knowhere-python-sdk 0.2.1__tar.gz → 0.3.1__tar.gz

knowhere_python_sdk-0.3.1/.release-please-manifest.json

@@ -0,0 +1,3 @@
+{
+  ".": "0.3.1"
+}

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/CHANGELOG.md

@@ -1,5 +1,21 @@
 # Changelog
 
+## [0.3.1](https://github.com/Ontos-AI/knowhere-python-sdk/compare/v0.3.0...v0.3.1) (2026-04-22)
+
+
+### Documentation
+
+* clarify ParseResult document scope ([861084e](https://github.com/Ontos-AI/knowhere-python-sdk/commit/861084e34144987994fa618ac0db262ce681b5a8))
+* clarify ParseResult document scope ([bb14ad4](https://github.com/Ontos-AI/knowhere-python-sdk/commit/bb14ad4077c41cbe74a5dd155995d6f9937962b8))
+
+## [0.3.0](https://github.com/Ontos-AI/knowhere-python-sdk/compare/v0.2.1...v0.3.0) (2026-04-21)
+
+
+### Features
+
+* add retrieval service sdk clients ([bceef5c](https://github.com/Ontos-AI/knowhere-python-sdk/commit/bceef5cf379dba39543244bd6ca86262a536fb9b))
+* integrate retrieval service v1 in Python SDK ([bce7aa8](https://github.com/Ontos-AI/knowhere-python-sdk/commit/bce7aa8dbf069d5880b92c6f9d8996878251f7cb))
+
 ## [0.2.1](https://github.com/Ontos-AI/knowhere-python-sdk/compare/v0.2.0...v0.2.1) (2026-04-09)
 
 

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: knowhere-python-sdk
-Version: 0.2.1
+Version: 0.3.1
 Summary: Official Python SDK for the Knowhere document parsing API
 Project-URL: Homepage, https://knowhereto.ai
 Project-URL: Documentation, https://docs.knowhereto.ai
@@ -64,6 +64,84 @@ for chunk in result.text_chunks:
     print(chunk.content[:80])
 ```
 
+## Retrieval and document lifecycle
+
+New documents are published into a retrieval namespace. The server returns a
+stable `document_id` after the job is published. `client.jobs.create(...)`
+does not return a usable `document_id`; persist `job_result.document_id` if you
+need to update or archive the same document later.
+
+```python
+job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual.pdf",
+    namespace="support-center",
+)
+
+job_result = client.jobs.wait(job.job_id)
+document_id = job_result.document_id
+
+if document_id is None:
+    raise RuntimeError("Expected document_id after successful publication.")
+```
+
+After the job is done and published, query the canonical document content:
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="How do I reset Bluetooth pairing?",
+    top_k=5,
+    channels=["path", "term"],
+    filter_mode="keep",
+    signal_paths=["Bluetooth", "Pairing"],
+)
+
+print(response.router_used)
+
+for result in response.results:
+    print(result.content)
+    print(result.score)
+    print(result.source.source_file_name, result.source.section_path)
+```
+
+Use `document_id` to update or archive a document:
+
+```python
+update_job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual-v2.pdf",
+    document_id=document_id,
+)
+
+document = client.documents.get(document_id)
+print(document.status)
+
+client.documents.archive(document_id)
+```
+
+You can also list documents in a namespace:
+
+```python
+documents = client.documents.list(namespace="support-center")
+for document in documents.documents:
+    print(document.document_id, document.status)
+```
+
+Retrieval supports exclusions when clients want follow-up results that avoid
+previously used documents or sections:
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="battery charging",
+    exclude_document_ids=["doc_old"],
+    exclude_sections=[
+        {"document_id": "doc_123", "section_path": "Appendix / Legal"}
+    ],
+)
+```
+
 While you can provide an `api_key` keyword argument, we recommend using [python-dotenv](https://pypi.org/project/python-dotenv/) to add `KNOWHERE_API_KEY="sk_..."` to your `.env` file so that your API key is not stored in source control.
 
 ### Parse a local file
@@ -78,6 +156,8 @@ result = client.parse(
 
 print(result.manifest.source_file_name)  # "report.pdf"
 print(len(result.chunks))                # 152
+print(result.namespace)                  # "default" or your explicit namespace
+print(result.document_id)                # Published canonical document id
 ```
 
 ### Access different chunk types
@@ -137,6 +217,7 @@ from pathlib import Path
 job = client.jobs.create(
     source_type="file",
     file_name="report.pdf",
+    namespace="support-center",
     parsing_params={"model": "advanced", "ocr_enabled": True},
 )
 
@@ -146,6 +227,8 @@ client.jobs.upload(job, file=Path("report.pdf"))
 # Step 3: Poll until done (adaptive backoff)
 job_result = client.jobs.wait(job.job_id, poll_interval=10.0, poll_timeout=1800.0)
 
+print(job_result.document_id)  # Persist this to update/archive the document later.
+
 # Step 4: Download and parse results
 result = client.jobs.load(job_result)
 print(result.statistics)

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/README.md

@@ -32,6 +32,84 @@ for chunk in result.text_chunks:
     print(chunk.content[:80])
 ```
 
+## Retrieval and document lifecycle
+
+New documents are published into a retrieval namespace. The server returns a
+stable `document_id` after the job is published. `client.jobs.create(...)`
+does not return a usable `document_id`; persist `job_result.document_id` if you
+need to update or archive the same document later.
+
+```python
+job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual.pdf",
+    namespace="support-center",
+)
+
+job_result = client.jobs.wait(job.job_id)
+document_id = job_result.document_id
+
+if document_id is None:
+    raise RuntimeError("Expected document_id after successful publication.")
+```
+
+After the job is done and published, query the canonical document content:
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="How do I reset Bluetooth pairing?",
+    top_k=5,
+    channels=["path", "term"],
+    filter_mode="keep",
+    signal_paths=["Bluetooth", "Pairing"],
+)
+
+print(response.router_used)
+
+for result in response.results:
+    print(result.content)
+    print(result.score)
+    print(result.source.source_file_name, result.source.section_path)
+```
+
+Use `document_id` to update or archive a document:
+
+```python
+update_job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual-v2.pdf",
+    document_id=document_id,
+)
+
+document = client.documents.get(document_id)
+print(document.status)
+
+client.documents.archive(document_id)
+```
+
+You can also list documents in a namespace:
+
+```python
+documents = client.documents.list(namespace="support-center")
+for document in documents.documents:
+    print(document.document_id, document.status)
+```
+
+Retrieval supports exclusions when clients want follow-up results that avoid
+previously used documents or sections:
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="battery charging",
+    exclude_document_ids=["doc_old"],
+    exclude_sections=[
+        {"document_id": "doc_123", "section_path": "Appendix / Legal"}
+    ],
+)
+```
+
 While you can provide an `api_key` keyword argument, we recommend using [python-dotenv](https://pypi.org/project/python-dotenv/) to add `KNOWHERE_API_KEY="sk_..."` to your `.env` file so that your API key is not stored in source control.
 
 ### Parse a local file
@@ -46,6 +124,8 @@ result = client.parse(
 
 print(result.manifest.source_file_name)  # "report.pdf"
 print(len(result.chunks))                # 152
+print(result.namespace)                  # "default" or your explicit namespace
+print(result.document_id)                # Published canonical document id
 ```
 
 ### Access different chunk types
@@ -105,6 +185,7 @@ from pathlib import Path
 job = client.jobs.create(
     source_type="file",
     file_name="report.pdf",
+    namespace="support-center",
     parsing_params={"model": "advanced", "ocr_enabled": True},
 )
 
@@ -114,6 +195,8 @@ client.jobs.upload(job, file=Path("report.pdf"))
 # Step 3: Poll until done (adaptive backoff)
 job_result = client.jobs.wait(job.job_id, poll_interval=10.0, poll_timeout=1800.0)
 
+print(job_result.document_id)  # Persist this to update/archive the document later.
+
 # Step 4: Download and parse results
 result = client.jobs.load(job_result)
 print(result.statistics)

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/docs/usage.md

@@ -12,6 +12,7 @@ Comprehensive reference for every feature, parameter, and pattern in the SDK.
 - [Working with Results](#working-with-results)
 - [Chunk Types](#chunk-types)
 - [Step-by-Step Control (Jobs API)](#step-by-step-control-jobs-api)
+- [Retrieval and Document Lifecycle](#retrieval-and-document-lifecycle)
 - [Async Usage](#async-usage)
 - [Progress Callbacks](#progress-callbacks)
 - [Error Handling](#error-handling)
@@ -316,8 +317,10 @@ from pathlib import Path
 job = client.jobs.create(
     source_type="file",
     file_name="report.pdf",
+    namespace="support-center",
     parsing_params={"model": "advanced", "ocr_enabled": True},
 )
+print(job.document_id)  # Persist this value for update/archive flows.
 
 # Step 2: Upload file to the presigned URL
 client.jobs.upload(job, file=Path("report.pdf"))
@@ -341,6 +344,8 @@ print(result.statistics)
 | `source_type` | `"url" \| "file"` | — | Required. Whether parsing from URL or uploaded file. |
 | `source_url` | `str \| None` | `None` | URL to parse (required when `source_type="url"`). |
 | `file_name` | `str \| None` | `None` | Original filename (used when `source_type="file"`). |
+| `namespace` | `str \| None` | `None` | Retrieval namespace. The server defaults to `"default"` when omitted. |
+| `document_id` | `str \| None` | `None` | Existing document ID when creating an update job. Omit for a new document. |
 | `data_id` | `str \| None` | `None` | Your own correlation/idempotency identifier. |
 | `parsing_params` | `ParsingParams \| None` | `None` | Parsing configuration. |
 | `webhook` | `WebhookConfig \| None` | `None` | Webhook for completion notification. |
@@ -351,6 +356,8 @@ Returns a `Job` object:
 job.job_id          # "abc-123"
 job.status          # "pending"
 job.source_type     # "file"
+job.namespace       # "support-center"
+job.document_id     # "doc_..." — persist this for updates and archive calls
 job.upload_url      # presigned URL (for file uploads)
 job.upload_headers  # headers to include in the upload request
 job.expires_in      # seconds until upload URL expires
@@ -407,6 +414,119 @@ result = client.jobs.load("https://storage.example.com/result.zip")
 
 ---
 
+## Retrieval and Document Lifecycle
+
+The retrieval APIs operate on canonical documents that are published after a
+job completes. For new documents, the server generates `document_id` during
+`jobs.create()`. Store that ID in your application if you need to update or
+archive the same document later.
+
+### Create a retrievable document
+
+```python
+job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual.pdf",
+    namespace="support-center",
+)
+
+print(job.document_id)  # "doc_..."
+```
+
+For file uploads, the flow is the same except that you upload the file before
+polling:
+
+```python
+job = client.jobs.create(
+    source_type="file",
+    file_name="manual.pdf",
+    namespace="support-center",
+)
+client.jobs.upload(job, file=Path("manual.pdf"))
+job_result = client.jobs.wait(job.job_id)
+```
+
+### Update an existing document
+
+Pass the prior `document_id` to create an update job. If `namespace` is omitted,
+the API resolves the namespace from the existing document.
+
+```python
+update_job = client.jobs.create(
+    source_type="url",
+    source_url="https://example.com/manual-v2.pdf",
+    document_id=job.document_id,
+)
+```
+
+The API rejects concurrent non-terminal jobs for the same document with a
+retryable `ConflictError` using the server error code `ABORTED`.
+
+### Query retrieval results
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="How do I pair a Bluetooth headset?",
+    top_k=5,
+)
+
+for result in response.results:
+    print(result.content)
+    print(result.score)
+    print(result.source.document_id)
+    print(result.source.source_file_name)
+    print(result.source.section_path)
+```
+
+Retrieval results expose `content`, not the older parse-result `text` field.
+Media results may include `asset_url` when the server can sign the referenced
+artifact.
+
+Each retrieval result uses one canonical source reference shape:
+
+```python
+result.content
+result.chunk_type
+result.score
+result.asset_url  # Optional[str]
+result.source.document_id
+result.source.source_file_name
+result.source.section_path
+```
+
+### Exclude documents or sections
+
+Use exclusions for follow-up queries that should avoid already-used context.
+
+```python
+response = client.retrieval.query(
+    namespace="support-center",
+    query="battery charging",
+    top_k=10,
+    exclude_document_ids=["doc_old"],
+    exclude_sections=[
+        {"document_id": "doc_123", "section_path": "Appendix / Legal"}
+    ],
+)
+```
+
+### List, get, and archive documents
+
+```python
+document_list = client.documents.list(namespace="support-center")
+for document in document_list.documents:
+    print(document.document_id, document.status, document.source_file_name)
+
+document = client.documents.get("doc_123")
+print(document.current_job_result_id)
+
+archived = client.documents.archive("doc_123")
+print(archived.status)  # "archived"
+```
+
+---
+
 ## Async Usage
 
 Every method available on `Knowhere` has an async counterpart on `AsyncKnowhere`:
@@ -429,6 +549,13 @@ async def main():
         job_result = await client.jobs.wait(job.job_id)
         result = await client.jobs.load(job_result)
 
+        retrieval = await client.retrieval.query(
+            namespace="support-center",
+            query="refund policy",
+            top_k=5,
+        )
+        print(retrieval.results[0].content)
+
 asyncio.run(main())
 ```
 

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "knowhere-python-sdk"
-version = "0.2.1"
+version = "0.3.1"
 description = "Official Python SDK for the Knowhere document parsing API"
 readme = "README.md"
 license = "MIT"

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/src/knowhere/__init__.py

@@ -35,8 +35,17 @@ from knowhere._exceptions import (
 )
 from knowhere._types import PollProgressCallback, UploadProgressCallback
 from knowhere._version import __version__
+from knowhere.types.document import Document, DocumentListResponse
 from knowhere.types.job import Job, JobError, JobProgress, JobResult
 from knowhere.types.params import ParsingParams, WebhookConfig
+from knowhere.types.retrieval import (
+    RetrievalChannel,
+    RetrievalFilterMode,
+    RetrievalSectionExclusion,
+    RetrievalSource,
+    RetrievalQueryResponse,
+    RetrievalResult,
+)
 from knowhere.types.result import (
     BaseChunk,
     Checksum,
@@ -87,6 +96,16 @@ __all__: list[str] = [
     "JobError",
     "JobProgress",
     "JobResult",
+    # Document types
+    "Document",
+    "DocumentListResponse",
+    # Retrieval types
+    "RetrievalChannel",
+    "RetrievalFilterMode",
+    "RetrievalSectionExclusion",
+    "RetrievalSource",
+    "RetrievalQueryResponse",
+    "RetrievalResult",
     # Result types
     "ParseResult",
     "Manifest",

{knowhere_python_sdk-0.2.1 → knowhere_python_sdk-0.3.1}/src/knowhere/_client.py

@@ -19,7 +19,9 @@ from knowhere._types import (
     PollProgressCallback,
     UploadProgressCallback,
 )
+from knowhere.resources.documents import AsyncDocuments, Documents
 from knowhere.resources.jobs import AsyncJobs, Jobs
+from knowhere.resources.retrieval import AsyncRetrieval, Retrieval
 from knowhere.types.job import Job, JobResult
 from knowhere.types.params import ParsingParams, WebhookConfig
 from knowhere.types.result import ParseResult
@@ -42,6 +44,16 @@ class Knowhere(SyncAPIClient):
         """Access the jobs resource namespace."""
         return Jobs(self)
 
+    @cached_property
+    def retrieval(self) -> Retrieval:
+        """Access the retrieval resource namespace."""
+        return Retrieval(self)
+
+    @cached_property
+    def documents(self) -> Documents:
+        """Access the documents resource namespace."""
+        return Documents(self)
+
     # -- overloaded parse signatures --
 
     @overload
@@ -50,6 +62,8 @@ class Knowhere(SyncAPIClient):
         *,
         url: str,
         data_id: Optional[str] = ...,
+        namespace: Optional[str] = ...,
+        document_id: Optional[str] = ...,
         parsing_params: Optional[ParsingParams] = ...,
         webhook: Optional[WebhookConfig] = ...,
         poll_interval: float = ...,
@@ -66,6 +80,8 @@ class Knowhere(SyncAPIClient):
         file: Union[Path, BinaryIO, bytes],
         file_name: Optional[str] = ...,
         data_id: Optional[str] = ...,
+        namespace: Optional[str] = ...,
+        document_id: Optional[str] = ...,
         parsing_params: Optional[ParsingParams] = ...,
         webhook: Optional[WebhookConfig] = ...,
         poll_interval: float = ...,
@@ -82,6 +98,8 @@ class Knowhere(SyncAPIClient):
         file: Optional[Union[Path, BinaryIO, bytes]] = None,
         file_name: Optional[str] = None,
         data_id: Optional[str] = None,
+        namespace: Optional[str] = None,
+        document_id: Optional[str] = None,
         parsing_params: Optional[ParsingParams] = None,
         webhook: Optional[WebhookConfig] = None,
         poll_interval: float = DEFAULT_POLL_INTERVAL,
@@ -105,6 +123,8 @@ class Knowhere(SyncAPIClient):
                 source_type="url",
                 source_url=url,
                 data_id=data_id,
+                namespace=namespace,
+                document_id=document_id,
                 parsing_params=parsing_params,
                 webhook=webhook,
             )
@@ -116,6 +136,8 @@ class Knowhere(SyncAPIClient):
                 source_type="file",
                 file_name=resolved_name,
                 data_id=data_id,
+                namespace=namespace,
+                document_id=document_id,
                 parsing_params=parsing_params,
                 webhook=webhook,
             )
@@ -149,12 +171,24 @@ class AsyncKnowhere(AsyncAPIClient):
         """Access the async jobs resource namespace."""
         return AsyncJobs(self)
 
+    @cached_property
+    def retrieval(self) -> AsyncRetrieval:
+        """Access the async retrieval resource namespace."""
+        return AsyncRetrieval(self)
+
+    @cached_property
+    def documents(self) -> AsyncDocuments:
+        """Access the async documents resource namespace."""
+        return AsyncDocuments(self)
+
     @overload
     async def parse(
         self,
         *,
         url: str,
         data_id: Optional[str] = ...,
+        namespace: Optional[str] = ...,
+        document_id: Optional[str] = ...,
         parsing_params: Optional[ParsingParams] = ...,
         webhook: Optional[WebhookConfig] = ...,
         poll_interval: float = ...,
@@ -171,6 +205,8 @@ class AsyncKnowhere(AsyncAPIClient):
         file: Union[Path, BinaryIO, bytes],
         file_name: Optional[str] = ...,
         data_id: Optional[str] = ...,
+        namespace: Optional[str] = ...,
+        document_id: Optional[str] = ...,
         parsing_params: Optional[ParsingParams] = ...,
         webhook: Optional[WebhookConfig] = ...,
         poll_interval: float = ...,
@@ -187,6 +223,8 @@ class AsyncKnowhere(AsyncAPIClient):
         file: Optional[Union[Path, BinaryIO, bytes]] = None,
         file_name: Optional[str] = None,
         data_id: Optional[str] = None,
+        namespace: Optional[str] = None,
+        document_id: Optional[str] = None,
         parsing_params: Optional[ParsingParams] = None,
         webhook: Optional[WebhookConfig] = None,
         poll_interval: float = DEFAULT_POLL_INTERVAL,
@@ -206,6 +244,8 @@ class AsyncKnowhere(AsyncAPIClient):
                 source_type="url",
                 source_url=url,
                 data_id=data_id,
+                namespace=namespace,
+                document_id=document_id,
                 parsing_params=parsing_params,
                 webhook=webhook,
             )
@@ -217,6 +257,8 @@ class AsyncKnowhere(AsyncAPIClient):
                 source_type="file",
                 file_name=resolved_name,
                 data_id=data_id,
+                namespace=namespace,
+                document_id=document_id,
                 parsing_params=parsing_params,
                 webhook=webhook,
             )
@@ -232,4 +274,4 @@ class AsyncKnowhere(AsyncAPIClient):
 
         return await self.jobs.load(
             job_result, verify_checksum=verify_checksum
-        )
+        )

knowhere_python_sdk-0.3.1/src/knowhere/_version.py

@@ -0,0 +1 @@
+__version__ = "0.3.1"  # x-release-please-version

knowhere_python_sdk-0.3.1/src/knowhere/resources/__init__.py

@@ -0,0 +1,16 @@
+"""Resource namespace re-exports."""
+
+from __future__ import annotations
+
+from knowhere.resources.documents import AsyncDocuments, Documents
+from knowhere.resources.jobs import AsyncJobs, Jobs
+from knowhere.resources.retrieval import AsyncRetrieval, Retrieval
+
+__all__: list[str] = [
+    "AsyncDocuments",
+    "AsyncJobs",
+    "AsyncRetrieval",
+    "Documents",
+    "Jobs",
+    "Retrieval",
+]