lines that contain non resolvable IDs will become ambiguous in the output feed (e.g., if you 4 IDs and one failed to resolve, you cannot determine which one failed to in the output). IMO that's fine, as inputs with non resolvable IDs are doomed to fail anyway.