label-studio-sdk 1.0.13__py3-none-any.whl → 1.0.15__py3-none-any.whl

label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py

@@ -3,6 +3,7 @@ import io
 import logging
 import os
 import shutil
+import base64
 from contextlib import contextmanager
 from tempfile import mkdtemp
 from urllib.parse import urlparse
@@ -201,7 +202,7 @@ def download_and_cache(
 
     # local storage: /data/local-files?d=dir/1.jpg => 1.jpg
     if is_local_storage_file:
-        url_filename = os.path.basename(url.split('?d=')[1])
+        url_filename = os.path.basename(url.split("?d=")[1])
     # cloud storage: s3://bucket/1.jpg => 1.jpg
     elif is_cloud_storage_file:
         url_filename = os.path.basename(url)
@@ -213,7 +214,11 @@ def download_and_cache(
     filepath = os.path.join(cache_dir, url_hash + "__" + url_filename)
 
     if not os.path.exists(filepath):
-        logger.info("Download {url} to {filepath}. download_resources: {download_resources}".format(url=url, filepath=filepath, download_resources=download_resources))
+        logger.info(
+            "Download {url} to {filepath}. download_resources: {download_resources}".format(
+                url=url, filepath=filepath, download_resources=download_resources
+            )
+        )
         if download_resources:
             headers = {
                 # avoid requests.exceptions.HTTPError: 403 Client Error: Forbidden. Please comply with the User-Agent policy:
@@ -256,3 +261,123 @@ def get_all_files_from_dir(d):
         if os.path.isfile(filepath):
             out.append(filepath)
     return out
+
+
+def get_base64_content(
+    url,
+    hostname=None,
+    access_token=None,
+    task_id=None,
+):
+    """This helper function is used to download a file and return its base64 representation without saving to filesystem.
+
+    :param url: File URL to download, it can be a uploaded file, local storage, cloud storage file or just http(s) url
+    :param hostname: Label Studio Hostname, it will be used for uploaded files, local storage files and cloud storage files
+      if not provided, it will be taken from LABEL_STUDIO_URL env variable
+    :param access_token: Label Studio access token, it will be used for uploaded files, local storage files and cloud storage files
+      if not provided, it will be taken from LABEL_STUDIO_API_KEY env variable
+    :param task_id: Label Studio Task ID, required for cloud storage files
+      because the URL will be rebuilt to `{hostname}/tasks/{task_id}/presign/?fileuri={url}`
+
+    :return: base64 encoded file content
+    """
+    # get environment variables
+    hostname = (
+        hostname
+        or os.getenv("LABEL_STUDIO_URL", "")
+        or os.getenv("LABEL_STUDIO_HOST", "")
+    )
+    access_token = (
+        access_token
+        or os.getenv("LABEL_STUDIO_API_KEY", "")
+        or os.getenv("LABEL_STUDIO_ACCESS_TOKEN", "")
+    )
+    if "localhost" in hostname:
+        logger.warning(
+            f"Using `localhost` ({hostname}) in LABEL_STUDIO_URL, "
+            f"`localhost` is not accessible inside of docker containers. "
+            f"You can check your IP with utilities like `ifconfig` and set it as LABEL_STUDIO_URL."
+        )
+    if hostname and not (
+        hostname.startswith("http://") or hostname.startswith("https://")
+    ):
+        raise ValueError(
+            f"Invalid hostname in LABEL_STUDIO_URL: {hostname}. "
+            "Please provide full URL starting with protocol (http:// or https://)."
+        )
+
+    # fix file upload url
+    if url.startswith("upload") or url.startswith("/upload"):
+        url = "/data" + ("" if url.startswith("/") else "/") + url
+
+    is_uploaded_file = url.startswith("/data/upload")
+    is_local_storage_file = url.startswith("/data/") and "?d=" in url
+    is_cloud_storage_file = (
+        url.startswith("s3:") or url.startswith("gs:") or url.startswith("azure-blob:")
+    )
+
+    # Local storage file: try to load locally
+    if is_local_storage_file:
+        filepath = url.split("?d=")[1]
+        filepath = safe_build_path(LOCAL_FILES_DOCUMENT_ROOT, filepath)
+        if os.path.exists(filepath):
+            logger.debug(
+                f"Local Storage file path exists locally, read content directly: {filepath}"
+            )
+            with open(filepath, "rb") as f:
+                return base64.b64encode(f.read()).decode("utf-8")
+
+    # Upload or Local Storage file
+    if is_uploaded_file or is_local_storage_file or is_cloud_storage_file:
+        # hostname check
+        if not hostname:
+            raise FileNotFoundError(
+                f"Can't resolve url, hostname not provided: {url}. "
+                "You can set LABEL_STUDIO_URL environment variable to use it as a hostname."
+            )
+        # uploaded and local storage file
+        elif is_uploaded_file or is_local_storage_file:
+            url = concat_urls(hostname, url)
+            logger.info("Resolving url using hostname [" + hostname + "]: " + url)
+        # s3, gs, azure-blob file
+        elif is_cloud_storage_file:
+            if task_id is None:
+                raise Exception(
+                    "Label Studio Task ID is required for cloud storage files"
+                )
+            url = concat_urls(hostname, f"/tasks/{task_id}/presign/?fileuri={url}")
+            logger.info(
+                "Cloud storage file: Resolving url using hostname ["
+                + hostname
+                + "]: "
+                + url
+            )
+
+        # check access token
+        if not access_token:
+            raise FileNotFoundError(
+                "To access uploaded and local storage files you have to "
+                "set LABEL_STUDIO_API_KEY environment variable."
+            )
+
+    # Download the content but don't save to filesystem
+    headers = {
+        "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36"
+    }
+
+    # check if url matches hostname - then uses access token to this Label Studio instance
+    parsed_url = urlparse(url)
+    if access_token and hostname and parsed_url.netloc == urlparse(hostname).netloc:
+        headers["Authorization"] = "Token " + access_token
+        logger.debug("Authorization token is used for get_base64_content")
+
+    try:
+        r = requests.get(url, headers=headers, verify=VERIFY_SSL)
+        r.raise_for_status()
+        return base64.b64encode(r.content).decode("utf-8")
+    except requests.exceptions.SSLError as e:
+        logger.error(
+            f"SSL error during requests.get('{url}'): {e}\n"
+            f"Try to set VERIFY_SSL=False in environment variables to bypass SSL verification."
+        )
+        raise e

label_studio_sdk/core/client_wrapper.py

@@ -18,7 +18,7 @@ class BaseClientWrapper:
 
         # even in the async case, refreshing access token (when the existing one is expired) should be sync
         from ..tokens.client_ext import TokensClientExt
-        self._tokens_client = TokensClientExt(base_url=base_url, api_key=api_key)
+        self._tokens_client = TokensClientExt(base_url=base_url, api_key=api_key, client_wrapper=self)
 
 
     def get_timeout(self) -> typing.Optional[float]:

label_studio_sdk/label_interface/interface.py

@@ -8,9 +8,11 @@ import re
 import json
 import jsonschema
 
+from functools import cached_property
 from typing import Dict, Optional, List, Tuple, Any, Callable, Union
 from pydantic import BaseModel
 
+
 # from typing import Dict, Optional, List, Tuple, Any
 from collections import defaultdict, OrderedDict
 from lxml import etree
@@ -517,6 +519,19 @@ class LabelInterface:
 
         return lst
     
+    @cached_property
+    def ner_tags(self):
+        return self.find_tags('controls', lambda t: t.tag.lower() in ('labels', 'hypertextlabels'))
+    
+    @cached_property
+    def image_tags(self):
+        return self.find_tags('objects', lambda t: t.tag.lower() == 'image')
+    
+    @cached_property
+    def pdf_tags(self):
+        return self.find_tags('objects', lambda t: t.tag.lower() == 'pdf')
+
+
     def load_task(self, task):
         """Loads a task and substitutes the value in each object tag
         with actual data from the task, returning a copy of the

label_studio_sdk/label_interface/object_tags.py

@@ -189,12 +189,17 @@ class AudioTag(ObjectTag):
 
 
 class ImageTag(ObjectTag):
-    """ """
+    """Image tag"""
     tag: str = "Image"
     
     def _generate_example(self, examples, only_urls=False):
         """ """
         return examples.get("Image")
+    
+    @property
+    def is_image_list(self):
+        """Check if the tag is an image list, i.e. it has a valueList attribute that accepts list of images"""
+        return bool(self.attr.get("valueList")) if self.attr else False
 
 
 class TableTag(ObjectTag):

label_studio_sdk/tokens/client_ext.py

@@ -1,6 +1,7 @@
 import threading
 import typing
 from datetime import datetime, timezone
+import ssl
 
 import httpx
 import jwt
@@ -12,9 +13,10 @@ from ..types.access_token_response import AccessTokenResponse
 class TokensClientExt:
     """Client for managing authentication tokens."""
 
-    def __init__(self, base_url: str, api_key: str):
+    def __init__(self, base_url: str, api_key: str, client_wrapper=None):
         self._base_url = base_url
         self._api_key = api_key
+        self._client_wrapper = client_wrapper
         self._use_legacy_token = not self._is_valid_jwt_token(api_key, raise_if_expired=True)
 
         # cache state for access token when using jwt-based api_key
@@ -76,19 +78,63 @@ class TokensClientExt:
         
         return self._access_token
 
+    def _get_client_params(self, existing_client: httpx.AsyncClient) -> dict:
+        """Extract parameters from an existing client to create a new one.
+
+        Args:
+            existing_client: The existing client to extract parameters from.
+
+        Returns:
+            dict: Parameters for creating a new client.
+        """
+        return {
+            'auth': existing_client.auth,
+            'params': existing_client.params,
+            'headers': existing_client.headers,
+            'cookies': existing_client.cookies,
+            'timeout': existing_client.timeout,
+            'follow_redirects': existing_client.follow_redirects,
+            'max_redirects': existing_client.max_redirects,
+            'event_hooks': existing_client.event_hooks,
+            'base_url': existing_client.base_url,
+            'trust_env': existing_client.trust_env,
+            'default_encoding': existing_client._default_encoding,
+            'verify': existing_client._transport._pool._ssl_context.verify_mode != ssl.CERT_NONE,
+            'http1': existing_client._transport._pool._http1,
+            'http2': existing_client._transport._pool._http2,
+            'limits': httpx.Limits(
+                max_connections=existing_client._transport._pool._max_connections,
+                max_keepalive_connections=existing_client._transport._pool._max_keepalive_connections,
+                keepalive_expiry=existing_client._transport._pool._keepalive_expiry
+            )
+        }
+
     def refresh(self) -> AccessTokenResponse:
         """Refresh the access token and return the token response."""
-        # We don't do this often, just use a separate httpx client for simplicity here
-        # (avoids complicated state management and sync vs async handling)
-        with httpx.Client() as sync_client:
-            response = sync_client.request(
+        existing_client = self._client_wrapper.httpx_client.httpx_client
+
+        # For sync client, use it directly
+        if isinstance(existing_client, httpx.Client):
+            print(f"\nverify:{existing_client._transport._pool._ssl_context.verify_mode != ssl.CERT_NONE}")
+            response = existing_client.request(
                 method="POST",
                 url=f"{self._base_url}/api/token/refresh/",
                 json={"refresh": self._api_key},
                 headers={"Content-Type": "application/json"},
             )
+        else:
+            # If an async client was used, get all parameters from the client to init a new sync client
+            client_params = self._get_client_params(existing_client)
+
+            with httpx.Client(**client_params) as sync_client:
+                response = sync_client.request(
+                    method="POST",
+                    url=f"{self._base_url}/api/token/refresh/",
+                    json={"refresh": self._api_key},
+                    headers={"Content-Type": "application/json"},
+                )
             
-            if response.status_code == 200:
-                return AccessTokenResponse.parse_obj(response.json())
-            else:
-                raise ApiError(status_code=response.status_code, body=response.json())
+        if response.status_code == 200:
+            return AccessTokenResponse.parse_obj(response.json())
+        else:
+            raise ApiError(status_code=response.status_code, body=response.json())

{label_studio_sdk-1.0.13.dist-info → label_studio_sdk-1.0.15.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: label-studio-sdk
-Version: 1.0.13
+Version: 1.0.15
 Summary: 
 Requires-Python: >=3.9,<4
 Classifier: Intended Audience :: Developers

{label_studio_sdk-1.0.13.dist-info → label_studio_sdk-1.0.15.dist-info}/RECORD

@@ -6,7 +6,7 @@ label_studio_sdk/_extensions/label_studio_tools/core/__init__.py,sha256=47DEQpj8
 label_studio_sdk/_extensions/label_studio_tools/core/label_config.py,sha256=P1S7dPjFkqF2zIQzk11iljhharrUc9qQRM_rUN38iJQ,6406
 label_studio_sdk/_extensions/label_studio_tools/core/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 label_studio_sdk/_extensions/label_studio_tools/core/utils/exceptions.py,sha256=JxaXUMghUp1YvL--s8KFC4mCHVbV39giE3kSBHCmuFU,66
-label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py,sha256=FO0fBVvffuDjbQQcvqLsXmGUn1gCP1YmA-tmNsvX8oo,9650
+label_studio_sdk/_extensions/label_studio_tools/core/utils/io.py,sha256=NJHLJA8Q93_MLmo_Yx9F4_Z-kkWCZmF_Ahi3MtDQCPo,14800
 label_studio_sdk/_extensions/label_studio_tools/core/utils/json_schema.py,sha256=_Lg3DxhRqGhzlk3egGUDufx-iaoEWec19upZKp-Cwic,3378
 label_studio_sdk/_extensions/label_studio_tools/core/utils/params.py,sha256=ZSUb-IXG5OcPQ7pJ8NDRLon-cMxnjVq6XtinxvTuJso,1244
 label_studio_sdk/_extensions/label_studio_tools/etl/__init__.py,sha256=SdN7JGLJ1araqbx-nL2fVdhm6E6CNyru-vWVs6sMswI,31
@@ -68,7 +68,7 @@ label_studio_sdk/converter/main.py,sha256=gfe5zPV2dnIk4ifG1AT95ExkzOSLzje0EOjnW0
 label_studio_sdk/converter/utils.py,sha256=VshPBwZLu2VPIGVsShKAkZwB_zKz0VvMkNRCwWeEqWg,18702
 label_studio_sdk/core/__init__.py,sha256=-t9txgeQZL_1FDw_08GEoj4ft1Cn9Dti6X0Drsadlr0,1519
 label_studio_sdk/core/api_error.py,sha256=RE8LELok2QCjABadECTvtDp7qejA1VmINCh6TbqPwSE,426
-label_studio_sdk/core/client_wrapper.py,sha256=fThAUV72_apQ4I2bS60iM84082lgXo4YMM6yvl07Kpg,2209
+label_studio_sdk/core/client_wrapper.py,sha256=lAGxJnIC7HMfysZS9sLHX1lnc82jw_n3eo_pTeljlFc,2230
 label_studio_sdk/core/datetime_utils.py,sha256=nBys2IsYrhPdszxGKCNRPSOCwa-5DWOHG95FB8G9PKo,1047
 label_studio_sdk/core/file.py,sha256=d4NNbX8XvXP32z8KpK2Xovv33nFfruIrpz0QWxlgpZk,2663
 label_studio_sdk/core/http_client.py,sha256=siUQ6UV0ARZALlxubqWSSAAPC9B4VW8y6MGlHStfaeo,19552
@@ -157,9 +157,9 @@ label_studio_sdk/label_interface/base.py,sha256=NCgY7ntk0WSc9O9iXu3g37-CxbZgCx_W
 label_studio_sdk/label_interface/control_tags.py,sha256=qLe4gsRxvppuNtrxfmgZHFX1ahM-XhePlrchZfnJiL0,30141
 label_studio_sdk/label_interface/create.py,sha256=c3h5_FF4u5J62_mqq1oK2mjqXL-I1559C6MfoxkgO6s,6993
 label_studio_sdk/label_interface/data_examples.json,sha256=uCYvCtMIxPi1-jLlFhwJPh01tLyMIRwTjINeAeW-JzE,8195
-label_studio_sdk/label_interface/interface.py,sha256=U_2IkpQXxw8fBW_buwfWIOvgdk1z_xLHFdk8YPUC1eE,45519
+label_studio_sdk/label_interface/interface.py,sha256=nEC_RQJ9VCCtYRKqJ7AYhggbLqvTKvJkPIRTQeIm8vc,45959
 label_studio_sdk/label_interface/label_tags.py,sha256=nWEo21Gd8IPzIO72UqraLrChIbvrSMCA_eEhzYGnGCc,2282
-label_studio_sdk/label_interface/object_tags.py,sha256=EGe3bYTZr92SezzWka8grYnvOQNtyfEYa5-yoM4a7Es,8705
+label_studio_sdk/label_interface/object_tags.py,sha256=9k3DEYEh7aXSLh2JjH-SWNVupP1qgwHFte85Ix7-4dQ,8944
 label_studio_sdk/label_interface/objects.py,sha256=V1Spp0S9qE7iA-5kPCi0QyHrJ80Du9BUuYMsQUAQqc0,1535
 label_studio_sdk/label_interface/region.py,sha256=th39WeQk8ypi-4krEpsW0BZnoygu4XgvP4w7NkRQp2M,1755
 label_studio_sdk/ml/__init__.py,sha256=J4ncAcAOU_qriOx_Im9eFmXyupKM19SXMcpMcXSmw-I,455
@@ -215,7 +215,7 @@ label_studio_sdk/tasks/types/tasks_list_request_fields.py,sha256=5YXxQgyzoaL0QjS
 label_studio_sdk/tasks/types/tasks_list_response.py,sha256=j1pNluAWQOQ8-d9YXQyRQAefnrl8uLQEB7_L55Z8DME,1136
 label_studio_sdk/tokens/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
 label_studio_sdk/tokens/client.py,sha256=SvBcKXIsrTihMJC72Ifxv0U1N3gtLGz3JxqdXYA_hD4,19101
-label_studio_sdk/tokens/client_ext.py,sha256=chhzBuVYp0YeUrAnYVwDX5yJq5IJlomGhBQ1zvjZAkI,3976
+label_studio_sdk/tokens/client_ext.py,sha256=Rma4ROcJPQOcBBGsA8Iteqk7WrbW7DTU381Ay5pJMiY,6134
 label_studio_sdk/types/__init__.py,sha256=fQykjzHpX04ftslk5I_hWSJQ_H9Kd8XJAmSv18EVOIc,8905
 label_studio_sdk/types/access_token_response.py,sha256=RV9FqkIiFR_9kmKueB-KiqjVyneiqUkMVueAlk5fUyc,624
 label_studio_sdk/types/annotation.py,sha256=AnHm2VjMasWZsaNXVSUzLYbpYrmM4NPZgWQh7WGa6ZQ,3157
@@ -364,7 +364,7 @@ label_studio_sdk/workspaces/members/client.py,sha256=IVM52Yq_9zMQ3TUHT0AkZ5BTQ9a
 label_studio_sdk/workspaces/members/types/__init__.py,sha256=ZIa_rd7d6K9ZITjTU6fptyGgvjNDySksJ7Rbn4wyhD4,252
 label_studio_sdk/workspaces/members/types/members_create_response.py,sha256=7Hp5FSWm4xR5ZOEmEIglq5HYtM9KWZZBDp87jw7jYFg,668
 label_studio_sdk/workspaces/members/types/members_list_response_item.py,sha256=DIc5DJoVahI9olBis_iFgOJrAf05m2fCE8g4R5ZeDko,712
-label_studio_sdk-1.0.13.dist-info/LICENSE,sha256=ymVrFcHiJGjHeY30NWZgdV-xzNEtfuC63oK9ZeMDjhs,11341
-label_studio_sdk-1.0.13.dist-info/METADATA,sha256=-1jBn5dwOCPtA4yi0RN5kq7HY373qPlf5dSWNP1r7YU,6033
-label_studio_sdk-1.0.13.dist-info/WHEEL,sha256=fGIA9gx4Qxk2KDKeNJCbOEwSrmLtjWCwzBz351GyrPQ,88
-label_studio_sdk-1.0.13.dist-info/RECORD,,
+label_studio_sdk-1.0.15.dist-info/LICENSE,sha256=ymVrFcHiJGjHeY30NWZgdV-xzNEtfuC63oK9ZeMDjhs,11341
+label_studio_sdk-1.0.15.dist-info/METADATA,sha256=Ct3zUaMfhUAHG0IuTDHQ3jCjtD0HcPVxN0Ocj-JPajI,6033
+label_studio_sdk-1.0.15.dist-info/WHEEL,sha256=fGIA9gx4Qxk2KDKeNJCbOEwSrmLtjWCwzBz351GyrPQ,88
+label_studio_sdk-1.0.15.dist-info/RECORD,,