Kristian Flikka
Kristian Flikka
Tag names are not open, and should not be in the repo. Examples: https://github.com/equinor/gordo/blob/9e3e511067910540b1ba4dd1c47c5974e5acb2ec/tests/gordo/workflow/test_config_elements.py#L41 https://github.com/equinor/gordo/blob/0fb93beb0bc175ee7773bc23a1f981467e5ba315/tests/gordo/workflow/test_workflow_generator/data/config-test-with-log-key.yml#L6 https://github.com/equinor/gordo/blob/a21f254dcc616f387be1ae51d0082d8d5449f1a8/tests/gordo/machine/dataset/data_provider/test_data_provider_datalake.py#L20 And this, particulary nasty one, with both loads of tags and more https://github.com/equinor/gordo/blob/3bf760be2cc6f912695a69c34774c03e81fc5cf8/examples/metadata.json
https://github.com/equinor/gordo/blob/e04c446f04488394de6ea8f30cde9c40b6869ed2/gordo/client/client.py#L543 Previously, the row_filter was not used when doing predictions. After https://github.com/equinor/gordo/commit/994a9f805a8a5081c05a56735043365f0b38e1da#diff-e7f1b0bbb4d76747e62b1fb5d978a1feL505 I believe it is. The argument for not using the filter at prediction time is that it is...
Today, one either uses an unauthenticated endpoint, or supplies a Session object. It could be even easier, by supplying directly to the Gordo client the stuff needed to obtain the...
Today, you supply a data provider. But some users we have would like to directly provide the needed "dataframe" for a given Gordo model when doing .predict() Note - be...
It _appears_ that the reason quite a few of the tests in tests/ert/unit_tests/services/test_storage_service.py takes very long time (and some fail) is that the socket.getfqdn() command takes long time (or perhaps...
**Issue** Resolves #12236 **Approach** Revert back the IP selection with private and public, but use public by default This paves way for a configurable site-config variable. Adds a test that...