# LitmusEdgeManager 2.31.x API Documentation/Edge Lifecycle Management/AI/ML Models - LE, LEM, LUNS API Docs ## Step 1 of Upload New AI/ML Model **POST** `{{LEM_URL}}/api/v1/file/{{project_id}}/mlModel/motionModel.zip/upload-url` # Step 1 of Upload New AI/ML Model Step 1 of a three-step pre-signed-URL upload flow. LEM returns an upload URL signed by its object storage; step 2 uploads the file to that URL. ## Endpoint ```http POST {{LEM_URL}}/api/v1/file/{{project_id}}/mlModel/motionModel.zip/upload-url ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Request body ```json { "projectId": "{{project_id}}", "fileName": "motionModel.zip" } ``` ## Response `200 OK` -- `application/json` ```json { "url": "https://10.17.3.103/files/.../motionModel.zip?..." } ``` ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Request Body ```json { "projectId": "{{project_id}}", "fileName": "motionModel.zip" } ``` ### Response **Status**: 200 OK ```json { "url": "https://10.17.3.103/files/462j1mq70mph31wa100t1yuxa/mlModel/motionModel.zip?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ZBE3Hey7R88GtiYw%2F20241218%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20241218T141532Z&X-Amz-Expires=604800&X-Amz-SignedHeaders=host&X-Amz-Signature=8a7fd062731b25a2bd3887e11fa790a5e26436430ee1509424108b0252e494ca" } ``` --- ## Step 2 of Upload New AI?ML Model **PUT** `{{LEM_URL}}/files/{{project_id}}/mlModel/motionModel.zip?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ZBE3Hey7R88GtiYw%2F20241218%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20241218T141532Z&X-Amz-Expires=604800&X-Amz-SignedHeaders=host&X-Amz-Signature=8a7fd062731b25a2bd3887e11fa790a5e26436430ee1509424108b0252e494ca` # Step 2 of Upload New AI/ML Model Step 2: PUT the file bytes to the pre-signed URL returned by step 1. This call goes **directly to the storage backend**, not the LEM API. (The Postman item's name has a typo: `AI?ML`.) ## Endpoint ```http PUT {{LEM_URL}}/files/{{project_id}}/mlModel/motionModel.zip? ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Request body Raw binary upload of the model ZIP. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Response **Status**: 200 OK --- ## Step 3 of Upload New Model **POST** `{{LEM_URL}}/api/v1/file/{{project_id}}/mlModel/motionModel.zip` # Step 3 of Upload New Model Step 3: notify LEM that the upload is complete. LEM records the metadata and makes the file available for `Deploy ML model`. ## Endpoint ```http POST {{LEM_URL}}/api/v1/file/{{project_id}}/mlModel/motionModel.zip ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Request body ```json { "description": "" } ``` ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Request Body ```json { "description": "" } ``` ### Response **Status**: 200 OK ```json { "code": "S", "message": null, "messageDetails": null } ``` --- ## Deploy ML model **POST** `{{LEM_URL}}/api/v1/async-task/task-with-subtasks` # Deploy ML Model Deploys a staged AI/ML model to one or more devices via an async task. Body is the same async-task envelope as `Applications > Launch Application to Edge Device`; the task type is `ANALYTICS_ML_MODEL_UPLOAD`. ## Endpoint ```http POST {{LEM_URL}}/api/v1/async-task/task-with-subtasks ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Request body ```json { "name": "Upload AI/ML Models (1 edge device)", "projectId": "{{project_id}}", "subtasks": [ { "deviceId": "{{edge_device_id}}", "num": 0, "params": { "fileName": "motionModel.zip" }, "type": "ANALYTICS_ML_MODEL_UPLOAD" } ], "type": "ANALYTICS_ML_MODEL_UPLOAD", "waitForOnline": false } ``` ## Response `200 OK` -- `application/json` ```json { "taskId": "ca549da1-e645-4f21-8865-487cb5a4aac1" } ``` ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Request Body ```json { "name": "Upload AI/ML Models (1 edge device)", "description": "", "params": {}, "projectId": "{{project_id}}", "subtasks": [ { "deviceId": "{{edge_device_id}}", "num": 0, "params": { "fileName": "motionModel.zip" }, "type": "ANALYTICS_ML_MODEL_UPLOAD", "plannedExecTime": 0 } ], "type": "ANALYTICS_ML_MODEL_UPLOAD", "waitForOnline": false } ``` ### Response **Status**: 200 OK ```json { "taskId": "ca549da1-e645-4f21-8865-487cb5a4aac1" } ``` --- ## Verify If Deploy Task is Complete **GET** `{{LEM_URL}}/api/v1/async-task/{{task_id}}/subtasks` # Verify If Deploy Task is Complete Returns the subtasks for one async task with their progress, status, and per-device messages. Use to monitor a deployment. ## Endpoint ```http GET {{LEM_URL}}/api/v1/async-task/{{task_id}}/subtasks ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Response **Status**: 200 OK ```json [ { "id": "5dda70d0-5d2a-489f-85d9-c3169620f1f1", "taskId": "20bb4e40-5656-43c5-880b-4cf0fa9c4d07", "num": 0, "deviceId": "vkrih85ka6fn7zcofirk20vs", "devicePrettyName": "VM", "type": "ANALYTICS_ML_MODEL_UPLOAD", "params": { "fileName": "motionModel.zip" }, "plannedExecTime": 0, "progress": 1, "status": "SUCCESS", "msg": null, "startTime": 1734532298000, "stopTime": 1734532298000, "context": null, "createdOn": 1734532294000, "updatedOn": 1734532298000, "createdBy": "parthshah", "updatedBy": "parthshah", "done": true } ] ``` --- ## Get Models **GET** `{{LEM_URL}}/api/v1/analytics/{{project_id}}/models` # Get Models Returns the AI/ML models currently installed on every device in the project. Each device record carries `totalSize` plus an array of `models[]` with name, absolute path, and size. ## Endpoint ```http GET {{LEM_URL}}/api/v1/analytics/{{project_id}}/models ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Response **Status**: 200 OK ```json { "pageNum": 0, "pagesCount": 1, "size": 3, "totalSize": 3, "elements": [ { "projectId": "462j1mq70mph31wa100t1yuxa", "deviceId": "1a5wiavluwxuhwai5leb1v0hc", "hostname": "litmus-edge-94c6911b9c77", "prettyName": "Litmus Edge (94:c6:91:1b:9c:77)- SJ office sensor", "isOnline": true, "totalSize": 141852176, "models": [ { "name": "coco", "absolutePath": "/var/lib/loopedge-analytics2/models/coco", "size": 136145199, "deploymentSource": "LE" }, { "name": "gasFive.tflite", "absolutePath": "/var/lib/loopedge-analytics2/models/gasFive.tflite", "size": 101132, "deploymentSource": "LE" }, { "name": "motionModel", "absolutePath": "/var/lib/loopedge-analytics2/models/motionModel", "size": 5094648, "deploymentSource": "LE" }, { "name": "savedmodel4", "absolutePath": "/var/lib/loopedge-analytics2/models/savedmodel4", "size": 511197, "deploymentSource": "LE" }, { "name": ".ipynb_checkpoints", "absolutePath": "/var/lib/loopedge-analytics2/models/.ipynb_checkpoints", "size": 0, "deploymentSource": "LE" } ], "online": true }, { "projectId": "462j1mq70mph31wa100t1yuxa", "deviceId": "3y0vcdju38zvqjrlm5lcpm1l", "hostname": "litmus-edge-080027b4a1c8", "prettyName": "Litmus Edge (08:00:27:b4:a1:c8)", "isOnline": false, "totalSize": 5094648, "models": [ { "name": "motionModel", "absolutePath": "/var/lib/loopedge-analytics2/models/motionModel", "size": 5094648, "deploymentSource": "LE" } ], "online": false }, { "projectId": "462j1mq70mph31wa100t1yuxa", "deviceId": "vkrih85ka6fn7zcofirk20vs", "hostname": "litmus-edge-005056b965c2", "prettyName": "VM", "isOnline": true, "totalSize": 0, "models": [], "online": true } ] } ``` --- ## Models List **GET** `{{LEM_URL}}/api/v1/analytics/{{project_id}}/model-list` # Models List Returns a flat **aggregated** list of model names across the project's devices, with total size and the number of devices running each model. ## Endpoint ```http GET {{LEM_URL}}/api/v1/analytics/{{project_id}}/model-list ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Response `200 OK` -- `application/json`. Array of `{name, size, count}` objects. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Response **Status**: 200 OK ```json [ { "name": ".ipynb_checkpoints", "size": 0, "count": 1 }, { "name": "savedmodel4", "size": 511197, "count": 1 }, { "name": "coco", "size": 136145199, "count": 1 }, { "name": "gasFive.tflite", "size": 101132, "count": 1 }, { "name": "motionModel", "size": 5094648, "count": 2 } ] ``` --- ## Get Models on Edge Manager **GET** `{{LEM_URL}}/api/v1/file/{{project_id}}/list` # Get Models on Edge Manager Returns the AI/ML models **uploaded to LEM** (not yet deployed). These are staging files in LEM's file store ready for deployment. ## Endpoint ```http GET {{LEM_URL}}/api/v1/file/{{project_id}}/list ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ### Response **Status**: 200 OK ```json { "pageNum": 0, "pagesCount": 1, "size": 1, "totalSize": 1, "elements": [ { "name": "motionModel.zip", "size": 0, "lastModified": 1734531634590, "fileType": "mlModel", "description": "" } ] } ``` --- ## Delete Model on LEM **DELETE** `{{LEM_URL}}/api/v1/file/{{project_id}}/{fileType}/{fileName}` # Delete Model on LEM Removes a staged file from LEM's file store. Pass `fileType` (`mlModel`) and `fileName` in the path. ## Endpoint ```http DELETE {{LEM_URL}}/api/v1/file/{{project_id}}/{fileType}/{fileName} ``` ## Authentication HTTP Basic Auth. **Username** is your API token, **password** is empty. Tokens are managed under `System > Access Control > Tokens`. ## Errors | HTTP status | When it happens | |---------------------|--------------------------------------------------------------------------------| | `400 Bad Request` | Missing or malformed query/body parameter. | | `401 Unauthorized` | Missing or invalid credentials. | | `403 Forbidden` | Token lacks permission for this operation. | | `404 Not Found` | Target entity does not exist. | | `5xx` | Service is unreachable, restarting, or internally errored. Inspect device logs under `System > Support`. | > **TLS note**: edge devices use a self-signed certificate by default. Either install the device CA in your client trust store or disable certificate verification when calling this endpoint directly. ---