Tamr Unify - Python Client¶

Example¶

from tamr_unify_client import Client
from tamr_unify_client.auth import UsernamePasswordAuth
import os

# grab credentials from environment variables
username = os.environ['UNIFY_USERNAME']
password = os.environ['UNIFY_PASSWORD']
auth = UsernamePasswordAuth(username, password)

host = 'localhost' # replace with your Tamr Unify host
unify = Client(auth, host=host)

# programmatically interace with Tamr Unify!
# e.g. refresh your project's Unified Dataset
project = unify.projects.by_resource_id('3')
ud = project.unified_dataset()
op = ud.refresh()
assert op.succeeded()

User Guide¶

FAQ¶

What version of the Python Client should I use?¶

If you are starting a new project or your existing project does not yet use the Python Client, we encourage you to use the latest stable version of the Python Client.

If you are already using the Python Client, you have 3 options:

“I like my project’s code the way it is.”

Keep using the version you are on.

“I want some new features released in versions with the same major version that I’m currently using.”

Upgrade to the latest stable version with the same major version as what you currently use.

“I want all new features and I’m willing to modify my code to get those features!”

Upgrade to the latest stable version even if it has a different major version from what you currently use.

Note that you do not need to reason about the Unify API version nor the the Unify version.

How does this the Python Client accomplish this?

The short answer is that the Python Client just cares about features, and will try everything it knows to implement those features correctly, independent of the API version.

We’ll illustrate with an example.

Let’s say you want to get a dataset by name in your Python code.

1. If no such feature exists, you can file a Feature Request. Note that the Python Client is limited by what the Unify API enables. So you should check if the Unify API docs to see if the feature you want is even possible.

2. If this feature already exists, you can try it out!

E.g. unify.datasets.by_name(some_dataset_name)

2.a It works! 🎉

2.b If it fails with an HTTP error, it could be for 2 reasons:

2.a.i It might be impossible to support that feature in the Python Client because your Unify API version does not have the necessary endpoints to support it.

2.a.ii Your Unify API version does support this feature with some endpoints, but the Python Client know how to correctly implement this feature for this version of the API. In this case, you should submit a Feature Request.

2.c If it fails with any other error, you should submit a Bug Report. 🐛

Note

To see how to submit Bug Reports / Feature Requests, see 🐛 Bug Reports / 🙋 Feature Requests.

To check what endpoints your version of the Unify API supports, see docs.tamr.com/reference (be sure to select the correct version in the top left!).

How do I call custom endpoints, e.g. endpoints outside the Unify API?¶

To call a custom endpoint within the Unify API, use the client.request() method, and provide an endpoint described by a path relative to base_path. For example, if base_path is /api/versioned/v1/ (the default), and you want to get /api/versioned/v1/projects/1, you only need to provide projects/1 (the relative ID provided by the project) as the endpoint, and the Client will resolve that into /api/versioned/v1/projects/1.

There are various APIs outside the /api/versioned/v1/ prefix that are often useful or necessary to call - e.g. /api/service/health, or other un-versioned / unsupported APIs. To call a custom endpoint outside the Unify API, use the client.request() method, and provide an endpoint described by an absolute path (a path starting with /). For example, to get /api/service/health (no matter what base_path is), call client.request() with /api/service/health as the endpoint. The Client will ignore base_path and send the request directly against the absolute path provided.

For additional detail, see Custom HTTP requests and Unversioned API Access.

Installation¶

tamr-unify-client is compatible with Python 3.6 or newer.

Stable releases¶

Installation is as simple as:

pip install tamr-unify-client

Or:

poetry add tamr-unify-client

Note

If you don’t use poetry, we recommend you use a virtual environment for your project and install the Python Client into that virtual environment.

You can create a virtual environment with Python 3 via:

python3 -m venv my-venv

For more, see The Hitchhiker’s Guide to Python .

Latest (unstable)¶

Note

This project uses the new pyproject.toml file, not a setup.py file, so make sure you have the latest version of pip installed: pip install -U pip.

To install the bleeding edge:

git clone https://github.com/Datatamer/unify-client-python
cd unify-client-python
pip install .

Offline installs¶

First, download tamr-unify-client and its dependencies on a machine with online access to PyPI:

pip download tamr-unify-client -d tamr-unify-client-requirements
zip -r tamr-unify-client-requirements.zip tamr-unify-client-requirements

Then, ship the .zip file to the target machine where you want tamr-unify-client installed. You can do this via email, cloud drives, scp or any other mechanism.

Finally, install tamr-unify-client from the saved dependencies:

unzip tamr-unify-client-requirements.zip
pip install --no-index --find-links=tamr-unify-client-requirements tamr-unify-client

If you are not using a virtual environment, you may need to specify the --user flag if you get permissions errors:

pip install --user --no-index --find-links=tamr-unify-client-requirements tamr-unify-client

Quickstart¶

Client configuration¶

Start by importing the Python Client and authentication provider:

from tamr_unify_client import Client
from tamr_unify_client.auth import UsernamePasswordAuth

Next, create an authentication provider and use that to create an authenticated client:

import os

username = os.environ['UNIFY_USERNAME']
password = os.environ['UNIFY_PASSWORD']

auth = UsernamePasswordAuth(username, password)
unify = Client(auth)

Warning

For security, it’s best to read your credentials in from environment variables or secure files instead of hardcoding them directly into your code.

For more, see User Guide > Secure Credentials .

By default, the client tries to find the Unify instance on localhost. To point to a different host, set the host argument when instantiating the Client.

For example, to connect to 10.20.0.1:

unify = Client(auth, host='10.20.0.1')

Top-level collections¶

The Python Client exposes 2 top-level collections: Projects and Datasets.

You can access these collections through the client and loop over their members with simple for-loops.

E.g.:

for project in unify.projects:
  print(project.name)

for dataset in unify.datasets:
  print(dataset.name)

Fetch a specific resource¶

If you know the identifier for a specific resource, you can ask for it directly via the by_resource_id methods exposed by collections.

E.g. To fetch the project with ID '1':

project = unify.projects.by_resource_id('1')

Resource relationships¶

Related resources (like a project and its unified dataset) can be accessed through specific methods.

E.g. To access the Unified Dataset for a particular project:

ud = project.unified_dataset()

Kick-off Unify Operations¶

Some methods on Model objects can kick-off long-running Unify operations.

Here, kick-off a “Unified Dataset refresh” operation:

operation = project.unified_dataset().refresh()
assert op.succeeded()

By default, the API Clients expose a synchronous interface for Unify operations.

Secure Credentials¶

This section discusses ways to pass credentials securely to UsernamePasswordAuth. Specifically, you should not hardcode your password(s) in your source code. Instead, you should use environment variables or secure files to store your credentials and simple Python code to read your credentials.

Environment variables¶

You can use os.environ to read in your credentials from environment variables:

# my_script.py
import os

from tamr_unify_client.auth import UsernamePasswordAuth

username = os.environ['UNIFY_USERNAME'] # replace with your username environment variable name
password = os.environ['UNIFY_PASSWORD'] # replace with your password environment variable name

auth = UsernamePasswordAuth(username, password)

You can pass in the environment variables from the terminal by including them before your command:

UNIFY_USERNAME="my Unify username" UNIFY_PASSWORD="my Unify password" python my_script.py

You can also create an .sh file to store your environment variables and simply source that file before running your script.

Config files¶

You can also store your credentials in a secure credentials file:

# credentials.yaml
---
username: "my unify username"
password: "my unify password"

Then pip install pyyaml read the credentials in your Python code:

# my_script.py
from tamr_unify_client.auth import UsernamePasswordAuth
import yaml

creds = yaml.load("path/to/credentials.yaml") # replace with your credentials.yaml path

auth = UsernamePasswordAuth(creds.username, creds.password)

As in this example, we recommend you use YAML as your format since YAML has support for comments and is more human-readable than JSON.

Important

You should not check these credentials files into your version control system (e.g. git). Do not share this file with anyone who should not have access to the password stored in it.

Workflows¶

Continuous Categorization¶

from tamr_unify_client import Client
from tamr_unify_client.auth import UsernamePasswordAuth
import os

username = os.environ['UNIFY_USERNAME']
password = os.environ['UNIFY_PASSWORD']
auth = UsernamePasswordAuth(username, password)

host = 'localhost' # replace with your host
unify = Client(auth)

project_id = "1" # replace with your project ID
project = unify.projects.by_resource_id(project_id)
project = project.as_categorization()

unified_dataset = project.unified_dataset()
op = unified_dataset.refresh()
assert op.succeeded()

model = project.model()
op = model.train()
assert op.succeeded()

op = model.predict()
assert op.succeeded()

Continuous Mastering¶

from tamr_unify_client import Client
from tamr_unify_client.auth import UsernamePasswordAuth
import os

username = os.environ['UNIFY_USERNAME']
password = os.environ['UNIFY_PASSWORD']
auth = UsernamePasswordAuth(username, password)

host = 'localhost' # replace with your host
unify = Client(auth)

project_id = "1" # replace with your project ID
project = unify.projects.by_resource_id(project_id)
project = project.as_mastering()

unified_dataset = project.unified_dataset()
op = unified_dataset.refresh()
assert op.succeeded()

op = project.pairs().refresh()
assert op.succeeded()

model = project.pair_matching_model()
op = model.train()
assert op.succeeded()

op = model.predict()
assert op.succeeded()

op = project.record_clusters().refresh()
assert op.succeeded()

op = project.published_clusters().refresh()
assert op.succeeded()

Geospatial Data¶

What geospatial data is supported?¶

In general, the Python Geo Interface is supported; see https://gist.github.com/sgillies/2217756

There are three layers of information, modeled after GeoJSON; see https://tools.ietf.org/html/rfc7946 :

The outermost layer is a FeatureCollection
Within a FeatureCollection are Features, each of which represents one “thing”, like a building or a river. Each feature has:
- type (string; required)
- id (object; required)
- geometry (Geometry, see below; optional)
- bbox (“bounding box”, 4 doubles; optional)
- properties (map[string, object]; optional)
Within a Feature is a Geometry, which represents a shape, like a point or a polygon. Each geometry has:
- type (one of “Point”, “MultiPoint”, “LineString”, “MultiLineString”, “Polygon”, “MultiPolygon”; required)
- coordinates (doubles; exactly how these are structured depends on the type of the geometry)

Although the Python Geo Interface is non-prescriptive when it comes to the data types of the id and properties, Unify has a more restricted set of supported types. See https://docs.tamr.com/reference#attribute-types

The Dataset class supports the __geo_interface__ property. This will produce one FeatureCollection for the entire dataset.

There is a companion iterator itergeofeatures() that returns a generator that allows you to stream the records in the dataset as Geospatial features.

To produce a GeoJSON representation of a dataset:

dataset = client.datasets.by_name("my_dataset")
with open("my_dataset.json", "w") as f:
  json.dump(dataset.__geo_interface__, f)

By default, itergeofeatures() will use the first dataset attribute with geometry type to fill in the feature geometry. You can override this by specifying the geometry attribute to use in the geo_attr parameter to itergeofeatures.

Dataset can also be updated from a feature collection that supports the Python Geo Interface:

import geopandas
geodataframe = geopandas.GeoDataFrame(...)
dataset = client.dataset.by_name("my_dataset")
dataset.from_geo_features(geodataframe)

By default the features’ geometries will be placed into the first dataset attribute with geometry type. You can override this by specifying the geometry attribute to use in the geo_attr parameter to from_geo_features.

Rules for converting from Unify records to Geospatial Features¶

The record’s primary key will be used as the feature’s id. If the primary key is a single attribute, then the value of that attribute will be the value of id. If the primary key is composed of multiple attributes, then the value of the id will be an array with the values of the key attributes in order.

Unify allows any number of geometry attributes per record; the Python Geo Interface is limited to one. When converting Unify records to Python Geo Features, the first geometry attribute in the schema will be used as the geometry; all other geometry attributes will appear as properties with no type conversion. In the future, additional control over the handling of multiple geometries may be provided; the current set of capabilities is intended primarily to support the use case of working with FeatureCollections within Unify, and FeatureCollection has only one geometry per feature.

An attribute is considered to have geometry type if it has type RECORD and contains an attribute named point, multiPoint, lineString, multiLineString, polygon, or multiPolygon.

If an attribute named bbox is available, it will be used as bbox. No conversion is done on the value of bbox. In the future, additional control over the handling of bbox attributes may be provided.

All other attributes will be placed in properties, with no type conversion. This includes all geometry attributes other than the first.

Rules for converting from Geospatial Features to Unify records¶

The Feature’s id will be converted into the primary key for the record. If the record uses a simple key, no value translation will be done. If the record uses a composite key, then the value of the Feature’s id must be an array of values, one per attribute in the key.

If the Feature contains keys in properties that conflict with the record keys, bbox, or geometry, those keys are ignored (omitted).

If the Feature contains a bbox, it is copied to the record’s bbox.

All other keys in the Feature’s properties are propagated to the same-name attribute on the record, with no type conversion.

Streaming data access¶

The Dataset method itergeofeatures() returns a generator that allows you to stream the records in the dataset as Geospatial features:

my_dataset = client.datasets.by_name("my_dataset")
for feature in my_dataset.itergeofeatures():
  do_something(feature)

Note that many packages that consume the Python Geo Interface will be able to consume this iterator directly. For example:

from geopandas import GeoDataFrame
df = GeoDataFrame.from_features(my_dataset.itergeofeatures())

This allows construction of a GeoDataFrame directly from the stream of records, without materializing the intermediate dataset.

Advanced Usage¶

Asynchronous Operations¶

You can opt-in to an asynchronous interface via the asynchronous keyword argument for methods that kick-off Unify operations.

E.g.:

operation = project.unified_dataset().refresh(asynchronous=True)
# do asynchronous stuff while operation is running
operation.wait() # hangs until operation finishes
assert op.succeeded()

Logging API calls¶

It can be useful (e.g. for debugging) to log the API calls made on your behalf by the Python Client.

You can set up HTTP-API-call logging on any client via standard Python logging mechanisms

from tamr_unify_client import Client
from unify_api_v1.auth import UsernamePasswordAuth
import logging

auth = UsernamePasswordAuth("username", "password")
unify = Client(auth)

# Reload the `logging` library since other libraries (like `requests`) already
# configure logging differently. See: https://stackoverflow.com/a/53553516/1490091
import imp
imp.reload(logging)

logging.basicConfig(
  level=logging.INFO, format="%(message)s", filename=log_path, filemode="w"
)
unify.logger = logging.getLogger(name)

By default, when logging is set up, the client will log {method} {url} : {response_status} for each API call.

You can customize this by passing in a value for log_entry:

def log_entry(method, url, response):
# custom logging function
# use the method, url, and response to construct the logged `str`
# e.g. for logging out machine-readable JSON:
import json
return json.dumps({
  "request": f"{method} {url}",
  "status": response.status_code,
  "json": response.json(),
})

# after configuring `unify.logger`
unify.log_entry = log_entry

Custom HTTP requests and Unversioned API Access¶

We encourage you to use the high-level, object-oriented interface offered by the Python Client. If you aren’t sure whether you need to send low-level HTTP requests, you probably don’t.

But sometimes it’s useful to directly send HTTP requests to Unify; for example, Unify has many APIs that are not covered by the higher-level interface (most of which are neither versioned nor supported). You can still call these endpoints using the Python Client, but you’ll need to work with raw Response objects.

Custom endpoint¶

The client exposes a request method with the same interface as requests.request:

# import Python Client library and configure your client

unify = Client(auth)
# do stuff with the `unify` client

# now I NEED to send a request to a specific endpoint
response = unify.request('GET', 'relative/path/to/resource')

This will send a request relative to the base_path registered with the client. If you provide an absolute path to the resource, the base_path will be ignored when composing the request:

# import Python Client library and configure your client

unify = Client(auth)

# request a resource outside the configured base_path
response = unify.request('GET', '/absolute/path/to/resource')

You can also use the get, post, put, delete convenience methods:

# e.g. `get` convenience method
response = unify.get('relative/path/to/resource')

Custom Host / Port / Base API path¶

If you need to repeatedly send requests to another port or base API path (i.e. not /api/versioned/v1/), you can simply instantiate a different client.

Then just call request as described above:

# import Python Client library and configure your client

unify = api.Client(auth)
# do stuff with the `unify` client

# now I NEED to send requests to a different host/port/base API path etc..
# NOTE: in this example, we reuse `auth` from the first client, but we could
# have made a new Authentication provider if this client needs it.
custom_client = api.Client(
  auth,
  host="10.10.0.1",
  port=9090,
  base_path="/api/some_service/",
)
response = custom_client.get('relative/path/to/resource')

One-off authenticated request¶

All of the Python Client Authentication providers adhere to the requests.auth.BaseAuth interface.

This means that you can pass in an Authentication provider directly to the requests library:

from tamr_unify_client.auth import UsernamePasswordAuth
import os
import requests

username = os.environ['UNIFY_USERNAME']
password =  os.environ['UNIFY_PASSWORD']
auth = UsernamePasswordAuth(username, password)

response = requests.request('GET', 'some/specific/endpoint', auth=auth)

Contributor Guide¶

Code of Conduct¶

See CODE_OF_CONDUCT.md

🐛 Bug Reports / 🙋 Feature Requests¶

Please leave bug reports and feature requests as Github issues .

Be sure to check through existing issues (open and closed) to confirm that the bug hasn’t been reported before.

Duplicate bug reports are a huge drain on the time of other contributors, and should be avoided as much as possible.

↪️ Pull Requests¶

For larger, new features:

Open an RFC issue . Discuss the feature with project maintainers to be sure that your change fits with the project vision and that you won’t be wasting effort going in the wrong direction.

Once you get the green light 🚦 from maintainers, you can proceed with the PR.

Contributions / PRs should follow the Forking Workflow :

Fork it: https://github.com/[your-github-username]/unify-client-python/fork
Create your feature branch:
git checkout -b my-new-feature
Commit your changes:
git commit -am 'Add some feature'
Push to the branch:
git push origin my-new-feature
Create a new Pull Request

We optimize for PR readability, so please squash commits before and during the PR review process if you think it will help reviewers and onlookers navigate your changes.

Don’t be afraid to push -f on your PRs when it helps our eyes read your code.

Install¶

This project uses poetry as its package manager. For details on poetry, see the official documentation .

Install pyenv:
curl https://pyenv.run | bash
Clone your fork and cd into the project:
git clone https://github.com/<your-github-username>/unify-client-python
cd unify-client-python
Use pyenv to install a compatible Python version (3.6 or newer; e.g. 3.7.3):
pyenv install 3.7.3
Set that Python version to be your version for this project(e.g. 3.7.3):
pyenv local 3.7.3
Check that your Python version matches the version specified in .python-version:
cat .python-version
python --version
Install poetry as described here:
curl -sSL https://raw.githubusercontent.com/sdispater/poetry/master/get-poetry.py | python
Install dependencies via poetry:
poetry install

Run tests¶

To run all tests:

poetry run pytest .

To run specific tests, see these pytest docs .

Run style checks¶

To run linter:

poetry run flake8 .

To run formatter:

poetry run black --check .

Run the formatter without the –check flag to fix formatting in-place.

Build docs¶

To build the docs:

cd docs/
poetry run make html

After docs are build, view them by:

cd docs/ # unless you are there already
open -a 'Google Chrome' _build/html/index.html # open in your favorite browser

Editor config¶

Atom :

Developer Interface¶

Authentication¶

class tamr_unify_client.auth.UsernamePasswordAuth(username, password)[source]¶

Provides username/password authentication for Unify. Specifically, sets the Authorization HTTP header with Unify’s custom BasicCreds format.

Parameters:	username (str) – password (str) –

Usage:

>>> from tamr_unify_client.auth import UsernamePasswordAuth
>>> auth = UsernamePasswordAuth('my username', 'my password')
>>> import tamr_unify_client as api
>>> unify = api.Client(auth)

Client¶

class tamr_unify_client.Client(auth, host='localhost', protocol='http', port=9100, base_path='/api/versioned/v1/', session=None)[source]¶

Python Client for Unify API. Each client is specific to a specific origin (protocol, host, port).

Parameters:

auth (requests.auth.AuthBase) – Unify-compatible Authentication provider. Recommended: use one of the classes described in Authentication
host (str) – Host address of remote Unify instance (e.g. 10.0.10.0). Default: ‘localhost’
protocol (str) – Either ‘http’ or ‘https’. Default: ‘http’
port (int) – Unify instance main port. Default: 9100
base_path (str) – Base API path. Requests made by this client will be relative to this path. Default: ‘api/versioned/v1/’
session (requests.Session) – Session to use for API calls. Default: A new default requests.Session().

Usage:

>>> import tamr_unify_client as api
>>> from tamr_unify_client.auth import UsernamePasswordAuth
>>> auth = UsernamePasswordAuth('my username', 'my password')
>>> local = api.Client(auth) # on http://localhost:9100
>>> remote = api.Client(auth, protocol='https', host='10.0.10.0') # on https://10.0.10.0:9100

origin¶

HTTP origin i.e. <protocol>://<host>[:<port>]. For additional information, see MDN web docs .

Type:	str

request(method, endpoint, **kwargs)[source]¶

Sends an authenticated request to the server. The URL for the request will be "<origin>/<base_path>/<endpoint>".

Parameters:	method (str) – The HTTP method for the request to be sent. endpoint (str) – API endpoint to call (relative to the Base API path for this client).
Returns:	HTTP response
Return type:	`requests.Response`

get(endpoint, **kwargs)[source]¶: Calls request() with the "GET" method.

post(endpoint, **kwargs)[source]¶: Calls request() with the "POST" method.

put(endpoint, **kwargs)[source]¶: Calls request() with the "PUT" method.

delete(endpoint, **kwargs)[source]¶: Calls request() with the "DELETE" method.

projects¶

Collection of all projects on this Unify instance.

Returns:	Collection of all projects.
Return type:	`ProjectCollection`

datasets¶

Collection of all datasets on this Unify instance.

Returns:	Collection of all datasets.
Return type:	`DatasetCollection`

Attribute¶

class tamr_unify_client.attribute.resource.Attribute(client, data, alias=None)[source]¶

A Unify Attribute.

See https://docs.tamr.com/reference#attribute-types

relative_id¶

Type:	str

name¶

Type:	str

description¶

Type:	str

type¶

Type:	`AttributeType`

is_nullable¶

Type:	bool

resource_id¶

Type:	str

Attribute Collection¶

class tamr_unify_client.attribute.collection.AttributeCollection(client, api_path)[source]¶

Collection of Attribute s.

Parameters:	client (`Client`) – Client for API call delegation. api_path (str) – API path used to access this collection. E.g. `"datasets/1/attributes"`.

by_resource_id(resource_id)[source]¶

Retrieve an attribute by resource ID.

Parameters:	resource_id (str) – The resource ID. E.g. `"AttributeName"`
Returns:	The specified attribute.
Return type:	`Attribute`

by_relative_id(relative_id)[source]¶

Retrieve an attribute by relative ID.

Parameters:	relative_id (str) – The resource ID. E.g. `"datasets/1/attributes/AttributeName"`
Returns:	The specified attribute.
Return type:	`Attribute`

by_external_id(external_id)[source]¶

Retrieve an attribute by external ID.

Since attributes do not have external IDs, this method is not supported and will raise a NotImplementedError .

Parameters:	external_id (str) – The external ID.
Returns:	The specified attribute, if found.
Return type:	`Attribute`
Raises:	KeyError – If no attribute with the specified external_id is found LookupError – If multiple attributes with the specified external_id are found

stream()[source]¶

Stream attributes in this collection. Implicitly called when iterating over this collection.

Returns:	Stream of attributes.
Return type:	Python generator yielding `Attribute`

Usage:

>>> for attribute in collection.stream(): # explicit
>>>     do_stuff(attribute)
>>> for attribute in collection: # implicit
>>>     do_stuff(attribute)

by_name(attribute_name)[source]¶

Lookup a specific attribute in this collection by exact-match on name.

Parameters:	attribute_name (str) – Name of the desired attribute.
Returns:	Attribute with matching name in this collection.
Return type:	`Attribute`
Raises:	KeyError – If no attribute with specified name was found.

create(creation_spec)[source]¶

Create an Attribute in this collection

Parameters:	creation_spec (dict[str, str]) – Attribute creation specification should be formatted as specified in the Public Docs for adding an Attribute.
Returns:	The created Attribute
Return type:	`Attribute`

Attribute Type¶

class tamr_unify_client.attribute.type.AttributeType(data)[source]¶

The type of an Attribute or SubAttribute.

See https://docs.tamr.com/reference#attribute-types

Parameters:	data (`dict`) – JSON data representing this type

base_type¶

Type:	str

inner_type¶

Type:	`AttributeType`

attributes¶

Type:	list[`SubAttribute`]

SubAttribute¶

class tamr_unify_client.attribute.subattribute.SubAttribute(data)[source]¶

An attribute which is itself a property of another attribute.

See https://docs.tamr.com/reference#attribute-types

Parameters:	data (`dict`) – JSON data representing this attribute

name¶

Type:	str

description¶

Type:	str

type¶

Type:	`AttributeType`

is_nullable¶

Type:	bool

Categorization¶

Categorization Project¶

class tamr_unify_client.categorization.project.CategorizationProject(client, data, alias=None)[source]¶

A Categorization project in Unify.

model()[source]¶

Machine learning model for this Categorization project. Learns from verified labels and predicts categorization labels for unlabeled records.

Returns:	The machine learning model for categorization.
Return type:	`MachineLearningModel`

create_taxonomy(creation_spec)[source]¶

Creates a Taxonomy for this project.

A taxonomy cannot already be associated with this project.

Parameters:	creation_spec (dict) – The creation specification for the taxonomy, which can include name.
Returns:	The new Taxonomy
Return type:	`Taxonomy`

taxonomy()[source]¶

Retrieves the Taxonomy associated with this project. If a taxonomy is not already associated with this project, call create_taxonomy() first.

Returns:	The project’s Taxonomy
Return type:	`Taxonomy`

add_input_dataset(dataset)¶

Associate a dataset with a project in Unify.

By default, datasets are not associated with any projects. They need to be added as input to a project before they can be used as part of that project

Parameters:	dataset (`Dataset`) – The dataset to associate with the project.
Returns:	HTTP response from the server
Return type:	`requests.Response`

as_categorization()¶

Convert this project to a CategorizationProject

Returns:	This project.
Return type:	`CategorizationProject`
Raises:	TypeError – If the `type` of this project is not `"CATEGORIZATION"`

as_mastering()¶

Convert this project to a MasteringProject

Returns:	This project.
Return type:	`MasteringProject`
Raises:	TypeError – If the `type` of this project is not `"DEDUP"`

attribute_configurations()¶

Project’s attribute’s configurations.

Returns:	The configurations of the attributes of a project.
Return type:	`AttributeConfigurationCollection`

attribute_mappings()¶

Project’s attribute’s mappings.

Returns:	The attribute mappings of a project.
Return type:	`AttributeMappingCollection`

attributes¶

Attributes of this project.

Returns:	Attributes of this project.
Return type:	`AttributeCollection`

description¶

Type:	str

external_id¶

Type:	str

input_datasets()¶

Retrieve a collection of this project’s input datasets.

Returns:	The project’s input datasets.
Return type:	`DatasetCollection`

name¶

Type:	str

relative_id¶

Type:	str

resource_id¶

Type:	str

type¶

A Unify project type, listed in https://docs.tamr.com/reference#create-a-project.

Type:	str

unified_dataset()¶

Unified dataset for this project.

Returns:	Unified dataset for this project.
Return type:	`Dataset`

Category¶

class tamr_unify_client.categorization.category.resource.Category(client, data, alias=None)[source]¶

A category of a taxonomy

name¶

Type:	str

description¶

Type:	str

path¶

Type:	list[str]

parent()[source]¶

Gets the parent Category of this one, or None if it is a tier 1 category

Returns:	The parent Category or None
Return type:	`Category`

relative_id¶

Type:	str

resource_id¶

Type:	str

Category Collection¶

class tamr_unify_client.categorization.category.collection.CategoryCollection(client, api_path)[source]¶

Collection of Category s.

Parameters:	client (`Client`) – Client for API call delegation. api_path (str) – API path used to access this collection. E.g. `"projects/1/taxonomy/categories"`.

by_resource_id(resource_id)[source]¶

Retrieve a category by resource ID.

Parameters:	resource_id (str) – The resource ID. E.g. `"1"`
Returns:	The specified category.
Return type:	`Category`

by_relative_id(relative_id)[source]¶

Retrieve a category by relative ID.

Parameters:	relative_id (str) – The relative ID. E.g. `"projects/1/categories/1"`
Returns:	The specified category.
Return type:	`Category`

by_external_id(external_id)[source]¶

Retrieve an attribute by external ID.

Since categories do not have external IDs, this method is not supported and will raise a NotImplementedError .

Parameters:	external_id (str) – The external ID.
Returns:	The specified category, if found.
Return type:	`Category`
Raises:	KeyError – If no category with the specified external_id is found LookupError – If multiple categories with the specified external_id are found

stream()[source]¶

Stream categories in this collection. Implicitly called when iterating over this collection.

Returns:	Stream of categories.
Return type:	Python generator yielding `Category`

Usage:

>>> for category in collection.stream(): # explicit
>>>     do_stuff(category)
>>> for category in collection: # implicit
>>>     do_stuff(category)

create(creation_spec)[source]¶

Creates a new category.

Parameters:	creation_spec (dict) – Category creation specification, formatted as specified in the Public Docs for Creating a Category.
Returns:	The newly created category.
Return type:	`Category`

bulk_create(creation_specs)[source]¶

Creates new categories in bulk.

Parameters:	creation_specs (iterable[dict]) – A collection of creation specifications, as detailed for create.
Returns:	JSON response from the server
Return type:	`dict`

Taxonomy¶

class tamr_unify_client.categorization.taxonomy.Taxonomy(client, data, alias=None)[source]¶

A project’s taxonomy

name¶

Type:	str

categories()[source]¶

Retrieves the categories of this taxonomy.

Returns:	A collection of the taxonomy categories.
Return type:	`CategoryCollection`

relative_id¶

Type:	str

resource_id¶

Type:	str

Dataset¶

class tamr_unify_client.dataset.resource.Dataset(client, data, alias=None)[source]¶

A Unify dataset.

name¶

Type:	str

external_id¶

Type:	str

description¶

Type:	str

version¶

Type:	str

tags¶

Type:	list[str]

key_attribute_names¶

Type:	list[str]

attributes¶

Attributes of this dataset.

Returns:	Attributes of this dataset.
Return type:	`AttributeCollection`

upsert_records(records, primary_key_name, **json_args)[source]¶

Creates or updates the specified records.

Parameters:	records (iterable[dict]) – The records to update, as dictionaries. primary_key_name (str) – The name of the primary key for these records, which must be a key in each record dictionary. **json_args – Arguments to pass to the JSON dumps function, as documented here. Some of these, such as indent, may not work with Unify.
Returns:	JSON response body from the server.
Return type:	dict

delete_records(records, primary_key_name)[source]¶

Deletes the specified records.

Parameters:	records (iterable[dict]) – The records to delete, as dictionaries. primary_key_name (str) – The name of the primary key for these records, which must be a key in each record dictionary.
Returns:	JSON response body from the server.
Return type:	dict

delete_records_by_id(record_ids)[source]¶

Deletes the specified records.

Parameters:	record_ids (iterable) – The IDs of the records to delete.
Returns:	JSON response body from the server.
Return type:	dict

delete_all_records()[source]¶

Removes all records from the dataset.

Returns:	HTTP response from the server
Return type:	`requests.Response`

refresh(**options)[source]¶

Brings dataset up-to-date if needed, taking whatever actions are required.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The refresh operation.
Return type:	`Operation`

profile()[source]¶

Returns profile information for a dataset.

If profile information has not been generated, call create_profile() first. If the returned profile information is out-of-date, you can call refresh() on the returned object to bring it up-to-date.

Returns:	Dataset Profile information.
Return type:	`DatasetProfile`

create_profile(**options)[source]¶

Create a profile for this dataset.

If a profile already exists, the existing profile will be brought up to date.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The operation to create the profile.
Return type:	`Operation`

records()[source]¶

Stream this dataset’s records as Python dictionaries.

Returns:	Stream of records.
Return type:	Python generator yielding `dict`

status()[source]¶

Retrieve this dataset’s streamability status.

Returns:	Dataset streamability status.
Return type:	`DatasetStatus`

usage()[source]¶

Retrieve this dataset’s usage by recipes and downstream datasets.

Returns:	The dataset’s usage.
Return type:	`DatasetUsage`

from_geo_features(features, geo_attr=None)[source]¶

Upsert this dataset from a geospatial FeatureCollection or iterable of Features.

features can be:

An object that implements __geo_interface__ as a FeatureCollection (see https://gist.github.com/sgillies/2217756)
An iterable of features, where each element is a feature dictionary or an object that implements the __geo_interface__ as a Feature
A map where the “features” key contains an iterable of features

See: geopandas.GeoDataFrame.from_features()

If geo_attr is provided, then the named Unify attribute will be used for the geometry. If geo_attr is not provided, then the first attribute on the dataset with geometry type will be used for the geometry.

Parameters:	features – geospatial features geo_attr (str) – (optional) name of the Unify attribute to use for the feature’s geometry

upstream_datasets()[source]¶

The Dataset’s upstream datasets.

API returns the URIs of the upstream datasets, resulting in a list of DatasetURIs, not actual Datasets.

Returns:	A list of the Dataset’s upstream datasets.
Return type:	list[`DatasetURI`]

itergeofeatures(geo_attr=None)[source]¶

Returns an iterator that yields feature dictionaries that comply with __geo_interface__

See https://gist.github.com/sgillies/2217756

Parameters:	geo_attr (str) – (optional) name of the Unify attribute to use for the feature’s geometry
Returns:	stream of features
Return type:	Python generator yielding `dict[str, object]`

relative_id¶

Type:	str

resource_id¶

Type:	str

Dataset Collection¶

class tamr_unify_client.dataset.collection.DatasetCollection(client, api_path='datasets')[source]¶

Collection of Dataset s.

Parameters:	client (`Client`) – Client for API call delegation. api_path (str) – API path used to access this collection. E.g. `"projects/1/inputDatasets"`. Default: `"datasets"`.

by_resource_id(resource_id)[source]¶

Retrieve a dataset by resource ID.

Parameters:	resource_id (str) – The resource ID. E.g. `"1"`
Returns:	The specified dataset.
Return type:	`Dataset`

by_relative_id(relative_id)[source]¶

Retrieve a dataset by relative ID.

Parameters:	relative_id (str) – The resource ID. E.g. `"datasets/1"`
Returns:	The specified dataset.
Return type:	`Dataset`

by_external_id(external_id)[source]¶

Retrieve a dataset by external ID.

Parameters:	external_id (str) – The external ID.
Returns:	The specified dataset, if found.
Return type:	`Dataset`
Raises:	KeyError – If no dataset with the specified external_id is found LookupError – If multiple datasets with the specified external_id are found

stream()[source]¶

Stream datasets in this collection. Implicitly called when iterating over this collection.

Returns:	Stream of datasets.
Return type:	Python generator yielding `Dataset`

Usage:

>>> for dataset in collection.stream(): # explicit
>>>     do_stuff(dataset)
>>> for dataset in collection: # implicit
>>>     do_stuff(dataset)

by_name(dataset_name)[source]¶

Lookup a specific dataset in this collection by exact-match on name.

Parameters:	dataset_name (str) – Name of the desired dataset.
Returns:	Dataset with matching name in this collection.
Return type:	`Dataset`
Raises:	KeyError – If no dataset with specified name was found.

create(creation_spec)[source]¶

Create a Dataset in Unify

Parameters:	creation_spec (dict[str, str]) – Dataset creation specification should be formatted as specified in the Public Docs for Creating a Dataset.
Returns:	The created Dataset
Return type:	`Dataset`

Dataset Profile¶

class tamr_unify_client.dataset.profile.DatasetProfile(client, data, alias=None)[source]¶

Profile info of a Unify dataset.

dataset_name¶

The name of the associated dataset.

Type:	str

relative_dataset_id¶

The relative dataset ID of the associated dataset.

Type:	str

is_up_to_date¶

Whether the associated dataset is up to date.

Type:	bool

profiled_data_version¶

The profiled data version.

Type:	str

profiled_at¶

Info about when profile info was generated.

Type:	dict

simple_metrics¶

Simple metrics for profiled dataset.

Type:	list

attribute_profiles¶

Simple metrics for profiled dataset.

Type:	list

refresh(**options)[source]¶

Updates the dataset profile if needed.

The dataset profile is updated on the server; you will need to call profile() to retrieve the updated profile.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The refresh operation.
Return type:	`Operation`

relative_id¶

Type:	str

resource_id¶

Type:	str

Dataset Status¶

class tamr_unify_client.dataset.status.DatasetStatus(client, data, alias=None)[source]¶

Streamability status of a Unify dataset.

dataset_name¶

The name of the associated dataset.

Type:	str

relative_dataset_id¶

The relative dataset ID of the associated dataset.

Type:	str

is_streamable¶

Whether the associated dataset is available to be streamed.

Type:	bool

relative_id¶

Type:	str

resource_id¶

Type:	str

Dataset URI¶

class tamr_unify_client.dataset.uri.DatasetURI(client, uri)[source]¶

Indentifier of a dataset.

Parameters:	client (`Client`) – Queried dataset’s client. uri (`str`) – Queried dataset’s dataset ID.

resource_id¶

Type:	str

relative_id¶

Type:	str

uri¶

Type:	str

dataset()[source]¶

Fetch the dataset that this identifier points to.

Returns: A Unify dataset.

Return type:

class:	~tamr_unify_client.dataset.resource.Dataset

Dataset Usage¶

class tamr_unify_client.dataset.usage.DatasetUsage(client, data, alias=None)[source]¶

The usage of a dataset and its downstream dependencies.

See https://docs.tamr.com/reference#retrieve-downstream-dataset-usage

relative_id¶

Type:	str

usage¶

Type:	`DatasetUse`

dependencies¶

Type:	list[`DatasetUse`]

resource_id¶

Type:	str

Dataset Use¶

class tamr_unify_client.dataset.use.DatasetUse(client, data)[source]¶

The use of a dataset in project steps. This is not a BaseResource because it has no API path and cannot be directly retrieved or modified.

See https://docs.tamr.com/reference#retrieve-downstream-dataset-usage

Parameters:	client (`Client`) – Delegate underlying API calls to this client. data (`dict`) – The JSON body containing usage information.

dataset_id¶

Type:	str

dataset_name¶

Type:	str

input_to_project_steps¶

Type:	list[`ProjectStep`]

output_from_project_steps¶

Type:	list[`ProjectStep`]

dataset()[source]¶

Retrieves the Dataset this use represents.

Returns:	The dataset being used.
Return type:	`Dataset`

Machine Learning Model¶

class tamr_unify_client.base_model.MachineLearningModel(client, data, alias=None)[source]¶

A Unify Machine Learning model.

train(**options)[source]¶

Learn from verified labels.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The resultant operation.
Return type:	`Operation`

predict(**options)[source]¶

Suggest labels for unverified records.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The resultant operation.
Return type:	`Operation`

relative_id¶

Type:	str

resource_id¶

Type:	str

Mastering¶

Binning Model¶

class tamr_unify_client.mastering.binning_model.BinningModel(client, data, alias=None)[source]¶

A binning model object.

records()[source]¶

Stream this object’s records as Python dictionaries.

Returns:	Stream of records.
Return type:	Python generator yielding `dict`

update_records(records)[source]¶

Send a batch of record creations/updates/deletions to this dataset.

Parameters:	records (iterable[dict]) – Each record should be formatted as specified in the Public Docs for Dataset updates.
Returns:	JSON response body from server.
Return type:	`dict`

relative_id¶

Type:	str

resource_id¶

Type:	str

Estimated Pair Counts¶

class tamr_unify_client.mastering.estimated_pair_counts.EstimatedPairCounts(client, data, alias=None)[source]¶

Estimated Pair Counts info for Mastering Project

is_up_to_date¶

Whether an estimate pairs job has been run since the last edit to the binning model.

Return type:	bool

total_estimate¶

The total number of estimated candidate pairs and generated pairs for the model across all clauses.

Returns:

A dictionary containing candidate pairs and estimated pairs mapped to their corresponding estimated counts. For example:

{

“candidatePairCount”: “54321”,
”generatedPairCount”: “12345”

}

Return type: dict[str, str]

clause_estimates¶

The estimated candidate pair count and generated pair count for each clause in the model.

Returns:

A dictionary containing each clause name mapped to a dictionary containing the corresponding estimated candidate and generated pair counts. For example:

{

“Clause1”: {

“candidatePairCount”: “321”,
”generatedPairCount”: “123”

},

”Clause2”: {

“candidatePairCount”: “654”,
”generatedPairCount”: “456”

}

}

Return type: dict[str, dict[str, str]]

refresh(**options)[source]¶

Updates the estimated pair counts if needed.

The pair count estimates are updated on the server; you will need to call estimate_pairs() to retrieve the updated estimate.

Parameters:	**options – Options passed to underlying `Operation` . See `apply_options()` .
Returns:	The refresh operation.
Return type:	`Operation`

relative_id¶

Type:	str

resource_id¶

Type:	str

Mastering Project¶

class tamr_unify_client.mastering.project.MasteringProject(client, data, alias=None)[source]¶

A Mastering project in Unify.

pairs()[source]¶

Record pairs generated by Unify’s binning model. Pairs are displayed on the “Pairs” page in the Unify UI.

Call refresh() from this dataset to regenerate pairs according to the latest binning model.

Returns:	The record pairs represented as a dataset.
Return type:	`Dataset`

pair_matching_model()[source]¶

Machine learning model for pair-matching for this Mastering project. Learns from verified labels and predicts categorization labels for unlabeled pairs.

Calling predict() from this dataset will produce new (unpublished) clusters. These clusters are displayed on the “Clusters” page in the Unify UI.

Returns:	The machine learning model for pair-matching.
Return type:	`MachineLearningModel`

high_impact_pairs()[source]¶

High-impact pairs as a dataset. Unify labels pairs as “high-impact” if labeling these pairs would help it learn most quickly (i.e. “Active learning”).

High-impact pairs are displayed with a ⚡ lightning bolt icon on the “Pairs” page in the Unify UI.

Call refresh() from this dataset to produce new high-impact pairs according to the latest pair-matching model.

Returns:	The high-impact pairs represented as a dataset.
Return type:	`Dataset`

record_clusters()[source]¶

Record Clusters as a dataset. Unify clusters labeled pairs using pairs model. These clusters populate the cluster review page and get transient cluster ids, rather than published cluster ids (i.e., “Permanent Ids”)

Call refresh() from this dataset to generate clusters based on to the latest pair-matching model.

Returns:	The record clusters represented as a dataset.
Return type:	`Dataset`

published_clusters()[source]¶

Published record clusters generated by Unify’s pair-matching model.

Returns:	The published clusters represented as a dataset.
Return type:	`Dataset`

published_clusters_configuration()[source]¶

Retrieves published clusters configuration for this project.

Returns:	The published clusters configuration
Return type:	`PublishedClustersConfiguration`

published_cluster_ids()[source]¶

Retrieves published cluster IDs for this project.

Returns:	The published cluster ID dataset.
Return type:	`Dataset`

published_cluster_stats()[source]¶

Retrieves published cluster stats for this project.

Returns:	The published cluster stats dataset.
Return type:	`Dataset`

published_cluster_versions(cluster_ids)[source]¶

Retrieves version information for the specified published clusters. See https://docs.tamr.com/reference#retrieve-published-clusters-given-cluster-ids.

Parameters:	cluster_ids (iterable[str]) – The persistent IDs of the clusters to get version information for.
Returns:	A stream of the published clusters.
Return type:	Python generator yielding `PublishedCluster`

record_published_cluster_versions(record_ids)[source]¶

Retrieves version information for the published clusters of the given records. See https://docs.tamr.com/reference#retrieve-published-clusters-given-record-ids.

Parameters:	record_ids (iterable[str]) – The Tamr IDs of the records to get cluster version information for.
Returns:	A stream of the relevant published clusters.
Return type:	Python generator yielding `RecordPublishedCluster`

estimate_pairs()[source]¶

Returns pair estimate information for a mastering project

Returns:	Pairs Estimate information.
Return type:	`EstimatedPairCounts`

record_clusters_with_data()[source]¶

Project’s unified dataset with associated clusters.

Returns:	The record clusters with data represented as a dataset
Return type:	`Dataset`

published_clusters_with_data()[source]¶

Project’s unified dataset with associated clusters.

Returns:	The published clusters with data represented as a dataset
Return type:	`Dataset`

binning_model()[source]¶

Binning model for this project.

Returns:	Binning model for this project.
Return type:	`BinningModel`

add_input_dataset(dataset)¶

Associate a dataset with a project in Unify.

By default, datasets are not associated with any projects. They need to be added as input to a project before they can be used as part of that project

Parameters:	dataset (`Dataset`) – The dataset to associate with the project.
Returns:	HTTP response from the server
Return type:	`requests.Response`

as_categorization()¶

Convert this project to a CategorizationProject

Returns:	This project.
Return type:	`CategorizationProject`
Raises:	TypeError – If the `type` of this project is not `"CATEGORIZATION"`

as_mastering()¶

Convert this project to a MasteringProject

Returns:	This project.
Return type:	`MasteringProject`
Raises:	TypeError – If the `type` of this project is not `"DEDUP"`

attribute_configurations()¶

Project’s attribute’s configurations.

Returns:	The configurations of the attributes of a project.
Return type:	`AttributeConfigurationCollection`

attribute_mappings()¶

Project’s attribute’s mappings.

Returns:	The attribute mappings of a project.
Return type:	`AttributeMappingCollection`

attributes¶

Attributes of this project.

Returns:	Attributes of this project.
Return type:	`AttributeCollection`

description¶

Type:	str

external_id¶

Type:	str

input_datasets()¶

Retrieve a collection of this project’s input datasets.

Returns:	The project’s input datasets.
Return type:	`DatasetCollection`

name¶

Type:	str

relative_id¶

Type:	str

resource_id¶

Type:	str

type¶

A Unify project type, listed in https://docs.tamr.com/reference#create-a-project.

Type:	str

unified_dataset()¶

Unified dataset for this project.

Returns:	Unified dataset for this project.
Return type:	`Dataset`

Published Cluster¶

Metric¶

class tamr_unify_client.mastering.published_cluster.metric.Metric(data)[source]¶

A metric for a published cluster.

This is not a BaseResource because it does not have its own API endpoint.

Parameters:	data – The JSON entity representing this cluster.

name¶

Type:	str

value¶

Type:	str

Published Cluster¶

class tamr_unify_client.mastering.published_cluster.resource.PublishedCluster(data)[source]¶

A representation of a published cluster in a mastering project with version information. See https://docs.tamr.com/reference#retrieve-published-clusters-given-cluster-ids.

This is not a BaseResource because it does not have its own API endpoint.

Parameters:	data – The JSON entity representing this `PublishedCluster`.

id¶

Type:	str

versions¶

Type:	list[`PublishedClusterVersion`]

Published Cluster Configuration¶

class tamr_unify_client.mastering.published_cluster.configuration.PublishedClustersConfiguration(client, data, alias=None)[source]¶

The configuration of published clusters in a project.

See https://docs.tamr.com/reference#the-published-clusters-configuration-object

relative_id¶

Type:	str

versions_time_to_live¶

Type:	str

resource_id¶

Type:	str

Published Cluster Version¶

class tamr_unify_client.mastering.published_cluster.version.PublishedClusterVersion(data)[source]¶

A version of a published cluster in a mastering project.

This is not a BaseResource because it does not have its own API endpoint.

Parameters:	data – The JSON entity representing this version.

version¶

Type:	str

timestamp¶

Type:	str

name¶

Type:	str

metrics¶

Type:	list[`Metric`]

record_ids¶

Type:	list[dict[str, str]]

Record Published Cluster¶

class tamr_unify_client.mastering.published_cluster.record.RecordPublishedCluster(data)[source]¶

A representation of a published cluster of a record in a mastering project with version information. See https://docs.tamr.com/reference#retrieve-published-clusters-given-record-ids.

This is not a BaseResource because it does not have its own API endpoint.

Parameters:	data – The JSON entity representing this `RecordPublishedCluster`.

entity_id¶

Type:	str

source_id¶

Type:	str

origin_entity_id¶

Type:	str

origin_source_id¶

Type:	str

versions¶

Type:	list[`RecordPublishedClusterVersion`]

Record Published Cluster Version¶

class tamr_unify_client.mastering.published_cluster.record_version.RecordPublishedClusterVersion(data)[source]¶

A version of a published cluster in a mastering project.

This is not a BaseResource because it does not have its own API endpoint.

Parameters:	data – The JSON entity representing this version.

version¶

Type:	str

timestamp¶

Type:	str

cluster_id¶

Type:	str

Operation¶

class tamr_unify_client.operation.Operation(client, data, alias=None)[source]¶

A long-running operation performed by Unify. Operations appear on the “Jobs” page of the Unify UI.

By design, client-side operations represent server-side operations at a particular point in time (namely, when the operation was fetched from the server). In other words: Operations will not pick up on server-side changes automatically. To get an up-to-date representation, refetch the operation e.g. op = op.poll().

apply_options(asynchronous=False, **options)[source]¶

Applies operation options to this operation.

NOTE: This function should not be called directly. Rather, options should be passed in through a higher-level function e.g. refresh() .

Synchronous mode:: Automatically waits for operation to resolve before returning the operation.
asynchronous mode:: Immediately return the 'PENDING' operation. It is up to the user to coordinate this operation with their code via wait() and/or poll() .

Parameters:	asynchronous (bool) – Whether or not to run in asynchronous mode. Default: `False`. **options – When running in synchronous mode, these options are passed to the underlying `wait()` call.
Returns:	Operation with options applied.
Return type:	`Operation`

type¶

Type:	str

description¶

Type:	str

state¶

Server-side state of this operation.

Operation state can be unresolved (i.e. state is one of: 'PENDING', 'RUNNING'), or resolved (i.e. state is one of: 'CANCELED', 'SUCCEEDED', 'FAILED'). Unless opting into asynchronous mode, all exposed operations should be resolved.

Note: you only need to manually pick up server-side changes when opting into asynchronous mode when kicking off this operation.

Usage:

>>> op.state # operation is currently 'PENDING'
'PENDING'
>>> op.wait() # continually polls until operation resolves
>>> op.state # incorrect usage; operation object state never changes.
'PENDING'
>>> op = op.poll() # correct usage; use value returned by Operation.poll or Operation.wait
>>> op.state
'SUCCEEDED'

poll()[source]¶

Poll this operation for server-side updates.

Does not update the calling Operation object. Instead, returns a new Operation.

Returns:	Updated representation of this operation.
Return type:	`Operation`

wait(poll_interval_seconds=3, timeout_seconds=None)[source]¶

Continuously polls for this operation’s server-side state.

Parameters:	poll_interval_seconds (int) – Time interval (in seconds) between subsequent polls. timeout_seconds (int) – Time (in seconds) to wait for operation to resolve.
Raises:	TimeoutError – If operation takes longer than timeout_seconds to resolve.
Returns:	Resolved operation.
Return type:	`Operation`

succeeded()[source]¶

Convenience method for checking if operation was successful.

Returns:	`True` if operation’s state is `'SUCCEEDED'`, `False` otherwise.
Return type:	`bool`

relative_id¶

Type:	str

resource_id¶

Type:	str

Project¶

Attribute Configuration¶

class tamr_unify_client.project.attribute_configuration.resource.AttributeConfiguration(client, data, alias=None)[source]¶

The configurations of Unify Attributes.

See https://docs.tamr.com/reference#the-attribute-configuration-object

relative_id¶

Type:	str

id¶

Type:	str

relative_attribute_id¶

Type:	str

attribute_role¶

Type:	str

similarity_function¶

Type:	str

enabled_for_ml¶

Type:	bool

tokenizer¶

Type:	str

numeric_field_resolution¶

Type:	list

attribute_name¶

Type:	str

resource_id¶

Type:	str

Attribute Configuration Collection¶

class tamr_unify_client.project.attribute_configuration.collection.AttributeConfigurationCollection(client, api_path)[source]¶

Collection of AttributeConfiguration

Parameters:	client (`Client`) – Client for API call delegation. api_path (str) – API path used to access this collection. E.g. `"projects/1/attributeConfigurations"`

by_resource_id(resource_id)[source]¶

Retrieve an attribute configuration by resource ID.

Parameters:	resource_id (str) – The resource ID.
Returns:	The specified attribute configuration.
Return type:	`AttributeConfiguration`

by_relative_id(relative_id)[source]¶

Retrieve an attribute configuration by relative ID.

Parameters:	relative_id (str) – The relative ID.
Returns:	The specified attribute configuration.
Return type:	`AttributeConfiguration`

by_external_id(external_id)[source]¶

Retrieve an attribute configuration by external ID.

Since attributes do not have external IDs, this method is not supported and will raise a NotImplementedError .

Parameters:	external_id (str) – The external ID.
Returns:	The specified attribute, if found.
Return type:	`AttributeConfiguration`
Raises:	KeyError – If no attribute with the specified external_id is found LookupError – If multiple attributes with the specified external_id are found NotImplementedError – AttributeConfiguration does not support external_id

stream()[source]¶

Stream attribute configurations in this collection. Implicitly called when iterating over this collection.

Returns:	Stream of attribute configurations.
Return type:	Python generator yielding `AttributeConfiguration`

Usage:

>>> for attributeConfiguration in collection.stream(): # explicit
>>>     do_stuff(attributeConfiguration)
>>> for attributeConfiguration in collection: # implicit
>>>     do_stuff(attributeConfiguration)

create(creation_spec)[source]¶

Create an Attribute configuration in this collection

Parameters:	creation_spec (dict[str, str]) – Attribute configuration creation specification should be formatted as specified in the Public Docs for adding an AttributeConfiguration.
Returns:	The created Attribute configuration
Return type:	`AttributeConfiguration`

Attribute Mapping¶

class tamr_unify_client.project.attribute_mapping.resource.AttributeMapping(data)[source]¶

see https://docs.tamr.com/reference#retrieve-projects-mappings AttributeMapping and AttributeMappingCollection do not inherit from BaseResource and BaseCollection. BC and BR require a specific URL for each individual attribute mapping (ex: /projects/1/attributeMappings/1), but these types of URLs do not exist for attribute mappings

id¶

Type:	str

relative_id¶

Type:	str

input_attribute_id¶

Type:	str

relative_input_attribute_id¶

Type:	str

input_dataset_name¶

Type:	str

input_attribute_name¶

Type:	str

unified_attribute_id¶

Type:	str

relative_unified_attribute_id¶

Type:	str

unified_dataset_name¶

Type:	str

unified_attribute_name¶

Type:	str

resource_id¶

Type:	str

Attribute Mapping Collection¶

class tamr_unify_client.project.attribute_mapping.collection.AttributeMappingCollection(client, api_path)[source]¶

Collection of AttributeMapping :param map_url: API path used to access this collection. :type api_path: str :param client: Client for API call delegation. :type client: Client

stream()[source]¶: Stream items in this collection. :returns: Stream of attribute mappings.

by_resource_id(resource_id)[source]¶: Retrieve an item in this collection by resource ID. :param resource_id: The resource ID. :type resource_id: str :returns: The specified attribute mapping. :rtype: AttributeMapping

by_relative_id(relative_id)[source]¶: Retrieve an item in this collection by relative ID. :param relative_id: The relative ID. :type relative_id: str :returns: The specified attribute mapping. :rtype: AttributeMapping

create(creation_spec)[source]¶: Create an Attribute mapping in this collection :param creation_spec: Attribute mapping creation specification should be formatted as specified in the Public Docs for adding an AttributeMapping. :type creation_spec: dict[str, str] :returns: The created Attribute mapping :rtype: AttributeMapping

Project¶

class tamr_unify_client.project.resource.Project(client, data, alias=None)[source]¶

A Unify project.

name¶

Type:	str

external_id¶

Type:	str

description¶

Type:	str

type¶

A Unify project type, listed in https://docs.tamr.com/reference#create-a-project.

Type:	str

attributes¶

Attributes of this project.

Returns:	Attributes of this project.
Return type:	`AttributeCollection`

unified_dataset()[source]¶

Unified dataset for this project.

Returns:	Unified dataset for this project.
Return type:	`Dataset`

as_categorization()[source]¶

Convert this project to a CategorizationProject

Returns:	This project.
Return type:	`CategorizationProject`
Raises:	TypeError – If the `type` of this project is not `"CATEGORIZATION"`

as_mastering()[source]¶

Convert this project to a MasteringProject

Returns:	This project.
Return type:	`MasteringProject`
Raises:	TypeError – If the `type` of this project is not `"DEDUP"`

add_input_dataset(dataset)[source]¶

Associate a dataset with a project in Unify.

By default, datasets are not associated with any projects. They need to be added as input to a project before they can be used as part of that project

Parameters:	dataset (`Dataset`) – The dataset to associate with the project.
Returns:	HTTP response from the server
Return type:	`requests.Response`

input_datasets()[source]¶

Retrieve a collection of this project’s input datasets.

Returns:	The project’s input datasets.
Return type:	`DatasetCollection`

attribute_configurations()[source]¶

Project’s attribute’s configurations.

Returns:	The configurations of the attributes of a project.
Return type:	`AttributeConfigurationCollection`

attribute_mappings()[source]¶

Project’s attribute’s mappings.

Returns:	The attribute mappings of a project.
Return type:	`AttributeMappingCollection`

relative_id¶

Type:	str

resource_id¶

Type:	str

Project Collection¶

class tamr_unify_client.project.collection.ProjectCollection(client, api_path='projects')[source]¶

Collection of Project s.

Parameters:	client (`Client`) – Client for API call delegation. api_path (str) – API path used to access this collection. Default: `"projects"`.

by_resource_id(resource_id)[source]¶

Retrieve a project by resource ID.

Parameters:	resource_id (str) – The resource ID. E.g. `"1"`
Returns:	The specified project.
Return type:	`Project`

by_relative_id(relative_id)[source]¶

Retrieve a project by relative ID.

Parameters:	relative_id (str) – The resource ID. E.g. `"projects/1"`
Returns:	The specified project.
Return type:	`Project`

by_external_id(external_id)[source]¶

Retrieve a project by external ID.

Parameters:	external_id (str) – The external ID.
Returns:	The specified project, if found.
Return type:	`Project`
Raises:	KeyError – If no project with the specified external_id is found LookupError – If multiple projects with the specified external_id are found

stream()[source]¶

Stream projects in this collection. Implicitly called when iterating over this collection.

Returns:	Stream of projects.
Return type:	Python generator yielding `Project`

Usage:

>>> for project in collection.stream(): # explicit
>>>     do_stuff(project)
>>> for project in collection: # implicit
>>>     do_stuff(project)

create(creation_spec)[source]¶

Create a Project in Unify

Parameters:	creation_spec (dict[str, str]) – Project creation specification should be formatted as specified in the Public Docs for Creating a Project.
Returns:	The created Project
Return type:	`Project`

Project Step¶

class tamr_unify_client.project.step.ProjectStep(client, data)[source]¶

A step of a Unify project. This is not a BaseResource because it has no API path and cannot be directly retrieved or modified.

See https://docs.tamr.com/reference#retrieve-downstream-dataset-usage

Parameters:	client (`Client`) – Delegate underlying API calls to this client. data (`dict`) – The JSON body containing project step information.

project_step_id¶

Type:	str

project_step_name¶

Type:	str

project_name¶

Type:	str

type¶

A Unify project type, listed in https://docs.tamr.com/reference#create-a-project.

Type:	str

project()[source]¶

Retrieves the Project this step is associated with.

Returns:	This step’s project.
Return type:	`Project`
Raises:	KeyError – If no project with the specified name is found. LookupError – If multiple projects with the specified name are found.