label-studio-sdk 0.0.30__tar.gz → 0.0.34__tar.gz

label-studio-sdk-0.0.34/MANIFEST.in

@@ -0,0 +1,2 @@
+include requirements.txt
+include label_studio_sdk/schema/*.json

{label-studio-sdk-0.0.30/label_studio_sdk.egg-info → label-studio-sdk-0.0.34}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: label-studio-sdk
-Version: 0.0.30
+Version: 0.0.34
 Summary: Label Studio annotation tool
 Home-page: https://github.com/heartexlabs/label-studio-sdk
 Author: Heartex

{label-studio-sdk-0.0.30 → label-studio-sdk-0.0.34}/README.md

@@ -18,7 +18,7 @@ This is the first release of the Label Studio SDK. It supports Label Studio Ente
 
 - **Find a bug?** [Create a GitHub issue](https://github.com/heartexlabs/label-studio-sdk/issues)!
 - **Have a question?** [Join the Slack Community](https://slack.labelstud.io/?source=github-sdk)!
-- **Want to contribute?** [See the contributing guide](https://github.com/heartexlabs/label-studio-sdk/CONTRIBUTING.md)
+- **Want to contribute?** [See the contributing guide](CONTRIBUTING.md)
 
 ## Quickstart
 To start using the SDK in your machine learning and data science projects and pipelines, do the following:
@@ -57,15 +57,4 @@ If you want to extend the SDK:
 
 ## Examples
 
-These end-to-end examples demonstrate how to use the SDK for specific use cases.
-
-### Active learning example
-
-If you want to write a Python script to set up an active learning workflow for labeling and training, review this [working active learning example as a Jupyter notebook](https://github.com/heartexlabs/label-studio-sdk/blob/master/examples/active_learning/active_learning.ipynb) or start with the [active learning python script example](https://github.com/heartexlabs/label-studio-sdk/blob/master/examples/active_learning/active_learning.py).
-
-### Weak supervision example
-
-If you want to write a Python script to perform programmatic labeling and use weak supervision to correct the noisy labels, refer to this [working weak supervision example as a Jupyter notebook](https://github.com/heartexlabs/label-studio-sdk/blob/master/examples/weak_supervision/weak_supervision.ipynb) or start with the [weak supervision python script example](https://github.com/heartexlabs/label-studio-sdk/blob/master/examples/weak_supervision/weak_supervision.py).
-
-<img src="https://labelstud.io/images/opossum/other/5.svg" width="400px">
-
+Please check [the examples folder](examples), there are many of very useful codes that you can learn from. 

{label-studio-sdk-0.0.30 → label-studio-sdk-0.0.34}/label_studio_sdk/__init__.py

@@ -1,8 +1,11 @@
 """ .. include::../docs/index.md
 """
+
 from .client import Client
 from .project import Project
 from .utils import parse_config
 
+__pdoc__ = {"label_interface": False}
+
 
-__version__ = '0.0.30'
+__version__ = '0.0.34'

{label-studio-sdk-0.0.30 → label-studio-sdk-0.0.34}/label_studio_sdk/client.py

@@ -1,23 +1,22 @@
 """ .. include::../docs/client.md
 """
-import os
 
 import json
-import warnings
 import logging
+import os
+
 import requests
 
 from typing import Optional
 from pydantic import BaseModel, constr, root_validator
 from requests.adapters import HTTPAdapter
-from types import SimpleNamespace
 
 logger = logging.getLogger(__name__)
 
 MAX_RETRIES = 3
-TIMEOUT = (10.0, 180.0)
+TIMEOUT = (10.0, int(os.environ.get('TIMEOUT', 180)))
 HEADERS = {}
-LABEL_STUDIO_DEFAULT_URL = 'http://localhost:8080'
+LABEL_STUDIO_DEFAULT_URL = "http://localhost:8080"
 
 
 class ClientCredentials(BaseModel):
@@ -28,11 +27,11 @@ class ClientCredentials(BaseModel):
     @root_validator(pre=True, allow_reuse=True)
     def either_key_or_email_password(cls, values):
         assert (
-            'email' in values or 'api_key' in values
-        ), 'At least one of email or api_key should be included'
+            "email" in values or "api_key" in values
+        ), "At least one of email or api_key should be included"
         assert (
-            'email' not in values or 'password' in values
-        ), 'Provide both email and password for login auth'
+            "email" not in values or "password" in values
+        ), "Provide both email and password for login auth"
         return values
 
 
@@ -74,17 +73,27 @@ class Client(object):
             If true, make_request will raise exceptions on request errors
         """
         if not url:
-            url = os.getenv('LABEL_STUDIO_URL', LABEL_STUDIO_DEFAULT_URL)
-        self.url = url.rstrip('/')
+            url = os.getenv("LABEL_STUDIO_URL", LABEL_STUDIO_DEFAULT_URL)
+        self.url = url.rstrip("/")
         self.make_request_raise = make_request_raise
         self.session = session or self.get_session()
 
+        # set cookies
+        self.cookies = cookies
+
         # set api key or get it using credentials (username and password)
         if api_key is None and credentials is None:
-            api_key = os.getenv('LABEL_STUDIO_API_KEY')
+            api_key = os.getenv("LABEL_STUDIO_API_KEY")
 
         if api_key is not None:
             credentials = ClientCredentials(api_key=api_key)
+
+        if api_key is None and credentials is None:
+            raise RuntimeError(
+                "If neither 'api_key' nor 'credentials' are provided, 'LABEL_STUDIO_API_KEY' environment variable must "
+                "be set"
+            )
+
         self.api_key = (
             credentials.api_key
             if credentials.api_key
@@ -92,18 +101,15 @@ class Client(object):
         )
 
         # set headers
-        self.headers = {'Authorization': f'Token {self.api_key}'}
+        self.headers = {"Authorization": f"Token {self.api_key}"}
         if oidc_token:
-            self.headers.update({'Proxy-Authorization': f'Bearer {oidc_token}'})
+            self.headers.update({"Proxy-Authorization": f"Bearer {oidc_token}"})
         if extra_headers:
             self.headers.update(extra_headers)
 
-        # set cookies
-        self.cookies = cookies
-
         # set versions from /version endpoint
         self.versions = versions if versions else self.get_versions()
-        self.is_enterprise = 'label-studio-enterprise-backend' in self.versions
+        self.is_enterprise = "label-studio-enterprise-backend" in self.versions
 
     def get_versions(self):
         """Call /version api and get all Label Studio component versions
@@ -113,14 +119,14 @@ class Client(object):
         dict with Label Studio component names and their versions
 
         """
-        self.versions = self.make_request('GET', '/api/version').json()
+        self.versions = self.make_request("GET", "/api/version").json()
         return self.versions
 
     def get_api_key(self, credentials: ClientCredentials):
         login_url = self.get_url("/user/login")
         # Retrieve and set the CSRF token first
         self.session.get(login_url)
-        csrf_token = self.session.cookies.get('csrftoken', None)
+        csrf_token = self.session.cookies.get("csrftoken", None)
         login_data = dict(**credentials.dict(), csrfmiddlewaretoken=csrf_token)
         self.session.post(
             login_url,
@@ -143,7 +149,7 @@ class Client(object):
         dict
             Status string like "UP"
         """
-        response = self.make_request('GET', '/health')
+        response = self.make_request("GET", "/health")
         return response.json()
 
     def get_projects(self, **query_params):
@@ -164,7 +170,7 @@ class Client(object):
         dict
             Status string
         """
-        response = self.make_request('DELETE', f'/api/projects/{project_id}/')
+        response = self.make_request("DELETE", f"/api/projects/{project_id}/")
         return response
 
     def delete_all_projects(self):
@@ -176,7 +182,7 @@ class Client(object):
             List of (dict) status strings
         """
         responses = []
-        project_ids = [project.get_params()['id'] for project in self.list_projects()]
+        project_ids = [project.get_params()["id"] for project in self.list_projects()]
         for project_id in project_ids:
             response = self.delete_project(project_id)
             responses.append(response)
@@ -192,14 +198,14 @@ class Client(object):
         """
         from .project import Project
 
-        params = {'page_size': 10000000}
+        params = {"page_size": 10000000}
         params.update(query_params)
-        response = self.make_request('GET', '/api/projects', params=params)
+        response = self.make_request("GET", "/api/projects", params=params)
         if response.status_code == 200:
             projects = []
-            for data in response.json()['results']:
+            for data in response.json()["results"]:
                 project = Project._create_from_id(
-                    client=self, project_id=data['id'], params=data
+                    client=self, project_id=data["id"], params=data
                 )
                 projects.append(project)
                 logger.debug(
@@ -264,10 +270,10 @@ class Client(object):
         """
         from .users import User
 
-        response = self.make_request('GET', '/api/users')
+        response = self.make_request("GET", "/api/users")
         users = []
         for user_data in response.json():
-            user_data['client'] = self
+            user_data["client"] = self
             users.append(User(**user_data))
         return users
 
@@ -292,28 +298,28 @@ class Client(object):
 
         payload = (
             {
-                'username': user.username if user.username else user.email,
-                'email': user.email,
-                'first_name': user.first_name,
-                'last_name': user.last_name,
-                'phone': user.phone,
+                "username": user.username if user.username else user.email,
+                "email": user.email,
+                "first_name": user.first_name,
+                "last_name": user.last_name,
+                "phone": user.phone,
             }
             if isinstance(user, User)
             else user
         )
 
         response = self.make_request(
-            'POST', '/api/users', json=payload, raise_exceptions=False
+            "POST", "/api/users", json=payload, raise_exceptions=False
         )
         user_data = response.json()
-        user_data['client'] = self
+        user_data["client"] = self
 
         if response.status_code < 400:
             return User(**user_data)
         else:
-            if 'already exists' in response.text and exist_ok is True:
+            if "already exists" in response.text and exist_ok is True:
                 return None
-            logger.error('Create user error: ' + str(response.json()))
+            logger.error("Create user error: " + str(response.json()))
             return None
 
     def get_workspaces(self):
@@ -333,13 +339,48 @@ class Client(object):
             self.is_enterprise
         ), "Workspaces are available only for Enterprise instance of Label Studio"
 
-        response = self.make_request('GET', '/api/workspaces')
+        response = self.make_request("GET", "/api/workspaces")
         workspaces = []
         for workspace_data in response.json():
-            workspace_data['client'] = self
+            workspace_data["client"] = self
             workspaces.append(Workspace(**workspace_data))
         return workspaces
 
+    # write function get_workspace_by_title
+    def get_workspace_by_title(self, title):
+        """Return workspace by title from the current organization account
+
+        Parameters
+        ----------
+        title: str
+            Workspace title
+
+        Returns
+        -------
+        `label_studio_sdk.workspaces.Workspace` or None
+
+        """
+        workspaces = self.get_workspaces()
+        for workspace in workspaces:
+            if workspace.title == title:
+                return workspace
+        return None
+
+    def get_organization(self):
+        """Return active organization for the current user
+
+        Returns
+        -------
+        dict
+        """
+        # get organization id from the current user api
+        response = self.make_request('GET', '/api/current-user/whoami').json()
+        organization_id = response['active_organization']
+
+        # get organization data by id
+        response = self.make_request("GET", f"/api/organizations/{organization_id}")
+        return response.json()
+
     def get_session(self):
         """Create a session with requests.Session()
 
@@ -350,8 +391,8 @@ class Client(object):
         """
         session = requests.Session()
         session.headers.update(HEADERS)
-        session.mount('http://', HTTPAdapter(max_retries=MAX_RETRIES))
-        session.mount('https://', HTTPAdapter(max_retries=MAX_RETRIES))
+        session.mount("http://", HTTPAdapter(max_retries=MAX_RETRIES))
+        session.mount("https://", HTTPAdapter(max_retries=MAX_RETRIES))
         return session
 
     def get_url(self, suffix):
@@ -384,14 +425,14 @@ class Client(object):
         Response object for the relevant endpoint.
 
         """
-        if 'timeout' not in kwargs:
-            kwargs['timeout'] = TIMEOUT
+        if "timeout" not in kwargs:
+            kwargs["timeout"] = TIMEOUT
 
         raise_exceptions = self.make_request_raise
-        if 'raise_exceptions' in kwargs:  # kwargs have higher priority
-            raise_exceptions = kwargs.pop('raise_exceptions')
+        if "raise_exceptions" in kwargs:  # kwargs have higher priority
+            raise_exceptions = kwargs.pop("raise_exceptions")
 
-        logger.debug(f'{method}: {url} with args={args}, kwargs={kwargs}')
+        logger.debug(f"{method}: {url} with args={args}, kwargs={kwargs}")
         response = self.session.request(
             method,
             self.get_url(url),
@@ -403,49 +444,27 @@ class Client(object):
 
         if raise_exceptions:
             if response.status_code >= 400:
-                try:
-                    content = json.dumps(json.loads(response.content), indent=2)
-                except:
-                    content = response.text
-
-                logger.error(
-                    f'\n--------------------------------------------\n'
-                    f'Request URL: {response.url}\n'
-                    f'Response status code: {response.status_code}\n'
-                    f'Response content:\n{content}\n\n'
-                    f'SDK error traceback:'
-                )
+                self.log_response_error(response)
                 response.raise_for_status()
 
         return response
 
-    def sync_storage(self, storage_type, storage_id):
-        """Synchronize Cloud Storage.
-
-        Parameters
-        ----------
-        storage_type: string
-            Specify the type of the storage container.
-        storage_id: int
-            Specify the storage ID of the storage container.
-
-        Returns
-        -------
-        dict:
-            containing the same fields as in the original storage request and:
-
-        id: int
-            Storage ID
-        type: str
-            Type of storage
-        created_at: str
-            Creation time
-        last_sync: str
-            Time last sync finished, can be empty.
-        last_sync_count: int
-            Number of tasks synced in the last sync
-        """
+    def log_response_error(self, response):
+        try:
+            content = json.dumps(json.loads(response.content), indent=2)
+        except:
+            content = response.text
+
+        logger.error(
+            f"\n--------------------------------------------\n"
+            f"Request URL: {response.url}\n"
+            f"Response status code: {response.status_code}\n"
+            f"Response content:\n{content}\n\n"
+            f"SDK error traceback:"
+        )
 
+    def sync_storage(self, storage_type, storage_id):
+        """See project.sync_storage for more info"""
         response = self.make_request(
             "POST", f"/api/storages/{storage_type}/{str(storage_id)}/sync"
         )

{label-studio-sdk-0.0.30 → label-studio-sdk-0.0.34}/label_studio_sdk/data_manager.py

@@ -31,9 +31,10 @@
     tasks = project.get_tasks(filters=filters)
     ```
 """
+
 from datetime import datetime
 
-DATETIME_FORMAT = '%Y-%m-%dT%H:%M:%S.%fZ'
+DATETIME_FORMAT = "%Y-%m-%dT%H:%M:%S.%fZ"
 
 
 class Filters:
@@ -41,9 +42,9 @@ class Filters:
     Use the methods and variables in this class to create and combine filters for tasks on the Label Studio Data Manager.
     """
 
-    OR = 'or'
+    OR = "or"
     """Combine filters with an OR"""
-    AND = 'and'
+    AND = "and"
     """Combine filters with an AND"""
 
     @staticmethod
@@ -85,7 +86,7 @@ class Filters:
         dict
         """
         return {
-            "filter": 'filter:' + name,
+            "filter": "filter:" + name,
             "operator": operator,
             "type": column_type,
             "value": value,
@@ -106,7 +107,7 @@ class Filters:
             datetime in `'%Y-%m-%dT%H:%M:%S.%fZ'` format
 
         """
-        assert isinstance(dt, datetime), 'dt must be datetime type'
+        assert isinstance(dt, datetime), "dt must be datetime type"
         return dt.strftime(DATETIME_FORMAT)
 
     @classmethod
@@ -133,7 +134,7 @@ class Filters:
         if maximum is not None:
             if isinstance(maximum, datetime):
                 maximum = cls.datetime(maximum)
-            return {'min': value, 'max': maximum}
+            return {"min": value, "max": maximum}
 
         return value
 
@@ -160,13 +161,13 @@ class Operator:
 class Type:
     """Specify the type of data in a column."""
 
-    Number = 'Number'
-    Datetime = 'Datetime'
-    Boolean = 'Boolean'
-    String = 'String'
+    Number = "Number"
+    Datetime = "Datetime"
+    Boolean = "Boolean"
+    String = "String"
     List = "List"
 
-    Unknown = 'Unknown'
+    Unknown = "Unknown"
     """ Unknown is explicitly converted to string format. """
 
 
@@ -175,6 +176,8 @@ class Column:
 
     id = "tasks:id"
     """Task ID"""
+    inner_id = "tasks:inner_id"
+    """Task Inner ID, it starts from 1 for all projects"""
     ground_truth = "tasks:ground_truth"
     """Ground truth status of the tasks"""
     annotations_results = "tasks:annotations_results"
@@ -191,6 +194,8 @@ class Column:
     """Name of the file uploaded to create the tasks"""
     created_at = "tasks:created_at"
     """Time the task was created at"""
+    updated_at = "tasks:updated_at"
+    """Time the task was updated at (e.g. new annotation was created, review added, etc)"""
     annotators = "tasks:annotators"
     """Annotators that completed the task (Community). Can include assigned annotators (Enterprise only)"""
     total_predictions = "tasks:total_predictions"
@@ -209,6 +214,10 @@ class Column:
     """Number of annotations rejected for a task in review (Enterprise only)"""
     reviews_accepted = "tasks:reviews_accepted"
     """Number of annotations accepted for a task in review (Enterprise only)"""
+    comments = "tasks:comments"
+    """Number of comments in a task"""
+    unresolved_comment_count = "tasks:unresolved_comment_count"
+    """Number of unresolved comments in a task"""
 
     @staticmethod
     def data(task_field):
@@ -246,21 +255,21 @@ def _test():
     )
 
     assert filters == {
-        'conjunction': 'or',
-        'items': [
+        "conjunction": "or",
+        "items": [
             {
-                'filter': 'filter:tasks:id',
-                'operator': 'greater',
-                'type': 'Number',
-                'value': 42,
+                "filter": "filter:tasks:id",
+                "operator": "greater",
+                "type": "Number",
+                "value": 42,
             },
             {
-                'filter': 'filter:tasks:completed_at',
-                'operator': 'in',
-                'type': 'Datetime',
-                'value': {
-                    'min': '2021-11-01T00:00:00.000000Z',
-                    'max': '2021-11-05T00:00:00.000000Z',
+                "filter": "filter:tasks:completed_at",
+                "operator": "in",
+                "type": "Datetime",
+                "value": {
+                    "min": "2021-11-01T00:00:00.000000Z",
+                    "max": "2021-11-05T00:00:00.000000Z",
                 },
             },
         ],

label-studio-sdk-0.0.34/label_studio_sdk/exceptions.py

@@ -0,0 +1,10 @@
+class LSConfigParseException(Exception):
+    """ """
+
+
+class LabelStudioXMLSyntaxErrorSentryIgnored(Exception):
+    """ """
+
+
+class LabelStudioValidationErrorSentryIgnored(Exception):
+    """ """

label-studio-sdk-0.0.34/label_studio_sdk/label_interface/__init__.py

@@ -0,0 +1 @@
+from .interface import LabelInterface

label-studio-sdk-0.0.34/label_studio_sdk/label_interface/base.py

@@ -0,0 +1,77 @@
+from pydantic import BaseModel
+from typing import Dict, Optional, List, Tuple, Any, Callable
+
+
+class LabelStudioTag(BaseModel):
+    """
+    Base class for a LabelStudio Tag
+
+    Attributes:
+    -----------
+    attr: Optional[Dict]
+        A dictionary of attributes for the tag
+    tag: Optional[str]
+        The tag name
+    """
+
+    attr: Optional[Dict]
+    tag: Optional[str]
+
+    def match(
+        self,
+        tag_type: str,
+        name: Optional[str] = None,
+        name_filter_fn: Optional[Callable] = None,
+        to_name: Optional[str] = None,
+        to_name_filter_fn: Optional[Callable] = None,
+    ) -> bool:
+        """
+        This method checks if the current instance of LabelStudioTag matches the provided parameters.
+
+        Parameters:
+        -----------
+        tag_type : str
+            The type of the tag to match. It can be a string or a tuple of strings.
+        name : Optional[str]
+            The name of the tag to match. If provided, it should match the name of the current instance.
+        name_filter_fn : Optional[Callable]
+            A function to filter the name of the tag. If provided, the function should return True for a match.
+        to_name : Optional[str]
+            The 'to_name' attribute of the tag to match. If provided, it should match the 'to_name' of the current instance.
+        to_name_filter_fn : Optional[Callable]
+            A function to filter the 'to_name' of the tag. If provided, the function should return True for a match.
+
+        Returns:
+        --------
+        bool
+            True if the current instance matches the provided parameters, False otherwise.
+        """
+        if isinstance(tag_type, str):
+            tag_type = tag_type.lower()
+        elif isinstance(tag_type, tuple):
+            tag_type = tuple(t.lower() for t in tag_type)
+
+        if name:
+            name = name.lower()
+
+        if to_name:
+            to_name = to_name.lower()
+
+        if (isinstance(tag_type, str) and self.tag.lower() != tag_type) or (
+            isinstance(tag_type, tuple) and self.tag.lower() not in tag_type
+        ):
+            return False
+
+        if name and self.name.lower() != name:
+            return False
+
+        if name_filter_fn and not name_filter_fn(self.name.lower()):
+            return False
+
+        if to_name and self.to_name.lower() != to_name:
+            return False
+
+        if to_name_filter_fn and not to_name_filter_fn(self.name.lower()):
+            return False
+
+        return True