davidkhala.ai 0.2.1__py3-none-any.whl → 0.2.2__py3-none-any.whl

davidkhala/ai/model/chat.py

@@ -1,4 +1,6 @@
-from typing import Protocol, Any
+from typing import Protocol, Any, Iterable, TypedDict
+
+from davidkhala.ai.model import ModelAware
 
 
 class MessageProtocol(Protocol):
@@ -13,7 +15,61 @@ class ChoicesAware(Protocol):
     choices: list[ChoiceProtocol]
 
 
-def on_response(response: ChoicesAware, n: int):
+class ImagePromptDict(TypedDict):
+    text: str
+    image_url: list[str]
+
+
+def on_response(response: ChoicesAware, n: int | None):
     contents = [choice.message.content for choice in response.choices]
-    assert len(contents) == n
-    return contents
+    if n:
+        assert len(contents) == n, f"expected {n} choices, but got {len(contents)}"
+    return contents
+
+
+class MessageDict(TypedDict):
+    content: str | list | None
+    role: str
+
+
+def messages_from(*user_prompt: str | ImagePromptDict) -> Iterable[MessageDict]:
+    for _ in user_prompt:
+        message = MessageDict(role='user', content=None)
+        if type(_) == str:
+            message['content'] = _
+        elif type(_) == dict:
+            message['content'] = [{"type": "text", "text": _['text']}]
+            message['content'].extend({"type": "image_url", "image_url": {"url": i}} for i in _['image_url'])
+        yield message
+
+
+class ChatAware(ModelAware):
+    def __init__(self):
+        super().__init__()
+        self.messages: list[Any | MessageDict] = []
+        self.n: int = 1
+
+    def as_chat(self, model: str | None, sys_prompt: str = None):
+        self.model = model
+        if sys_prompt is not None:
+            self.messages = [MessageDict(role='system', content=sys_prompt)]
+
+    def chat(self, *user_prompt, **kwargs): ...
+
+    def messages_from(self, *user_prompt) -> list[MessageDict]:
+        messages = list(self.messages)
+        messages.extend(messages_from(*user_prompt))
+        return messages
+
+
+class CompareChatAware(ChatAware):
+    def __init__(self):
+        super().__init__()
+        self._models = None
+
+    def as_chat(self, *models: str, sys_prompt: str = None):
+        if len(models) > 1:
+            self._models = models
+            super().as_chat(None, sys_prompt)
+        elif len(models) == 1:
+            super().as_chat(models[0], sys_prompt)

davidkhala/ai/model/embed.py

@@ -0,0 +1,8 @@
+from davidkhala.ai.model import ModelAware
+
+
+class EmbeddingAware(ModelAware):
+    def as_embeddings(self, model: str):
+        self.model = model
+
+    def encode(self, *_input: str) -> list[list[float]]: ...

davidkhala/ai/model/garden.py

@@ -0,0 +1,9 @@
+from typing import Protocol
+
+
+class GardenAlike(Protocol):
+    @property
+    def models(self) -> list[str]: ...
+
+    @property
+    def free_models(self) -> list[str]: ...

davidkhala/ai/openai/__init__.py

@@ -1,20 +1,19 @@
 from typing import Literal
 
-from httpx import URL
 from openai import OpenAI
 
-from davidkhala.ai.model import AbstractClient
-from davidkhala.ai.model.chat import on_response
+from davidkhala.ai.model import SDKProtocol, Connectable
+from davidkhala.ai.model.embed import EmbeddingAware
+from davidkhala.ai.model.chat import on_response, ChatAware
+from davidkhala.ai.model.garden import GardenAlike
 
 
-class Client(AbstractClient):
+class Client(ChatAware, EmbeddingAware, SDKProtocol, GardenAlike, Connectable):
     def __init__(self, client: OpenAI):
         super().__init__()
-        self.client:OpenAI = client
-        self.base_url:URL = client.base_url
-        self.api_key = client.api_key
-        self.encoding_format:Literal["float", "base64"] = "float"
-        self.n:int = 1
+        self.client: OpenAI = client
+        self.encoding_format: Literal["float", "base64"] = "float"
+
     def connect(self):
         try:
             type(self).models.fget(self)
@@ -36,29 +35,9 @@ class Client(AbstractClient):
 
     def chat(self, *user_prompt, **kwargs):
 
-        messages = [
-            *self.messages,
-        ]
-        for prompt in user_prompt:
-            message = {
-                "role": "user"
-            }
-            if type(prompt) == str:
-                message['content'] = prompt
-            elif type(prompt) == dict:
-                message['content'] = [
-                    {"type": "text", "text": prompt['text']},
-                    {
-                        "type": "image_url",
-                        "image_url": {
-                            "url": prompt['image_url'],
-                        }
-                    },
-                ]
-            messages.append(message)
         response = self.client.chat.completions.create(
             model=self.model,
-            messages=messages,
+            messages=self.messages_from(*user_prompt),
             n=self.n,
             **kwargs
         )
@@ -67,6 +46,3 @@ class Client(AbstractClient):
 
     def close(self):
         self.client.close()
-
-
-

davidkhala/ai/openai/azure.py

@@ -1,6 +1,18 @@
+import base64
+import json
+from pathlib import Path
+
+from davidkhala.ml.ocr.interface import FieldProperties
 from davidkhala.utils.syntax.compat import deprecated
 from openai import AzureOpenAI, OpenAI
+from openai.types.chat import (
+    ChatCompletionUserMessageParam, ChatCompletionContentPartTextParam, ChatCompletionContentPartImageParam
+)
+from openai.types.chat.chat_completion_content_part_image_param import ImageURL
+from openai.types.shared_params import ResponseFormatJSONSchema
+from openai.types.shared_params.response_format_json_schema import JSONSchema
 
+from davidkhala.ai.model.chat import on_response
 from davidkhala.ai.openai import Client
 
 
@@ -19,6 +31,45 @@ class ModelDeploymentClient(AzureHosted):
             api_key=key,
         ))
 
+    def process(self, file: Path, schema: dict[str, FieldProperties])-> list[dict]:
+        with open(file, "rb") as f:
+            content = base64.b64encode(f.read()).decode("utf-8")
+        required = [k for k, _ in schema.items() if _.required]
+        properties = {k: {'type': v.type} for k, v in schema.items()}
+
+        json_schema = JSONSchema(
+            name='-',
+            schema={"type": "object",
+                    "properties": properties,
+                    "required": required,
+                    },
+        )
+
+        self.messages.append(ChatCompletionUserMessageParam(
+            role='user',
+            content=[
+                ChatCompletionContentPartTextParam(
+                    type='text',
+                    text="Extract the required fields from this image and return the output strictly following the provided JSON schema."),
+                ChatCompletionContentPartImageParam(
+                    type="image_url",
+                    image_url=ImageURL(
+                        url=f"data:image/jpeg;base64,{content}",
+                        detail='auto'
+                    )
+                )
+            ]
+        ))
+        response = self.client.chat.completions.create(
+            model=self.model,
+            messages=self.messages,
+            response_format=ResponseFormatJSONSchema(
+                type='json_schema',
+                json_schema=json_schema
+            ),
+            n=self.n,
+        )
+        return [json.loads(_) for _ in on_response(response, self.n)]
 
 @deprecated("Azure Open AI is deprecated. Please migrate to Microsoft Foundry")
 class OpenAIClient(AzureHosted):

davidkhala/ai/openai/native.py

@@ -1,4 +1,4 @@
-from typing import Optional, Literal
+from typing import Literal
 
 from openai import OpenAI
 
@@ -12,8 +12,7 @@ class NativeClient(Client):
             base_url=base_url
         ))
 
-
-    def chat(self, *user_prompt, web_search:Optional[Literal["low", "medium", "high"]]=None, **kwargs):
+    def chat(self, *user_prompt, web_search: Literal["low", "medium", "high"] = None, **kwargs):
         opts = {
             **kwargs
         }

davidkhala/ai/openrouter/__init__.py

@@ -1,36 +1,47 @@
+from openrouter import OpenRouter
+from openrouter.components import Model
 from openrouter.errors import UnauthorizedResponseError
+from openrouter.operations import ListData
 
-from davidkhala.ai.model import AbstractClient
-from openrouter import OpenRouter
+from davidkhala.ai.model import Connectable, SDKProtocol
+from davidkhala.ai.model.chat import CompareChatAware, on_response
+from davidkhala.ai.model.garden import GardenAlike
 
 
-class Client(AbstractClient):
+class Client(CompareChatAware, Connectable, SDKProtocol, GardenAlike):
     def __init__(self, api_key: str):
         super().__init__()
-        self.api_key = api_key
-        self.client = OpenRouter(api_key)
+        self.client: OpenRouter = OpenRouter(api_key)
 
     def chat(self, *user_prompt, **kwargs):
         r = self.client.chat.send(
             model=self.model,
-            messages=[
-                *self.messages,
-                *[{'role': 'user', 'content': _} for _ in user_prompt]
-            ]
+            models=self._models,
+            messages=self.messages_from(*user_prompt)
+            # openrouter has no n
         )
-        return [_.message.content for _ in r.choices]
+        return on_response(r, None)
+
     def connect(self):
         try:
-            self.client.models.list()
+            _ = self.models
             return True
         except UnauthorizedResponseError:
             return False
 
+    def list_models(self) -> list[Model]:
+        return self.client.models.list().data
+
+    @property
+    def models(self) -> list[str]:
+        return [m.id for m in self.list_models()]
+
 
 class Admin:
     def __init__(self, provisioning_key: str):
         self.provisioning_key = provisioning_key
         self.client = OpenRouter(provisioning_key)
+
     @property
-    def keys(self):
-        return self.client.api_keys.list().data
+    def keys(self) -> list[ListData]:
+        return self.client.api_keys.list().data

{davidkhala_ai-0.2.1.dist-info → davidkhala_ai-0.2.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: davidkhala.ai
-Version: 0.2.1
+Version: 0.2.2
 Summary: misc AI modules
 Requires-Python: >=3.12
 Provides-Extra: ali
@@ -9,7 +9,10 @@ Requires-Dist: davidkhala-utils; extra == 'ali'
 Requires-Dist: wuying-agentbay-sdk; extra == 'ali'
 Provides-Extra: api
 Requires-Dist: davidkhala-utils[http-request]; extra == 'api'
+Provides-Extra: atlas
+Requires-Dist: voyageai; extra == 'atlas'
 Provides-Extra: azure
+Requires-Dist: davidkhala-ml-ocr; extra == 'azure'
 Requires-Dist: davidkhala-utils; extra == 'azure'
 Requires-Dist: openai; extra == 'azure'
 Provides-Extra: dify
@@ -19,16 +22,15 @@ Requires-Dist: dify-plugin; extra == 'dify'
 Provides-Extra: google
 Requires-Dist: google-adk; extra == 'google'
 Requires-Dist: google-genai; extra == 'google'
-Provides-Extra: hf
-Requires-Dist: hf-xet; extra == 'hf'
-Requires-Dist: huggingface-hub; extra == 'hf'
-Requires-Dist: onnx; extra == 'hf'
-Requires-Dist: onnxruntime; extra == 'hf'
 Provides-Extra: langchain
 Requires-Dist: langchain; extra == 'langchain'
 Requires-Dist: langchain-openai; (python_version < '3.14') and extra == 'langchain'
 Requires-Dist: langgraph; extra == 'langchain'
+Provides-Extra: minimax
+Requires-Dist: anthropic; extra == 'minimax'
+Requires-Dist: openai; extra == 'minimax'
 Provides-Extra: mistral
+Requires-Dist: davidkhala-ml-ocr; extra == 'mistral'
 Requires-Dist: mistralai; extra == 'mistral'
 Provides-Extra: openrouter
 Requires-Dist: openrouter; extra == 'openrouter'

davidkhala_ai-0.2.2.dist-info/RECORD

@@ -0,0 +1,65 @@
+davidkhala/ai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/opik.py,sha256=YU1XuweMUAzUkhpjxhltt-SBBDBkR3z-PCNo0DqzBRs,39
+davidkhala/ai/you.py,sha256=fmMfJQZcG0e6sKfVQy2jOKZ-jkO-3H6GOGP2U19l4zQ,1759
+davidkhala/ai/agent/README.md,sha256=kIPsx3gOjrpOw7w2qhNEALuCEQkuh4nYp6uBnijdvHE,178
+davidkhala/ai/agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/agent/langgraph.py,sha256=vaiHpnYmkjfXiGq-VH0sUn-NXBacnbFW-6E6sUgdNLU,1050
+davidkhala/ai/agent/ragflow.py,sha256=UaK31us6V0NhAPCthGo07rQsm72vlR-McmihC_NDe1g,273
+davidkhala/ai/agent/dify/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/agent/dify/const.py,sha256=gU4lPBe4U2taakN2jhdPMRWXkqlyCg-YRE8JJmtsblo,218
+davidkhala/ai/agent/dify/interface.py,sha256=bTOI38ZjtkgoSw-ysgFwBZ1QkKVAa92gjOnERDoagQA,118
+davidkhala/ai/agent/dify/api/__init__.py,sha256=OevdexhW4rh40uGpzSxmeeff4pslKvygKpxjUQzMc-Y,861
+davidkhala/ai/agent/dify/api/app.py,sha256=XMLwNSpL5BhNZHKx4BEuk9NplIHCToMKXFssrn_7Hu4,3931
+davidkhala/ai/agent/dify/api/knowledge/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/agent/dify/api/knowledge/chunk.py,sha256=AurKJCZQQ2RkitLxMRcRPRKOE2WsTm7JSb6AnjJEgEw,470
+davidkhala/ai/agent/dify/api/knowledge/dataset.py,sha256=JO0Au4TTZIzSagkDI1lOULB5SM-WWGnIiQ7HmcZardM,2981
+davidkhala/ai/agent/dify/api/knowledge/document.py,sha256=sjgzwLWar2CsqO8gF93qeRfzw3fNCLgC4PGxD7QhAhE,1261
+davidkhala/ai/agent/dify/api/knowledge/model.py,sha256=fRv4qMsgPMbNU3Q9c03rU4JedoktQ38IjUDrAjQy-8w,3775
+davidkhala/ai/agent/dify/console/__init__.py,sha256=WKXcjjtnWyljhlgrReSneeD87K6GmgGoCmwKCPdGPdc,380
+davidkhala/ai/agent/dify/console/plugin.py,sha256=qwojlnYaIIRqjV8unN1mkMdOT5scG7ZhuVW8Rz8GgVM,2870
+davidkhala/ai/agent/dify/console/session.py,sha256=agS6i7qe-ZGMOawNTh23_bY2ovm0JtGMDLwEuBuRzyk,1604
+davidkhala/ai/agent/dify/console/knowledge/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/agent/dify/console/knowledge/dataset.py,sha256=QgHPJSPQmiosc7WI6zLPBXaWLIEOHHOEZlVWNlPljuc,2474
+davidkhala/ai/agent/dify/console/knowledge/pipeline.py,sha256=yg4pYx3PHKFl88_deexv6GgACyUcmnJXgvB91wXyfO4,4353
+davidkhala/ai/agent/dify/db/__init__.py,sha256=HYfJEnoFAoJJck2xvTDYx8zpw9Qao7sHXOGvW0diPqw,517
+davidkhala/ai/agent/dify/db/app.py,sha256=IRiSiR0v387p4p3J7M9xEkJ7pfQyO5DL6chpx7Z2IzA,1319
+davidkhala/ai/agent/dify/db/knowledge.py,sha256=GVaK5QmU_VxB8fDxV60uiYiIeR3JEn3IXJTlJHLiT5U,2917
+davidkhala/ai/agent/dify/db/orm.py,sha256=AJ2tVS18g-uyIPJGdEVIPAtoPYUm3KqzSl7av0S8uzA,2188
+davidkhala/ai/agent/dify/db/sys.py,sha256=U_qqopUMlgsilhHaG_ids6gtd-pNiR_Jm0kAr9hIL7M,188
+davidkhala/ai/agent/dify/model/__init__.py,sha256=VXKaNpQFK8J-j9zp5caOgiQxaQCIoUJjWn6NAjTTok0,104
+davidkhala/ai/agent/dify/model/knowledge.py,sha256=e4CbI3tjIn0gsh4t4ZOMp-JQtWFB7yzWvkioSztWJx0,397
+davidkhala/ai/agent/dify/model/workflow.py,sha256=I7_INuD5QAFT_xERmOwAMOKZEccGZ5COnvivv5fNisA,3044
+davidkhala/ai/agent/dify/plugins/__init__.py,sha256=iTWvutlkN9bXgptesi05M447nTeF5hKFAIfn4EviFj0,183
+davidkhala/ai/agent/dify/plugins/file.py,sha256=o-HjHSFwRTNIYs8IxqZUSnBbh-xr8f-xMUM3iU9wCCQ,390
+davidkhala/ai/agent/dify/plugins/firecrawl.py,sha256=lB_f8W_bdg-7PeBKmF0-HdwYyakV_0D3nET5iT-Z1KM,460
+davidkhala/ai/agent/dify/plugins/jina.py,sha256=dQ5iJxDLWtChXb1IjCtsHctgUtgjOiDfWOuR2u0aUIM,190
+davidkhala/ai/agent/dify/plugins/popular.py,sha256=XMuxqRcIko4gCQORg5HCcwCrbgLcaoRZxCYBIgtkrOo,812
+davidkhala/ai/ali/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/ali/agentbay.py,sha256=O5t71GGwtDgBE1zUXJDYe5djMVwSaNOwn5k8zg1xa18,1200
+davidkhala/ai/ali/dashscope.py,sha256=4KWxl8s6PrKtjgszYM6OIv6r5736EXC1HwLcfJo-x-8,2091
+davidkhala/ai/anthropic/__init__.py,sha256=zzELo9KUD9wKKzrEPYmt-wBHPIvInHzgU2ClfbdgzYo,112
+davidkhala/ai/api/__init__.py,sha256=FTsOpTnG5FxM80mFUYir2THfKmVQEGYnsoO_Egdypzo,1137
+davidkhala/ai/api/openrouter.py,sha256=BBM0HsBwo29PChI4ca8tNiEbzE16-WWc95TqW6k-xn8,2527
+davidkhala/ai/api/siliconflow.py,sha256=hxaEIQc2xi394SS8VdCXu2TWPxjpgc3hATzB03LsZQ0,2073
+davidkhala/ai/atlas/__init__.py,sha256=hNjmpp9fk-GGXVZOjrvs_x4j0b_Wlew-BlC5RMTUEMA,746
+davidkhala/ai/google/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+davidkhala/ai/google/adk.py,sha256=QwxYoOzT2Hol03V4NM0PF_HAzUGb4fB18VUAYacYbAY,657
+davidkhala/ai/google/gemini.py,sha256=Xf4HDOOcK4-jEBERzuLnQNFsU61P2fFx4K0z-ijvNHE,214
+davidkhala/ai/mistral/__init__.py,sha256=RpBOQFg67KFSuzaZ1LvuMtIpkfm-iTm6ZepkmHyggrg,371
+davidkhala/ai/mistral/agent.py,sha256=hXHQj6MtxmR86Db4YbtMcdMi5NebX_Tg2bxYo_lL0sY,1652
+davidkhala/ai/mistral/ai.py,sha256=SIIVjmR6oY4BjDTnYCiztkSJgfRF7_a3g21KoDQ0fVo,1324
+davidkhala/ai/mistral/file.py,sha256=Y8RnaONasZX6zJDWzrG0Ez7epQ_BS3CKiIK8EqqX9YE,1402
+davidkhala/ai/mistral/ocr.py,sha256=iImHiv784ZYdqKPwbUJhy7EL9WAXSpc4h4ixTZi3FHY,1736
+davidkhala/ai/model/__init__.py,sha256=ODvM3qaEIjNJcJeU6Nkozj77LUWlC1HcPM-ifYVjA5s,490
+davidkhala/ai/model/chat.py,sha256=1RgvdqMh7wK3bxiu5g3lDI89zVeuASSMKgeCJxqsiTI,2178
+davidkhala/ai/model/embed.py,sha256=2W5Z-qKJUaROxuqV4yqqy7w666PDoh6RBj9gWFKWhUY,217
+davidkhala/ai/model/garden.py,sha256=vcG6-fO0SJckNiPJ1wLJzsu6yKOIdQcVDARI9f9WQNg,178
+davidkhala/ai/openai/__init__.py,sha256=-9SayyencR2nvGycYt_HF-8VqUXmQW9lIwYJV95RXCc,1443
+davidkhala/ai/openai/azure.py,sha256=npF0GA-1jkKL6bsDyMrStkJODn1ShTsqdOKzsLu-mgc,3257
+davidkhala/ai/openai/databricks.py,sha256=Z_ES3KlZuwMjMiGnlAxxrirLInQRVXj2KZs32MHXBbI,736
+davidkhala/ai/openai/native.py,sha256=WhtpBCx_iGyy6tolGjJm1_gJtGo-dcmc_7QvnaxQ6WY,691
+davidkhala/ai/openai/opik.py,sha256=ovodFNDoDO_ygX8_EmeqbZcu8PRTntzsNNQg4JoZKWo,257
+davidkhala/ai/openrouter/__init__.py,sha256=8TLj1noeAvEzhowvkbLu2KUiR8KNijv8OMYRt6G7ApQ,1473
+davidkhala_ai-0.2.2.dist-info/METADATA,sha256=pa6tpzTxuokcqYOsmvCx_oLdhS_JZ_QInQGC0PN7RY0,1837
+davidkhala_ai-0.2.2.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+davidkhala_ai-0.2.2.dist-info/RECORD,,

davidkhala/ai/agent/dify/api/knowledge.py

@@ -1,191 +0,0 @@
-from __future__ import annotations
-
-import os
-from pathlib import Path
-from typing import Iterable, TypedDict, Optional
-from urllib.parse import urlparse
-
-import requests
-
-from davidkhala.ai.agent.dify.api import API, Iterator
-from davidkhala.ai.agent.dify.model import Document as DocumentBase
-
-
-class DatasetDict(TypedDict):
-    id: str
-    name: str
-    description: str
-    provider: str
-    permission: str
-    data_source_type: str
-    indexing_technique: str
-    doc_form: str
-    runtime_mode: str
-    is_published: bool
-    enable_api: bool
-    # stats
-    app_count: int
-    document_count: int
-    word_count: int
-    total_documents: int
-    total_available_documents: int
-    # embedding
-    embedding_available: bool
-    embedding_model: str
-    embedding_model_provider: str
-    retrieval_model_dict: dict
-    external_retrieval_model: dict
-    external_knowledge_info: dict
-
-
-class Document(DocumentBase):
-    data_source_info: dict[str, str]
-    data_source_detail_dict: dict[str, dict]
-    dataset_process_rule_id: str
-    created_from: str
-    created_by: str
-    created_at: int
-    tokens: int
-    archived: bool
-    display_status: str
-    word_count: int
-    hit_count: int
-    doc_form: str
-    doc_metadata: dict
-    disabled_at: int
-    disabled_by: str
-
-
-class Dataset(API):
-    def __init__(self, api_key: str, base_url="https://api.dify.ai/v1"):
-        super().__init__(api_key, f"{base_url}/datasets")
-
-    def paginate_datasets(self, page=1, size=20):
-        r = self.request(self.base_url, "GET", params={
-            'page': page,
-            'limit': size,
-        })
-        return r
-
-    def list_datasets(self) -> Iterable[list[DatasetDict]]:
-        return Iterator(self.paginate_datasets, None)
-
-    @property
-    def ids(self):
-        for sub_list in self.list_datasets():
-            for dataset in sub_list:
-                yield dataset['id']
-
-    class Instance(API):
-        def __init__(self, d: Dataset, dataset_id: str):
-            super().__init__(d.api_key, f"{d.base_url}/{dataset_id}")
-
-        def get(self):
-            return self.request(self.base_url, "GET")
-
-        def upload(self, filename, *, path=None, url=None, document_id=None):
-            """
-            don't work for .html
-            work for .md
-            """
-            files = {}
-            if path:
-                with open(path, 'rb') as f:
-                    content = f.read()
-                if not filename:
-                    filename = os.path.basename(path)
-            elif url:
-                r = requests.get(url)
-                r.raise_for_status()
-                if not filename:
-                    parsed_url = urlparse(url)
-                    filename = Path(parsed_url.path).name
-                content = r.content
-            files['file'] = (filename, content)
-            if document_id:
-                # don't work for html
-                r = requests.post(f"{self.base_url}/documents/{document_id}/update-by-file", files=files,
-                                  **self.options)
-            else:
-                r = requests.post(f"{self.base_url}/document/create-by-file", files=files, **self.options)
-            r = self.on_response(r)
-            return r['document']
-
-        def paginate_documents(self, page=1, size=20):
-            return self.request(f"{self.base_url}/documents", "GET", params={
-                'page': page,
-                'limit': size
-            })
-
-        def list_documents(self) -> Iterable[Document]:
-            for document_batch in Iterator(self.paginate_documents, None):
-                for document in document_batch:
-                    yield Document(**document)
-
-        def has_document(self, name) -> bool:
-            return any(name == item['name'] for row in self.list_documents() for item in row)
-
-
-class ChunkDict(TypedDict):
-    id: str
-    position: int
-    document_id: str
-    content: str
-    sign_content: str  # trimmed version of content
-    answer: Optional[str]  # only used in QA chunk
-    word_count: int
-    tokens: int
-    keywords: Optional[list[str]]
-    index_node_id: str  # chunk 在向量索引中的节点 ID
-    index_node_hash: str  # hash of sign_content
-    hit_count: int
-    enabled: bool
-    status: str  # 'completed'
-    created_at: int  # timestamp
-    updated_at: int  # timestamp
-    completed_at: int  # timestamp
-    created_by: str  # user id
-    child_chunks: list
-    error: Optional
-    stopped_at: Optional[int]  # timestamp
-    disabled_at: Optional[int]  # timestamp
-
-
-class Document(API):
-    def __init__(self, d: Dataset.Instance, document_id: str):
-        super().__init__(d.api_key, f"{d.base_url}/documents/{document_id}")
-
-    def exist(self):
-        try:
-            self.get()
-            return True
-        except requests.exceptions.HTTPError as e:
-            if e.response.status_code == 404:
-                return False
-            else:
-                raise e
-
-    def get(self):
-        return self.request(self.base_url, "GET")
-
-    def paginate_chunks(self, page=1, size=20):
-        return self.request(f"{self.base_url}/segments", "GET", params={
-            'page': page,
-            'limit': size
-        })
-
-    def list_chunks(self) -> Iterable[ChunkDict]:
-        for chunk_batch in Iterator(self.paginate_chunks, None):
-            for chunk in chunk_batch:
-                yield chunk
-
-    def delete(self):
-        if self.exist():
-            self.request(self.base_url, "DELETE")
-class Chunk(API):
-    def __init__(self, d: Document, segment_id: str):
-        super().__init__(d.api_key, f"{d.base_url}/segments/{segment_id}")
-    def get(self):
-        r=  self.request(self.base_url, "GET")
-        assert r['doc_form'] # optional value text_model
-        return r['data']

davidkhala/ai/agent/dify/ops/__init__.py

@@ -1 +0,0 @@
-