chore: 🐝 Update SDK - Generate

.speakeasy/gen.lock

@@ -1,26 +1,26 @@
 lockVersion: 2.0.0
 id: 8b5fa338-9106-4734-abf0-e30d67044a90
 management:
-  docChecksum: 903444f359d1dfa6342c692ae3e5c7ff
+  docChecksum: a112aea005467aa6818696fa4e99fcfe
   docVersion: 0.0.1
   speakeasyVersion: internal
-  generationVersion: 2.250.19
-  releaseVersion: 0.18.0
-  configChecksum: 938a4a39baa5695a3140be3b858483d4
+  generationVersion: 2.277.0
+  releaseVersion: 0.21.0
+  configChecksum: c5e7c8526f43272d7585627468d8c4e5
   repoURL: https://github.com/Unstructured-IO/unstructured-python-client.git
   repoSubDirectory: .
   installationURL: https://github.com/Unstructured-IO/unstructured-python-client.git
   published: true
 features:
   python:
-    core: 4.4.5
+    core: 4.5.0
     examples: 2.81.3
-    globalSecurity: 2.83.2
+    globalSecurity: 2.83.4
     globalServerURLs: 2.82.1
     nameOverrides: 2.81.1
     retries: 2.82.1
     serverIDs: 2.81.1
-    unions: 2.82.5
+    unions: 2.82.6
 generatedFiles:
   - src/unstructured_client/sdkconfiguration.py
   - src/unstructured_client/general.py
@@ -51,3 +51,6 @@ generatedFiles:
   - docs/models/shared/security.md
   - USAGE.md
   - .gitattributes
+  - src/unstructured_client/_hooks/sdkhooks.py
+  - src/unstructured_client/_hooks/types.py
+  - src/unstructured_client/_hooks/__init__.py

RELEASES.md

@@ -424,4 +424,34 @@ Based on:
 ### Generated
 - [python v0.18.0] .
 ### Releases
-- [PyPI v0.18.0] https://pypi.org/project/unstructured-client/0.18.0 - .
+- [PyPI v0.18.0] https://pypi.org/project/unstructured-client/0.18.0 - .
+
+## 2024-02-19 00:19:41
+### Changes
+Based on:
+- OpenAPI Doc 0.0.64 
+- Speakeasy CLI 1.183.2 (2.262.2) https://github.com/speakeasy-api/speakeasy
+### Generated
+- [python v0.19.0] .
+### Releases
+- [PyPI v0.19.0] https://pypi.org/project/unstructured-client/0.19.0 - .
+
+## 2024-02-22 00:18:37
+### Changes
+Based on:
+- OpenAPI Doc 0.0.1 
+- Speakeasy CLI 1.189.0 (2.263.3) https://github.com/speakeasy-api/speakeasy
+### Generated
+- [python v0.20.0] .
+### Releases
+- [PyPI v0.20.0] https://pypi.org/project/unstructured-client/0.20.0 - .
+
+## 2024-03-01 23:20:07
+### Changes
+Based on:
+- OpenAPI Doc 0.0.1 
+- Speakeasy CLI 1.200.0 (2.277.0) https://github.com/speakeasy-api/speakeasy
+### Generated
+- [python v0.21.0] .
+### Releases
+- [PyPI v0.21.0] https://pypi.org/project/unstructured-client/0.21.0 - .

USAGE.md

@@ -15,10 +15,6 @@ req = shared.PartitionParameters(
         'image',
         'table',
     ],
-    files=shared.Files(
-        content='0x2cC94b2FEF'.encode(),
-        file_name='um.shtml',
-    ),
     gz_uncompressed_content_type='application/pdf',
     hi_res_model_name='yolox',
     languages=[
@@ -31,9 +27,8 @@ req = shared.PartitionParameters(
     max_characters=1500,
     new_after_n_chars=1500,
     output_format='application/json',
-    skip_infer_table_types=[
-        'pdf',
-    ],
+    overlap=25,
+    overlap_all=True,
     strategy='hi_res',
 )
 

gen.yaml

@@ -7,8 +7,14 @@ generation:
     nameResolutionDec2023: false
     parameterOrderingFeb2024: false
     requestResponseComponentNamesFeb2024: false
+  auth:
+    oAuth2ClientCredentialsEnabled: false
 python:
-  version: 0.18.0
+  version: 0.21.0
+  additionalDependencies:
+    dependencies: {}
+    extraDependencies:
+      dev: {}
   author: Unstructured
   clientServerStatusCodesAsErrors: true
   description: Python Client SDK for Unstructured API

pylintrc

@@ -179,7 +179,8 @@ good-names=i,
            k,
            ex,
            Run,
-           _
+           _,
+           e
 
 # Good variable names regexes, separated by a comma. If names match any regex,
 # they will always be accepted

setup.py

@@ -10,7 +10,7 @@
 
 setuptools.setup(
     name="unstructured-client",
-    version="0.18.0",
+    version="0.21.0",
     author="Unstructured",
     description="Python Client SDK for Unstructured API",
     license = "MIT",
@@ -20,9 +20,9 @@
     install_requires=[
         "certifi>=2023.7.22",
         "charset-normalizer>=3.2.0",
-        "dataclasses-json-speakeasy>=0.5.11",
+        "dataclasses-json>=0.6.4",
         "idna>=3.4",
-        "jsonpath-python>=1.0.6 ",
+        "jsonpath-python>=1.0.6",
         "marshmallow>=3.19.0",
         "mypy-extensions>=1.0.0",
         "packaging>=23.1",
@@ -34,7 +34,9 @@
         "urllib3>=1.26.18",
     ],
     extras_require={
-        "dev":["pylint==2.16.2"]
+        "dev": [
+            "pylint==2.16.2",
+        ],
     },
     package_dir={'': 'src'},
     python_requires='>=3.8',

src/unstructured_client/_hooks/__init__.py

@@ -0,0 +1,4 @@
+"""Code generated by Speakeasy (https://speakeasyapi.dev). DO NOT EDIT."""
+
+from .sdkhooks import *
+from .types import *

src/unstructured_client/_hooks/sdkhooks.py

@@ -0,0 +1,55 @@
+"""Code generated by Speakeasy (https://speakeasyapi.dev). DO NOT EDIT."""
+
+import requests
+from .types import SDKInitHook, BeforeRequestContext, BeforeRequestHook, AfterSuccessContext, AfterSuccessHook, AfterErrorContext, AfterErrorHook, Hooks
+from typing import List, Optional, Tuple, Union
+
+
+class SDKHooks(Hooks):
+    sdk_init_hooks: List[SDKInitHook] = []
+    before_request_hooks: List[BeforeRequestHook] = []
+    after_success_hooks: List[AfterSuccessHook] = []
+    after_error_hooks: List[AfterErrorHook] = []
+
+    def __init__(self):
+        pass
+
+    def register_sdk_init_hook(self, hook: SDKInitHook) -> None:
+        self.sdk_init_hooks.append(hook)
+
+    def register_before_request_hook(self, hook: BeforeRequestHook) -> None:
+        self.before_request_hooks.append(hook)
+
+    def register_after_success_hook(self, hook: AfterSuccessHook) -> None:
+        self.after_success_hooks.append(hook)
+
+    def register_after_error_hook(self, hook: AfterErrorHook) -> None:
+        self.after_error_hooks.append(hook)
+
+    def sdk_init(self, base_url: str, client: requests.Session) -> Tuple[str, requests.Session]:
+        for hook in self.sdk_init_hooks:
+            base_url, client = hook.sdk_init(base_url, client)
+        return base_url, client
+
+    def before_request(self, hook_ctx: BeforeRequestContext, request: requests.PreparedRequest) -> Union[requests.PreparedRequest, Exception]:
+        for hook in self.before_request_hooks:
+            request = hook.before_request(hook_ctx, request)
+            if isinstance(request, Exception):
+                raise request
+
+        return request
+
+    def after_success(self, hook_ctx: AfterSuccessContext, response: requests.Response) -> requests.Response:
+        for hook in self.after_success_hooks:
+            response = hook.after_success(hook_ctx, response)
+            if isinstance(response, Exception):
+                raise response
+        return response
+
+    def after_error(self, hook_ctx: AfterErrorContext, response: Optional[requests.Response], error: Optional[Exception]) -> Tuple[Optional[requests.Response], Optional[Exception]]:
+        for hook in self.after_error_hooks:
+            result = hook.after_error(hook_ctx, response, error)
+            if isinstance(result, Exception):
+                raise result
+            response, error = result
+        return response, error

src/unstructured_client/_hooks/types.py

@@ -0,0 +1,70 @@
+"""Code generated by Speakeasy (https://speakeasyapi.dev). DO NOT EDIT."""
+
+import requests as requests_http
+from abc import ABC, abstractmethod
+from typing import Any, Callable, List, Optional, Tuple, Union
+
+
+class HookContext:
+    operation_id: str
+    oauth2_scopes: Optional[List[str]] = None
+    security_source: Optional[Union[Any, Callable[[], Any]]] = None
+
+    def __init__(self, operation_id: str, oauth2_scopes: Optional[List[str]], security_source: Optional[Union[Any, Callable[[], Any]]]):
+        self.operation_id = operation_id
+        self.oauth2_scopes = oauth2_scopes
+        self.security_source = security_source
+
+
+class BeforeRequestContext(HookContext):
+    pass
+
+
+class AfterSuccessContext(HookContext):
+    pass
+
+
+class AfterErrorContext(HookContext):
+    pass
+
+
+class SDKInitHook(ABC):
+    @abstractmethod
+    def sdk_init(self, base_url: str, client: requests_http.Session) -> Tuple[str, requests_http.Session]:
+        pass
+
+
+class BeforeRequestHook(ABC):
+    @abstractmethod
+    def before_request(self, hook_ctx: BeforeRequestContext, request: requests_http.PreparedRequest) -> Union[requests_http.PreparedRequest, Exception]:
+        pass
+
+
+class AfterSuccessHook(ABC):
+    @abstractmethod
+    def after_success(self, hook_ctx: AfterSuccessContext, response: requests_http.Response) -> Union[requests_http.PreparedRequest, Exception]:
+        pass
+
+
+class AfterErrorHook(ABC):
+    @abstractmethod
+    def after_error(self, hook_ctx: AfterErrorContext, response: Optional[requests_http.Response], error: Optional[Exception]) -> Union[Tuple[Optional[requests_http.PreparedRequest], Optional[Exception]], Exception]:
+        pass
+
+
+class Hooks(ABC):
+    @abstractmethod
+    def register_sdk_init_hook(self, hook: SDKInitHook):
+        pass
+
+    @abstractmethod
+    def register_before_request_hook(self, hook: BeforeRequestHook):
+        pass
+
+    @abstractmethod
+    def register_after_success_hook(self, hook: AfterSuccessHook):
+        pass
+
+    @abstractmethod
+    def register_after_error_hook(self, hook: AfterErrorHook):
+        pass

src/unstructured_client/general.py

@@ -1,8 +1,10 @@
 """Code generated by Speakeasy (https://speakeasyapi.dev). DO NOT EDIT."""
 
+import requests as requests_http
 from .sdkconfiguration import SDKConfiguration
 from typing import Any, List, Optional
 from unstructured_client import utils
+from unstructured_client._hooks import HookContext
 from unstructured_client.models import errors, operations, shared
 from unstructured_client.utils._human_utils import suggest_defining_url_if_401  # human code
 
@@ -13,15 +15,17 @@ def __init__(self, sdk_config: SDKConfiguration) -> None:
         self.sdk_configuration = sdk_config
 
 
+
     @suggest_defining_url_if_401  # human code
     def partition(self, request: Optional[shared.PartitionParameters], retries: Optional[utils.RetryConfig] = None) -> operations.PartitionResponse:
         r"""Pipeline 1"""
+        hook_ctx = HookContext(operation_id='partition', oauth2_scopes=[], security_source=self.sdk_configuration.security)
         base_url = utils.template_url(*self.sdk_configuration.get_server_details())
 
         url = base_url + '/general/v0/general'
         headers = {}
         req_content_type, data, form = utils.serialize_request_body(request, Optional[shared.PartitionParameters], "request", False, True, 'multipart')
-        if req_content_type not in ('multipart/form-data', 'multipart/mixed'):
+        if req_content_type is not None and req_content_type not in ('multipart/form-data', 'multipart/mixed'):
             headers['content-type'] = req_content_type
         headers['Accept'] = 'application/json'
         headers['user-agent'] = self.sdk_configuration.user_agent
@@ -31,6 +35,7 @@ def partition(self, request: Optional[shared.PartitionParameters], retries: Opti
         else:
             client = utils.configure_security_client(self.sdk_configuration.client, self.sdk_configuration.security)
 
+
         global_retry_config = self.sdk_configuration.retry_config
         retry_config = retries
         if retry_config is None:
@@ -40,11 +45,32 @@ def partition(self, request: Optional[shared.PartitionParameters], retries: Opti
                 retry_config = utils.RetryConfig('backoff', utils.BackoffStrategy(500, 60000, 1.5, 900000), True)
 
         def do_request():
-            return client.request('POST', url, data=data, files=form, headers=headers)
+            try:
+                req = self.sdk_configuration.get_hooks().before_request(
+                    hook_ctx, 
+                    requests_http.Request('POST', url, data=data, files=form, headers=headers).prepare(),
+                )
+                http_res = client.send(req)
+            except Exception as e:
+                _, e = self.sdk_configuration.get_hooks().after_error(hook_ctx, None, e)
+                raise e
+
+            if utils.match_status_codes(['422','4XX','5XX'], http_res.status_code):
+                http_res, e = self.sdk_configuration.get_hooks().after_error(hook_ctx, http_res, None)
+                if e:
+                    raise e
+            else:
+                result = self.sdk_configuration.get_hooks().after_success(hook_ctx, http_res)
+                if isinstance(result, Exception):
+                    raise result
+                http_res = result
+
+            return http_res
 
         http_res = utils.retry(do_request, utils.Retries(retry_config, [
             '5xx'
         ]))
+
         content_type = http_res.headers.get('Content-Type')
 
         res = operations.PartitionResponse(status_code=http_res.status_code, content_type=content_type, raw_response=http_res)
@@ -67,4 +93,4 @@ def do_request():
 
         return res
 
-    
+

src/unstructured_client/models/shared/partition_parameters.py

@@ -43,6 +43,10 @@ class PartitionParameters:
     r"""If chunking strategy is set, cut off new sections after reaching a length of n chars (soft max). Default: 1500"""
     output_format: Optional[str] = dataclasses.field(default=None, metadata={'multipart_form': { 'field_name': 'output_format' }})
     r"""The format of the response. Supported formats are application/json and text/csv. Default: application/json."""
+    overlap: Optional[int] = dataclasses.field(default=None, metadata={'multipart_form': { 'field_name': 'overlap' }})
+    r"""A prefix of this many trailing characters from prior text-split chunk is applied to second and later chunks formed from oversized elements by text-splitting. Default: None"""
+    overlap_all: Optional[bool] = dataclasses.field(default=None, metadata={'multipart_form': { 'field_name': 'overlap_all' }})
+    r"""When True, overlap is also applied to 'normal' chunks formed by combining whole elements. Use with caution as this can introduce noise into otherwise clean semantic units. Default: None"""
     pdf_infer_table_structure: Optional[bool] = dataclasses.field(default=None, metadata={'multipart_form': { 'field_name': 'pdf_infer_table_structure' }})
     r"""If True and strategy=hi_res, any Table Elements extracted from a PDF will include an additional metadata field, 'text_as_html', where the value (string) is a just a transformation of the data into an HTML <table>."""
     skip_infer_table_types: Optional[List[str]] = dataclasses.field(default=None, metadata={'multipart_form': { 'field_name': 'skip_infer_table_types' }})

src/unstructured_client/sdk.py

@@ -5,6 +5,7 @@
 from .sdkconfiguration import SDKConfiguration
 from typing import Callable, Dict, Union
 from unstructured_client import utils
+from unstructured_client._hooks import SDKHooks
 from unstructured_client.models import shared
 from unstructured_client.utils._human_utils import clean_server_url  # human code
 
@@ -52,6 +53,16 @@ def security():
                 server_url = utils.template_url(server_url, url_params)
 
         self.sdk_configuration = SDKConfiguration(client, security, server_url, server, retry_config=retry_config)
+
+        hooks = SDKHooks()
+
+        current_server_url, *_ = self.sdk_configuration.get_server_details()
+        server_url, self.sdk_configuration.client = hooks.sdk_init(current_server_url, self.sdk_configuration.client)
+        if current_server_url != server_url:
+            self.sdk_configuration.server_url = server_url
+
+        # pylint: disable=protected-access
+        self.sdk_configuration._hooks=hooks
 
         self._init_sdks()
 

src/unstructured_client/sdkconfiguration.py

@@ -1,10 +1,12 @@
 """Code generated by Speakeasy (https://speakeasyapi.dev). DO NOT EDIT."""
 
-import requests
-from dataclasses import dataclass
-from typing import Dict, Tuple, Callable, Union
-from .utils.retries import RetryConfig
+
+import requests as requests_http
+from ._hooks import SDKHooks
 from .utils import utils
+from .utils.retries import RetryConfig
+from dataclasses import dataclass
+from typing import Callable, Dict, Tuple, Union
 from unstructured_client.models import shared
 
 
@@ -21,16 +23,17 @@
 
 @dataclass
 class SDKConfiguration:
-    client: requests.Session
+    client: requests_http.Session
     security: Union[shared.Security,Callable[[], shared.Security]] = None
     server_url: str = ''
     server: str = ''
     language: str = 'python'
     openapi_doc_version: str = '0.0.1'
-    sdk_version: str = '0.18.0'
-    gen_version: str = '2.250.19'
-    user_agent: str = 'speakeasy-sdk/python 0.18.0 2.250.19 0.0.1 unstructured-client'
+    sdk_version: str = '0.21.0'
+    gen_version: str = '2.277.0'
+    user_agent: str = 'speakeasy-sdk/python 0.21.0 2.277.0 0.0.1 unstructured-client'
     retry_config: RetryConfig = None
+    _hooks: SDKHooks = None
 
     def get_server_details(self) -> Tuple[str, Dict[str, str]]:
         if self.server_url:
@@ -39,3 +42,7 @@ def get_server_details(self) -> Tuple[str, Dict[str, str]]:
             self.server = SERVER_PROD
 
         return SERVERS[self.server], {}
+
+
+    def get_hooks(self) -> SDKHooks:
+        return self._hooks

src/unstructured_client/utils/utils.py

@@ -21,15 +21,16 @@
 class SecurityClient:
     client: requests.Session
     query_params: Dict[str, str] = {}
+    headers: Dict[str, str] = {}
 
     def __init__(self, client: requests.Session):
         self.client = client
 
-    def request(self, method, url, **kwargs):
-        params = kwargs.get('params', {})
-        kwargs["params"] = {**self.query_params, **params}
+    def send(self, request: requests.PreparedRequest, **kwargs):
+        request.prepare_url(url=request.url, params=self.query_params)
+        request.headers.update(self.headers)
 
-        return self.client.request(method, url, **kwargs)
+        return self.client.send(request, **kwargs)
 
 
 def configure_security_client(client: requests.Session, security: dataclass):
@@ -102,20 +103,19 @@ def _parse_security_scheme_value(client: SecurityClient, scheme_metadata: Dict,
 
     if scheme_type == "apiKey":
         if sub_type == 'header':
-            client.client.headers[header_name] = value
+            client.headers[header_name] = value
         elif sub_type == 'query':
             client.query_params[header_name] = value
-        elif sub_type == 'cookie':
-            client.client.cookies[header_name] = value
         else:
             raise Exception('not supported')
     elif scheme_type == "openIdConnect":
-        client.client.headers[header_name] = _apply_bearer(value)
+        client.headers[header_name] = _apply_bearer(value)
     elif scheme_type == 'oauth2':
-        client.client.headers[header_name] = _apply_bearer(value)
+        if sub_type != 'client_credentials':
+            client.headers[header_name] = _apply_bearer(value)
     elif scheme_type == 'http':
         if sub_type == 'bearer':
-            client.client.headers[header_name] = _apply_bearer(value)
+            client.headers[header_name] = _apply_bearer(value)
         else:
             raise Exception('not supported')
     else:
@@ -145,7 +145,7 @@ def _parse_basic_auth_scheme(client: SecurityClient, scheme: dataclass):
             password = value
 
     data = f'{username}:{password}'.encode()
-    client.client.headers['Authorization'] = f'Basic {base64.b64encode(data).decode()}'
+    client.headers['Authorization'] = f'Basic {base64.b64encode(data).decode()}'
 
 
 def generate_url(clazz: type, server_url: str, path: str, path_params: dataclass,
@@ -264,7 +264,8 @@ def get_query_params(clazz: type, query_params: dataclass, gbls: Dict[str, Dict[
         f_name = metadata.get("field_name")
         serialization = metadata.get('serialization', '')
         if serialization != '':
-            serialized_parms = _get_serialized_params(metadata, field.type, f_name, value)
+            serialized_parms = _get_serialized_params(
+                metadata, field.type, f_name, value)
             for key, value in serialized_parms.items():
                 if key in params:
                     params[key].extend(value)
@@ -312,7 +313,8 @@ def _get_serialized_params(metadata: Dict, field_type: type, field_name: str, ob
 
     serialization = metadata.get('serialization', '')
     if serialization == 'json':
-        params[metadata.get("field_name", field_name)] = marshal_json(obj, field_type)
+        params[metadata.get("field_name", field_name)
+               ] = marshal_json(obj, field_type)
 
     return params
 
@@ -702,7 +704,8 @@ def unmarshal_json(data, typ, decoder=None):
 
 def marshal_json(val, typ, encoder=None):
     if not is_optional_type(typ) and val is None:
-        raise ValueError(f"Could not marshal None into non-optional type: {typ}")
+        raise ValueError(
+            f"Could not marshal None into non-optional type: {typ}")
 
     marshal = make_dataclass('Marshal', [('res', typ)],
                              bases=(DataClassJsonMixin,))
@@ -732,6 +735,16 @@ def match_content_type(content_type: str, pattern: str) -> boolean:
     return False
 
 
+def match_status_codes(status_codes: List[str], status_code: int) -> bool:
+    for code in status_codes:
+        if code == str(status_code):
+            return True
+
+        if code.endswith("XX") and code.startswith(str(status_code)[:1]):
+            return True
+    return False
+
+
 def datetimeisoformat(optional: bool):
     def isoformatoptional(val):
         if optional and val is None:
@@ -835,13 +848,15 @@ def list_decode(val: List):
 
     return list_decode
 
+
 def union_encoder(all_encoders: Dict[str, Callable]):
     def selective_encoder(val: any):
         if type(val) in all_encoders:
             return all_encoders[type(val)](val)
         return val
     return selective_encoder
 
+
 def union_decoder(all_decoders: List[Callable]):
     def selective_decoder(val: any):
         decoded = val
@@ -854,6 +869,7 @@ def selective_decoder(val: any):
         return decoded
     return selective_decoder
 
+
 def get_field_name(name):
     def override(_, _field_name=name):
         return _field_name