Vectice Python API cheat sheet

This API guide provides a quick overview of Vectice's simple Python API calls to help you get started auto-documenting your datasets, models, and notes.

Be sure to create your API key first!

Install Vectice

Start by installing the Vectice library.

# Install Vectice latest package
pip install vectice

# Install a specific version of the Vectice package
pip install vectice==<version number>

Connect to Vectice

Get started by connecting to the Vectice API and starting an iteration.

Connect with API key

#import and connect to Vectice
import vectice
connect = vectice.connect(api_token="your-api-key") #Paste your API key

# Connect to your project phase using your phase ID
phase = connect.phase("PHA-XXX") #You can fetch the relevant phase ID from your chosen Vectice project in the app.

#Create an iteration
iteration = phase.create_or_get_current_iteration()

Connect with JSON file

#import and connect to Vectice
import vectice
connect = vectice.connect(config="your-api-config.json") #Enter your API key JSON file

# Connect to your project phase using your phase ID
phase = connect.phase("PHA-XXX") #You can fetch the relevant phase ID from your chosen Vectice project in the app.

#Create an iteration
iteration = phase.create_or_get_current_iteration()

Auto-document your assets

Auto-document your notes, datasets, and models directly to Vectice.

Auto-document NOTES

# Auto-document your first comments or notes
iteration.log("this is a comment")

Auto-document DATASETS

For instructions on using these resources, refer to the Vectice API Reference guide's Resources section.

Local file

from vectice import FileResource, Dataset

# Auto-document your first dataset from a local file
file_resource = FileResource(paths="my/file/path", dataframe=your_df)
clean_dataset = Dataset.clean(resource=file_resource, name="your_dataset_name")
iteration.log(clean_dataset)

BigQuery

from vectice import BigQueryResource, Dataset

# Auto-document your first dataset from BigQuery
bq_resource = BigQueryResource(paths="your-bq-table", dataframes=your_df)
clean_dataset = Dataset.clean(resource=bq_resource, name="your_dataset_name")
iteration.log(clean_dataset)

S3

from vectice import S3Resource, Dataset

# Auto-document your first dataset from S3
s3_resource = S3Resource(uris="s3://.../<file_path_inside_bucket>", dataframes=your_df)
clean_dataset = Dataset.clean(resource=s3_resource, name="your_dataset_name")
iteration.log(clean_dataset)

GCS

from vectice import GCSResource, Dataset

# Auto-document your first dataset from GCS
gcs_resource = GCSResource(uris="gs://.../<file_path_inside_bucket>", dataframes=your_df)
clean_dataset = Dataset.clean(gcs_resource, name, attachments)
iteration.log(clean_dataset)

Databricks

from vectice import DatabricksTableResource, Dataset

# Auto-document your first dataset from Databricks
db_resource = DatabricksTableResource(paths="my-table", dataframes=your_df)
clean_dataset = Dataset.clean(resource=gcs_resource, name="your_dataset_name")
iteration.log(clean_dataset)

Auto-document MODELS

Default

from vectice import Model

# Auto-document your first model
model = Model(metrics, properties, attachments, predictor)
iteration.log(model)

MLFlow

import mlflow
from mlflow.client import MlflowClient
from vectice import Model

# Auto-document your first model
model = Model.mlflow(
        run_id="your-mlflow-run-id",
        client=MlflowClient() #also work with client=mlflow
    )
iteration.log(model)

Close your iteration

Once you are done logging your assets for an iteration, mark it complete.

# Completes and closes the current iteration once you are happy with it
iteration.complete()