45 lines
1.5 KiB
Gherkin
45 lines
1.5 KiB
Gherkin
Feature: HuggingFace Existence Checks
|
|
As a data engineer
|
|
I want to skip datasets that already exist on HuggingFace
|
|
So that I avoid redundant uploads
|
|
|
|
Background:
|
|
Given a configured HuggingFace organization
|
|
|
|
Scenario: Skip dataset that exists on HuggingFace
|
|
Given "wordnet" already exists on HuggingFace
|
|
When I check if "wordnet" exists on HF
|
|
Then the result should be True
|
|
And the repo ID should be set
|
|
|
|
Scenario: Process dataset that does not exist
|
|
Given "new-dataset" does not exist on HuggingFace
|
|
When I check if "new-dataset" exists on HF
|
|
Then the result should be False
|
|
And the repo ID should be None
|
|
|
|
Scenario: Check dataset with -v1 suffix
|
|
Given a configured HuggingFace organization
|
|
Given "wordnet-v1" exists on HuggingFace
|
|
When I check if "wordnet" exists on HF
|
|
Then the result should be True
|
|
And the repo ID should be "test-organization/wordnet-v1"
|
|
|
|
Scenario: Handle missing organization configuration
|
|
Given no organization is configured
|
|
When I check if "wordnet" exists on HF
|
|
Then the result should be False
|
|
|
|
Scenario: Handle authentication errors gracefully
|
|
Given HuggingFace authentication fails
|
|
When I check if "wordnet" exists on HF
|
|
Then the result should be False
|
|
|
|
Scenario: Skip processing when dataset exists
|
|
Given "wordnet" already exists on HuggingFace
|
|
And sheets tracking is enabled
|
|
When I process "wordnet"
|
|
Then processing should be skipped
|
|
And the sheet status should remain "DONE"
|
|
|