from edgeml import edgeml
import time
import math

Globals

READ_KEY = "YOUR_READ_KEY"  # Replace with your actual read key
WRITE_KEY = "YOUR_WRITE_KEY"  # Replace with your actual write key
BACKEND_URL = "YOUR_BACKEND_URL"  # Replace with your actual backend URL

Upload randomly generated data to the server using the edge-ml python library

To upload data to edge-ml, we can use the DatasetCollector

For this, we need to provide the following information:

Parameter	Description
url	The URL to the edge-ml instance.
write_key	The API key for writing into the system.
use_own_timestamps	If true, users can pass timestamps to the collection function. Otherwise, timestamps are set by the DatasetCollector.
timeSeries	An array containing the names of the time series to be used.
metaData	A dictionary with metadata. Must contain only key-value pairs where both keys and values are strings.

datasetName = "Example Dataset"
useOwnTimeStamps = False
timeSeries = ["Acc", "Mag"]
metaData = {}

collector = edgeml.DatasetCollector(BACKEND_URL,
                                    WRITE_KEY,
                                    datasetName,
                                    useOwnTimeStamps,
                                    timeSeries,
                                    metaData)

Now we can add data to this dataset

For this we can call the addDataPoint-function. Don't forget to call onComplete after inserting all the data.

timestamp = round(time.time() * 1000)
for i in range(100):
    timestamp += 40
    x = i / 10000        # Adjust the divisor to control the frequency of the wave
    y_acc = math.sin(x)  # Generate the y-coordinate for "Acc"
    y_mag = math.cos(x)  # Generate the y-coordinate for "Mag"
    await collector.addDataPoint(timestamp, "Acc", y_acc) 
    await collector.addDataPoint(timestamp, "Mag", y_mag) 

# signal data collection is complete. This uploads the remaining data to the server
collector.onComplete()

True

Retrieve data from edge-ml

It is also possible to obtain the datasets in a project. To do so use the DatasetReceiver

project = edgeml.DatasetReceiver(BACKEND_URL, READ_KEY)

# See a single dataset
print(project.datasets[0])
# Or get some attribute from the dataset
print(project.datasets[0].metaData)

# Until now, we have only the metdata of the datasets.
# We can also download the actual time-series data.

# Only for one timeSeries:
project.datasets[0].timeSeries[0].loadData()
# Or for one dataset:
project.datasets[0].loadData()
# Or for all datasets:
project.loadData()

Dataset - Name: Example Dataset, ID: 682ee7d8a3130d2327595758, Metadata: {}
{}

Get the data in the dataset

The datasets are provided as pandas dataframes

# Access the data of a dataset
print("Dataset")
print(project.datasets[0].data.head())

print("\nTimeseries")
# Or just one time series
print(project.datasets[0].timeSeries[0].data.head())

# Or get all dataset in a project as list
project_data = project.data
print("\n#datasts: ", len(project_data))

Dataset
                     time     Acc  Mag
0 2025-05-22 09:01:12.595  0.0000  1.0
1 2025-05-22 09:01:12.635  0.0001  1.0
2 2025-05-22 09:01:12.675  0.0002  1.0
3 2025-05-22 09:01:12.715  0.0003  1.0
4 2025-05-22 09:01:12.755  0.0004  1.0

Timeseries
                     time     Acc
0 2025-05-22 09:01:12.595  0.0000
1 2025-05-22 09:01:12.635  0.0001
2 2025-05-22 09:01:12.675  0.0002
3 2025-05-22 09:01:12.715  0.0003
4 2025-05-22 09:01:12.755  0.0004

#datasts:  11

Get the labels in the dataset

project.datasets[0].labelings

[]

Labeling in the project

To labelings in a project define the labels

project.labelings

[{'_id': '682f07918245a094a595cdf5',
  'name': 'test',
  'labels': [{'name': 't1',
    'color': '#0081DD',
    '_id': '682f07918245a094a595cdf3'},
   {'name': 't2', 'color': '#C24A5F', '_id': '682f07918245a094a595cdf4'}],
  'projectId': '682ec257f42749f02e3a325f'}]