After you have defined Flow Artifacts, you can inspect artifacts produced by executions and discover bookmarked artifacts within a project.
Flow Artifacts are surfaced in a dedicated section in the run view, where they can be easily discovered, inspected and downloaded.
Navigate to Flows > Flow name > Run name > Artifacts in order to inspect the artifacts produced by that particular run.
Bookmark artifacts
As many flows are executed, specific flow executions may generate outputs that are especially noteworthy to the project. These Flow Artifacts can be bookmarked by clicking on the Bookmark artifact icon on the right panel:
Flow Artifacts can also be bookmarked in bulk by selecting multiple artifacts on the left panel, and clicking on the Bookmark button above:
Access artifacts
Artifacts that have smaller file sizes may be viewed directly within Domino. The right panel provides two additional means of accessing artifacts, either by downloading them or by accessing them via code.
-
The Download button will download all the files that are part of the selected artifact.
In this example, all five files (
model.pkl
,python_env.yaml
,requirements.txt
,conda.yaml
, andMLmodel
) are downloaded at once: -
The Code snippet button generates Python code to directly download stored artifacts to a local workspace. Copy the code using the Copy button and run it inside a workspace:
Artifacts that have already been bookmarked are displayed at the top level view for Flows to allow for easier discoverability and reuse.
-
Navigate to Flows > Bookmarked Artifacts, and view all the artifacts that were bookmarked on the current project, across flows and runs. This view includes the artifact name, type, and version, as well as additional metadata.
-
Clicking on the Source column will navigate to the run that created the artifact, while clicking on the Artifact column will display a view of the artifact that includes all of its files.
-
This paginated view allows filtering files by name, as well as downloading one or more selected files at once. Clicking on a file name will display a larger file preview, alongside an option to download the file.
-
Maximize the reproducibility of Flows.