Read Endpoints
Note that access to the Comet REST endpoints require the
{'Authorization': ${COMET_API_KEY}}
header. To get your
API key, see Comet Python API Setup.
Get User Workspaces¶
Get a user's workspaces.
Example Response
{ "workspaceNames": [ "team name", "another team name" ] }
Get Projects¶
Get a workspace's projects.
Example Response
{ "projects": [ { "isPublic": "[boolean, true if this is a public project]", "lastUpdated": "[long, last time this project was updated in the DB]", "numberOfExperiments": "[integer, number of experiments in this project]", "ownerUserName": "Owner User Name", "projectDescription": "Project Description", "projectId": "aProjectId", "projectName": "Project Name", "workspaceId": "Team/workspace this project is in", "workspaceName": "Name of the team for team_id above" } ] }
Get Registry Models¶
Get a workspace's registry models.
Example Response
{ "registryModels": [ { "createdAt": "[long, when this model was created in the DB]", "description": "someDescription", "isPublic": "[Boolean]", "lastUpdated": "[long, last time this model was updated in the DB]", "latestVersion": { "comment": "someComment", "createdAt": "[long, when this model item was created in the DB]", "experimentModel": { "experimentKey": "someExperimentKey", "experimentModelId": "someExperimentModelId", "experimentModelName": "someExperimentModelName" }, "lastUpdated": "[long, last time this model item was updated in the DB]", "metadata": [ "metadata from first asset", "metadata from second asset" ], "registryModelItemId": "someRegistryModelItemId", "stages": [ "production", "staging" ], "userName": "someUserName", "version": "someVersion" }, "modelName": "someModelName", "registryModelId": "someRegistryModelId", "userName": "someUserName" } ] }
Get Registry Model Count¶
Get a count of a workspace's registry models.
Example Response
{ "registryModelCount": 7 }
Get Registry Model Details¶
Get a registry model's details.
Example Response
{ "createdAt": "[long, when this model was created in the DB]", "description": "someDescription", "isPublic": "[Boolean]", "lastUpdated": "[long, last time this model was updated in the DB]", "modelName": "someModelName", "registryModelId": "someRegistryModelId", "userName": "someUserName", "versions": [ { "assets": [ { "assetId": "someAssetId", "audio": "[Boolean, whether the asset is an audio file]", "canView": "[Boolean, whether the asset is viewable as an image]", "createdAt": "[Long, timestamp asset was created in DB]", "dir": "someDirectory", "fileName": "someFileName", "fileSize": "[Long, file size]", "histogram": "[Boolean, whether the asset is a histogram file]", "image": "[Boolean, whether the asset was stored as an image]", "link": "link to download asset file", "metadata": "Metadata associated with the asset", "runContext": "someRunContext", "step": "[Integer, step asset was logged during]", "type": "the type of asset" } ], "comment": "someComment", "createdAt": "[long, when this model item was created in the DB]", "experimentModel": { "experimentKey": "someExperimentKey", "experimentModelId": "someExperimentModelId", "experimentModelName": "someExperimentModelName" }, "lastUpdated": "[long, last time this model item was updated in the DB]", "registryModelItemId": "someRegistryModelItemId", "stages": [ "production", "staging" ], "userName": "someUserName", "version": "someVersion" } ] }
Get Registry Model Notes¶
Get a registry model's notes.
Example Response
{ "notes": "Potentially very long notes string" }
Download Registry Model Version¶
Download a zip file of a registry model version.
Example Response
"[Raw stream of zip file as octet-stream]"
Get Project¶
Get details on a project.
Example Response
{ "isPublic": "[boolean, true if this is a public project]", "lastUpdated": "[long, last time this project was updated in the DB]", "numberOfExperiments": "[integer, number of experiments in this project]", "ownerUserName": "Owner User Name", "projectDescription": "Project Description", "projectId": "aProjectId", "projectName": "Project Name", "workspaceId": "Team/workspace this project is in", "workspaceName": "Name of the team for team_id above" }
Get Experiments¶
Get a project's experiments.
Example Response
{ "experiments": [ { "durationMillis": "[long, How long the experiment ran for]", "endTimeMillis": "[long, As recorded by the backend, when this experiment ended]", "experimentKey": "SomeExperimentKey", "experimentName": "SomeExperimentName", "optimizationId": "optimization run associated with this experiment, if any", "startTimeMillis": "[long, As recorded by the backend, when this experiment started]" } ] }
Get Project Columns¶
Get a project's columns. These are the items that can be queried.
Example Response
{ "columns": [ { "id": "unique-column-id", "name": "Column Name", "source": "one of: metrics, params, log_other, metadata, tag, or env_details", "type": "one of: double, string, datetime, boolean, or timenumber" } ] }
Apply Project Query¶
Apply a query to a project.
Example Request
{ "archived": "[boolean: if true search for archived experiments. Optional]", "predicates": "Json Predicates for search here, example below", "projectId": "someProjectId", "projectName": "someProjectName", "workspaceName": "someWorkspaceName" }
Example Response
{ "experimentKeys": [ "KEY", "anotherExperimentKey" ], "experimentsInProject": "[int count of experiments in project, including all those filtered out. If archived was true this is the archived experiment count]" }
[ { "source": "metrics", "query": { "condition": "AND", "rules": [ { "id": "columnNameA", "field": "columnNameA", "type": "double", "input": "text", "operator": "greater", "value": 1 } ], "valid": true } } ]
Get Project Notes¶
Get a project's notes.
Example Response
{ "notes": "someProjectNotes" }
Get Multi Metric Chart¶
Get multi-metric chart data.
Example Request
{ "fetchFull": "[Boolean, whether to fetch all values (up to 15,000) or a sampled subset (about 500 points)]", "independentMetrics": "[Boolean, whether the metrics should be fetched individually or as a correlated whole (only return values for steps for which you have values for every requested metric name)]", "metrics": [ "someMetricName", "anotherMetricName" ], "params": [ "someParameterName", "anotherParameterName" ], "targetedExperiments": [ "KEY", "anotherExperimentKey" ] }
Example Response
{ "empty": "[Boolean, whether any metrics were returned]", "experiments": { "KEY": { "experimentKey": "KEY", "metrics": [ { "durations": [ "[Long, time between each value and the previous value, 0 for the first]" ], "metricName": "someMetricName", "steps": [ 1, 2, 3 ], "timestamps": [ "[Long, list of timestamps of metric values]" ], "values": [ 1.2, 1.3, 1.4 ] } ], "params": { "anotherParam": "whatever value you want", "someParam": "8675309" }, "steps": [ 1, 2, 3 ] } } }
Get Experiment Metadata¶
Get an experiment's metadata.
Example Response
{ "durationMillis": "someDurationMillis", "endTimeMillis": "someEndTimeMillis", "experimentKey": "KEY", "experimentName": "someExperimentName", "optimizationId": "someOptimizationId", "projectId": "someProjectId", "projectName": "someProjectName", "startTimeMillis": "someStartTimeMillis", "workspaceName": "someWorkspaceName" }
Get Git Metadata¶
Get the git metadata of an experiment.
Example Response
{ "branch": "gitBranch", "origin": "gitOrigin", "parent": "gitParent", "root": "gitRoot", "user": "gitUser" }
Get Git Patch¶
Get the git patch of an experiment.
Example Response
"[Raw stream of git patch as octet-stream]"
Get Html¶
Get an experiment's HTML.
Example Response
{ "html": "Html for Experiment" }
Get Code¶
Get an experiment's source code.
Example Response
{ "code": "Code for Experiment" }
Get Output¶
Get an experiment's standard output and error.
Example Response
{ "output": "stdout and stderr output for Experiment, with '\n' delimiters" }
Get Graph¶
Get an experiment's model graph.
Example Response
{ "graph": "Graph for Experiment" }
Get All Metrics For Name¶
Get all metric values for a metric name.
Example Response
{ "metrics": [ { "epoch": "[long, epoch this metric is from]", "metricName": "someMetricName", "metricValue": "someMetricValue", "offset": "[Long, offset that this metric was sent as]", "runContext": "someRunContext", "step": "[long, step this metric is from]", "timestamp": "[long, time this metric was recorded]" } ] }
Get Metrics Summary¶
Get an experiment's metrics summary.
Example Response
{ "values": [ { "name": "someName", "runContextCurrent": "run context of last value", "runContextMax": "run context of max value", "runContextMin": "run context of min value", "stepCurrent": "[long, step current value is from]", "stepMax": "[long, step max value is from]", "stepMin": "[long, step min value is from]", "timestampCurrent": "[long, time last value was recorded]", "timestampMax": "[long, time max value was recorded]", "timestampMin": "[long, time min value was recorded]", "valueCurrent": "last value", "valueMax": "max value", "valueMin": "min value" } ] }
Get Parameters¶
Get an experiment's parameters summary.
Example Response
{ "values": [ { "name": "someName", "runContextCurrent": "run context of last value", "runContextMax": "run context of max value", "runContextMin": "run context of min value", "stepCurrent": "[long, step current value is from]", "stepMax": "[long, step max value is from]", "stepMin": "[long, step min value is from]", "timestampCurrent": "[long, time last value was recorded]", "timestampMax": "[long, time max value was recorded]", "timestampMin": "[long, time min value was recorded]", "valueCurrent": "last value", "valueMax": "max value", "valueMin": "min value" } ] }
Get Others¶
Get an experiment's others (logged with log-other
) summary.
Example Response
{ "values": [ { "name": "someName", "runContextCurrent": "run context of last value", "runContextMax": "run context of max value", "runContextMin": "run context of min value", "stepCurrent": "[long, step current value is from]", "stepMax": "[long, step max value is from]", "stepMin": "[long, step min value is from]", "timestampCurrent": "[long, time last value was recorded]", "timestampMax": "[long, time max value was recorded]", "timestampMin": "[long, time min value was recorded]", "valueCurrent": "last value", "valueMax": "max value", "valueMin": "min value" } ] }
Get Tags¶
Get an experiment's tags.
Example Response
{ "tags": [ "a tag", "another tag" ] }
Get System Details¶
Get an experiment's system details.
Example Response
{ "command": [ "myScript.py", "-arg", "theArgument" ], "executable": "/usr/local/bin/python", "experimentKey": "KEY", "gpuStaticInfoList": [ { "gpuIndex": "[Integer, index]", "name": "name", "powerLimit": "[Integer, max power used????]", "totalMemory": "[Integer, total memory used???]", "uuid": "someUniqueId" } ], "hostname": "hostname", "installedPackages": [ "package", "anotherPackage" ], "ip": "ip address", "logAdditionalSystemInfoList": [ { "key": "someKey", "value": "someValue" } ], "maxTotalMemory": "[double, max memory used]", "networkInterfaceIps": [ "ip", "anotherIp" ], "os": "os with version info", "osPackages": [ "package", "anotherPackage" ], "osType": "os experiment ran on", "pid": "[Integer, pid]", "pythonVersion": "python version", "pythonVersionVerbose": "python version with verbose flag", "systemMetricNames": [ "name", "anotherName" ], "user": "system username" }
Get Asset List¶
Get an experiment's asset list.
Example Response
{ "assets": [ { "assetId": "someAssetId", "audio": "[Boolean, whether the asset is an audio file]", "canView": "[Boolean, whether the asset is viewable as an image]", "createdAt": "[Long, timestamp asset was created in DB]", "dir": "someDirectory", "fileName": "someFileName", "fileSize": "[Long, file size]", "histogram": "[Boolean, whether the asset is a histogram file]", "image": "[Boolean, whether the asset was stored as an image]", "link": "link to download asset file", "metadata": "Metadata associated with the asset", "runContext": "someRunContext", "step": "[Integer, step asset was logged during]", "type": "the type of asset" } ] }
all
, unknown
, audio
, video
, image
, histogram3d
, histogram_combined_3d
, confusion-matrix
Get Asset¶
Get an asset.
Example Response
"[Raw binary stream of asset or JSON {'msg':'no patch for experiment','code':200,'data':null,'sdk_error_code':0}]"
Get Experiment Model List¶
Get the list of models created by a given experiment.
Example Response
{ "models": [ { "experimentKey": "KEY", "experimentModelId": "someExperimentModelId", "modelName": "someExperimentModelName" } ] }
Get Experiment Model Asset List¶
Get the list of assets for a particular experiment model.
Example Response
{ "assets": [ { "assetId": "someAssetId", "audio": "[Boolean, whether the asset is an audio file]", "canView": "[Boolean, whether the asset is viewable as an image]", "createdAt": "[Long, timestamp asset was created in DB]", "dir": "someDirectory", "fileName": "someFileName", "fileSize": "[Long, file size]", "histogram": "[Boolean, whether the asset is a histogram file]", "image": "[Boolean, whether the asset was stored as an image]", "link": "link to download asset file", "metadata": "Metadata associated with the asset", "runContext": "someRunContext", "step": "[Integer, step asset was logged during]", "type": "the type of asset" } ] }
Download Experiment Model¶
Download a zip file of an experiment model.
Example Response
"[Raw stream of zip file as octet-stream]"
Get Store¶
Get an entire store.
Example Response
"{'store': '{key1: value1, ...}'}"
Get Account Details¶
Get account details from authenticated user's apiKey in header
Example Response
"{'userName': 'USERNAME', 'defaultWorkspaceName': 'WORKSPACE'}"