API Methods

This section describes each method provided by the dAIEdgeVLabAPI client API along with the expected inputs (arguments) and outputs (return values).

---

Client Initialization and Authentication

__init__(config_file=None, url=None, port=None, auto_refresh=False)

• Input:

  • config_file (str, optional): Path to a YAML configuration file.
  • url (str, optional): API host address (used if config_file is not provided).
  • port (int, optional): API port (used if config_file is not provided).
  • auto_refresh (bool, optional): Whether to automatically start the token refresh thread upon login (default is False).

• Output:

  • Initializes the API client instance.
  • Raises an exception if neither a configuration file nor both url and port are provided, or if required configuration fields are missing.
  • Automatically checks the API version and, if using a configuration file, attempts to log in (with optional auto-refresh enabled).

---

_check_version()

• Input: None.
• Output:
  • Checks if the API version is '0.1' and that the API status is 'OK'.
  • Raises an exception if the API version is unsupported or the API status is not OK.

---

login(username=None, password=None, auto_refresh=False) -> bool

• Input:

  • username (str, optional): User email to use for login.
  • password (str, optional): User password to use for login.
  • auto_refresh (bool, optional): Whether to start the token refresh thread after login (default is False).

• Output:

  • Returns True if login is successful and the access token is obtained.
  • Returns False if login fails (and logs an error message).
  • Raises an exception if no username or password is provided.

---

_refresh_token()

• Input: None.
• Output:
  • Runs on a separate daemon thread.
  • Periodically refreshes the access token every 10 minutes.
  • If an error occurs during refresh, logs the error and stops the refresh thread.

---

refresh() -> bool

• Input: None.
• Output:
  • Refreshes the access token using the current refresh token.
  • Returns True if the refresh is successful; otherwise, logs an error and raises an exception.

---

logout()

• Input: None.
• Output:
  • Invalidates the access token.
  • Stops the refresh thread, ensuring a clean shutdown of the client.

---

API Status

getAPIStatus() -> dict

• Input: None.
• Output:
  • Returns a dictionary with the API status information, including details such as the API version and overall status.

---

Benchmark Operations

startBenchmark(target, runtime, model_path, dataset=None, callback=None) -> str

• Input:

  • target (str): Identifier for the target device.
  • runtime (str): Runtime engine to use.
  • model_path (str): Path to the model file.
  • dataset (str or binary file, optional): Either a dataset name (string) or a binary file object.
  • callback (function, optional): A function that will be called with the benchmark result when the job completes.

• Output:

  • Returns the benchmark job ID (str).
  • If a callback is provided and the request is successful, a separate thread is started to wait for the benchmark result and pass it to the callback.

---

notifyOnBenchmarkResult(id, callback) -> None

• Input:

  • id (str): The benchmark job ID.
  • callback (function): A function to be called with the benchmark result.

• Output:

  • Does not return a value; waits for the benchmark to complete and then calls the provided callback with the result.

---

getBenchmarkStatus(id) -> dict

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a dictionary with the current status details of the benchmark job.

---

waitBenchmarkResult(id, interval=2, verbose=False) -> dict

• Input:

  • id (str): The benchmark job ID.
  • interval (int, optional): Time in seconds between status polls (default is 2 seconds).
  • verbose (bool, optional): If True, logs status updates during polling.

• Output:

  • Returns the complete benchmark result as a dictionary once the job status is success, canceled, or failed.

---

getBenchmarkResult(id) -> dict

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a dictionary aggregating:
    • report: Detailed benchmark report.
    • user_log: User log text.
    • error_log: Error log text.
    • raw_output: Binary output (bytes).

---

getBenchmarkReport(id) -> dict

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a dictionary with detailed benchmark report information.

---

getAvailableConfigurations() -> dict

• Input: None.
• Output:
  • Returns a dictionary containing the available benchmark runner configurations.

---

getBenchmarks() -> dict

• Input: None.
• Output:
  • Returns a dictionary containing all benchmark jobs associated with the authenticated user.
  • Returns None if the request fails.

---

getBenchmarkInfo(id) -> dict

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a dictionary with general information about the specified benchmark job.
  • Returns None if the request fails.

---

Logs and Output Retrieval

getErrorLog(id) -> str

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a string containing the error log for the specified benchmark job.
  • Returns None if the error log is not available.

---

getInfoLog(id) -> str

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a string containing the user log for the specified benchmark job.
  • Returns None if the user log is not available.

---

getBinaryOutput(id) -> bytes

• Input:

  • id (str): The benchmark job ID.

• Output:

  • Returns a bytes object containing the binary output of the benchmark job.
  • Returns None if the binary output is not available.

---

Dataset Operations

getDatasets() -> dict

• Input: None.
• Output:
  • Returns a dictionary containing information about the available datasets from the API.

---

uploadDataset(path) -> str

• Input:

  • path (str): The file path of the dataset to be uploaded.

• Output:

  • Returns the base name of the uploaded file if the upload is successful.
  • Returns None if the request fails.

---

deleteDataset(name) -> bool

• Input:

  • name (str): The name of the dataset to be deleted.

• Output:

  • Returns True if the dataset is successfully deleted.
  • Returns False otherwise.