Skip to content

migrating data results in wrong paths in the dataset metadata #636

@rokroskar

Description

@rokroskar

This was reported by @jen-thomas -

after migrating the dataset metadata and running renku dataset export zenodo:

File "[...]/site-packages/git/util.py", line 333, in finalize_process
    proc.wait(**kwargs)
  File "[...]/site-packages/git/cmd.py", line 412, in wait
    raise GitCommandError(self.args, status, errstr)
git.exc.GitCommandError: Cmd('git') failed due to: exit code(128)
  cmdline: git rev-list HEAD -- ../../../data/glonass/gpsdata-20161227.log
  stderr: 'fatal: ../../../data/glonass/gpsdata-20161227.log: '../../../data/glonass/gpsdata-20161227.log' is outside repository

Here is a small example to reproduce this:

renku init test-migration
cd test-migration
pipx run --spec renku==0.3.3 renku dataset create mydata
touch datafile
git add datafile
git commit -m 'added datafile'
pipx run --spec renku==0.3.3 renku dataset add mydata datafile
pipx run --spec renku==0.5.2 renku migrate datasets
git show

...

+files:
+- '@type':
+  - prov:Entity
+  - schema:DigitalDocument
+  - wfprov:Artifact
+  _id: blob/UNCOMMITTED/datafile
+  _label: datafile@UNCOMMITTED
+  _project: null
+  added: '2019-08-23 08:19:49.420919'
+  creator:
+  - '@type': schema:Person
+    _id: mailto:rok.roskar@sdsc.ethz.ch
+    affiliation: null
+    alternate_name: null
+    email: rok.roskar@sdsc.ethz.ch
+    name: Rok Roskar
+  dataset: null
+  path: ../../../data/mydata/datafile
+  url: null

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions