Data model
| attribute | explanation |
|---|---|
| package | The package where the algorithm is located e.g. bnlearn |
| function | The specific function from the package |
| script_filename | file to execute |
| docker_image | name of docker image |
| description | describes the algorithm |
| needs_gpu | True, if it is a gpu enabled algorithm |
| valid_parameters | parameters to choose for the experiments |
| docker_parameters | parameters for the docker image |
| attribute | description |
|---|---|
| name | of the dataset |
| description | of the dataset |
| load_query | SQL query that produces the observation matrix |
| data_source | database where it is stored |
| time_created | creation time |
| content_hash | hash of the dataset, is used to detect changes |
| attribute | description |
|---|---|
| annotation | can be approved, declined or missing
|
| from_node_id | start node |
| result_id | associated result |
| to_node_id | end node |
| attribute | description |
|---|---|
| result_id | associated result (can be null for uploaded ground truth graphs) |
| from_node_id | start node |
| to_node_id | end node |
| weight | weight |
| is_ground_truth | true if this edge exists in the real world |
| attribute | description |
|---|---|
| dataset_id | associated dataset |
| name | name |
| algorithm_id | algorithm to execute |
| description | description |
| parameters | algorithm parameters |
| last_job (computed) | the last job object |
| execution_time_statistics |
min, max, mean, median, lower_quantile, upper_quantile of the job execution times |
| attribute | description |
|---|---|
| experiment_id | associated experiment |
| start_time | time of job start |
| container_id | which pos executed the job |
| node_hostname | which cluster node it was executed on |
| status |
waiting, running, done, error, cancelled or hidden
|
| log | The output of the algorithm |
| parallel | If true, can be executed with other jobs running. If false, no other job can be running in the whole environment |
| enforce_cpus | Enforce core requirements. Local development with minikube cannot offer enough resources so set it to false
|
| gpus | the number of needed gpus, only needed for gpu experiments |
| error_code |
UNSCHEDULABLE, IMAGE_NOT_FOUND or UNKNOWN
|
| attribute | description |
|---|---|
| name | node name |
| dataset_id | associated dataset |
| attribute | description |
|---|---|
| job_id | associated job |
| start_time | start time of job |
| end_time | finish time of job |
| execution_time | actual execution time |
| dataset_loading_time | time needed to transfer the dataset |
| meta_results | meta results |
| attribute | description |
|---|---|
| result_id | associated result |
| from_node_id | start node |
| to_node_id | end node |
| statistic | statistic of the seperation set |
| level | size of the seperation set |