volcengine-audio 0.2.2__tar.gz → 0.2.4__tar.gz

volcengine_audio-0.2.4/.agents/skills/volcengine-audio-doc-sync/SKILL.md

@@ -0,0 +1,105 @@
+---
+name: volcengine-audio-doc-sync
+description: Refresh and reconcile Volcengine Audio SDK documentation snapshots and SDK metadata. Use when working on doc_sync/volcengine snapshots, scripts/sync_volcengine_docs.py, Volcengine upstream doc timestamp changes, README sync dates, or schema/helper drift from Volcengine STT, TTS, or realtime dialogue docs in the volcengine-audio package.
+---
+
+# Volcengine Audio Doc Sync
+
+## Scope
+
+Use this skill for the standalone `volcengine-audio` SDK. Keep all guidance,
+commands, and file paths package-local.
+
+Primary package files:
+
+- `doc_sync/volcengine/manifest.json`
+- `doc_sync/volcengine/*.md`
+- `scripts/sync_volcengine_docs.py`
+- `src/volcengine_audio/`
+- `tests/`
+- `README.md`
+- `README.zh-CN.md`
+
+## Workflow
+
+1. Inspect the current tracked snapshots before running network sync:
+
+```bash
+git diff -- doc_sync/volcengine
+sed -n '1,220p' doc_sync/volcengine/manifest.json
+```
+
+2. If the task asks to refresh upstream docs, run the direct API sync:
+
+```bash
+python scripts/sync_volcengine_docs.py
+```
+
+3. Diff the refreshed snapshots and manifest. Focus on request fields, enum
+   values, event IDs, resource/model names, auth headers, and response payload
+   shapes.
+
+4. Review linked docs in `manifest.json`:
+
+- Check each entry's `linked_document_ids` and the top-level
+  `untracked_linked_document_ids`.
+- Add a linked Volcengine doc to `scripts/sync_volcengine_docs.py` when the
+  target page carries SDK-relevant enum values, field descriptions,
+  resource/model compatibility, event IDs, response payload shapes, or
+  speaker/language support tables.
+- Keep credential FAQ, console-operation, setup-only, and marketing links out
+  of `DOCS` unless their content becomes part of SDK request/response behavior.
+- If a tracked doc stops being linked or becomes obsolete, verify whether it is
+  still a source of SDK behavior before removing it from `DOCS`.
+- For non-Volcengine links that move SDK-critical information, record the link
+  in the review notes and decide whether a separate fetch mechanism is needed;
+  do not silently ignore it.
+
+5. Update SDK code only when the snapshot diff proves API drift:
+
+- `stt.py`: request fields, language/audio enums, STT response payloads,
+  context data, old/new console auth guidance, and STT error event codes.
+- `tts.py`: resource IDs, `req_params.model`, additions, subtitle/timestamp
+  payloads, and response events.
+- `realtime.py`: session config, dialog/asr/tts extra fields, control events,
+  context management requests, and acknowledgement payloads.
+- `protocol.py`: shared message types, event IDs, serialization, compression,
+  and framing helpers.
+- `__init__.py`: public exports for any new public schema or helper.
+
+6. Update package tests for every behavior change. Prefer exact payload
+   assertions over partial checks.
+
+7. Update README files only with user-facing facts: install/development usage,
+   public examples, API reference, package version, and the local sync date or
+   source timestamps when they are meant to be public. Keep procedural sync
+   checklists in this skill, not in README files.
+
+## Snapshot Rules
+
+- Store only cleaned upstream `Result.Content` text in tracked `.md` snapshot
+  files.
+- Do not store full JSON responses in tracked snapshots.
+- Strip `<span ...>` and `</span>` tags before writing snapshots.
+- Keep source metadata such as `updated_time`, `source_url`, `api_url`, and
+  content hashes in `manifest.json`.
+- Keep `linked_document_ids` and `untracked_linked_document_ids` in
+  `manifest.json` so new, obsolete, or newly important links are visible in
+  future diffs.
+- Snapshot files are tracked for future diffs, but they are not included in the
+  wheel because the build packages only `src/volcengine_audio`.
+
+## Validation
+
+Run focused checks after changing code:
+
+```bash
+uv run pytest tests
+uv run ruff check src tests
+```
+
+For docs-only changes, at least run:
+
+```bash
+git diff --check
+```

volcengine_audio-0.2.4/.agents/skills/volcengine-audio-doc-sync/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "Volcengine Audio Doc Sync"
+  short_description: "Sync Volcengine audio SDK docs"
+  default_prompt: "Use $volcengine-audio-doc-sync to refresh Volcengine Audio SDK documentation snapshots and sync README metadata."

{volcengine_audio-0.2.2 → volcengine_audio-0.2.4}/AGENTS.md

@@ -1,7 +1,7 @@
 # AGENTS.md
 
-Guidance for coding agents working in
-`/Users/liangdeo/pyproj/dispatcher/packages/volcengine-audio`.
+Guidance for coding agents working in the `volcengine-audio` package
+repository.
 
 ## 1) What This Package Is
 
@@ -14,10 +14,9 @@ services. It provides typed schemas and request/response helpers for:
 - Shared websocket/binary protocol helpers
 
 Keep this guide in sync with `README.md`. The README is the user-facing
-package guide and records the local SDK/doc sync date, tracked upstream
-timestamps, and the latest upstream review summary; this file should carry the
-agent-facing workflow and code-boundary details that make those README claims
-actionable.
+package guide and records the local SDK/doc sync date and tracked upstream
+timestamps; this file carries the agent-facing workflow and code-boundary
+details that make those README claims actionable.
 
 Core maintenance entrypoints:
 
@@ -30,7 +29,7 @@ Core maintenance entrypoints:
 
 ## 2) Environment and Dependencies
 
-- Workspace runtime: Python 3.13
+- Local development runtime: Python 3.13
 - Package metadata: `requires-python = ">=3.10"`
 - Package manager: `uv`
 - Build backend: `hatchling`
@@ -39,11 +38,9 @@ Core maintenance entrypoints:
 Useful setup commands:
 
 ```bash
-# from package root
 uv sync --frozen --group dev
-
-# from repo root
-uv run pytest packages/volcengine-audio/tests
+uv run pytest tests
+uv run ruff check src tests
 ```
 
 ## 3) Source of Truth
@@ -57,6 +54,7 @@ first stop during maintenance:
 - `doc_sync/volcengine/1719100-tts_websocket_unidirectional_v3.md`
 - `doc_sync/volcengine/1598757-tts_http_chunked_sse_v3.md`
 - `doc_sync/volcengine/1354869-stt_streaming_bigmodel.md`
+- `doc_sync/volcengine/1257544-tts_voice_list.md`
 
 These files are tracked in git so future syncs can diff upstream changes
 quickly. They are repo-only maintenance artifacts and are not packed into
@@ -64,24 +62,30 @@ wheels because the wheel build only includes `src/volcengine_audio`.
 
 Current tracked upstream timestamps from `manifest.json` and `README.md`:
 
-- Realtime dialogue: `2026-04-29T07:34:45Z`
-- TTS WebSocket bidirectional V3: `2026-04-24T03:45:24Z`
-- TTS WebSocket unidirectional V3: `2026-04-24T03:44:56Z`
-- TTS HTTP Chunked/SSE V3: `2026-04-24T03:44:50Z`
-- STT streaming bigmodel: `2026-04-27T03:14:54Z`
+- Realtime dialogue: `2026-06-04T10:15:01Z`
+- TTS WebSocket bidirectional V3: `2026-05-25T08:51:30Z`
+- TTS WebSocket unidirectional V3: `2026-05-25T08:49:18Z`
+- TTS HTTP Chunked/SSE V3: `2026-05-25T09:03:36Z`
+- STT streaming bigmodel: `2026-05-29T02:49:48Z`
+- TTS voice list: `2026-05-26T05:41:00Z`
 
 Latest local sync review:
 
-- Local sync date in README: `2026-05-01`.
-- Realtime docs added concrete protocol surface that belongs in the SDK:
-  `UpdateConfig`, `EndASR`, `ConversationTruncate`, `ClientInterrupt`,
-  `ConfigUpdated`, `ConversationTruncated`, realtime `tts.extra`
-  `explicit_dialect`, and realtime `tts.extra.aigc_metadata`.
-- STT docs now separate old-console app/access-key headers from new-console
-  `X-Api-Key`, while the request schema still uses `context_data` for dialog
-  context and `EventReceive.WAITING_NEXT_PACKET_TIMEOUT` for `45000081`.
-- TTS docs only changed the `seed-tts-1.0-concurr` billing wording in this
-  refresh; the existing resource/model schema remains valid.
+- Local sync date in README: `2026-06-06`.
+- Realtime docs added 12K 2.0 context notes, O/SC convergence wording,
+  updated SC2.0 voice-list references, `tts.extra.tts_2.0_model`, and a
+  10w TPM default. The SDK exposes this as `tts_2_0_model` with the wire alias
+  `tts_2.0_model`.
+- STT docs added non-streaming `enable_auto_lang` for automatic language
+  detection across 25 locales. The SDK request schema includes
+  `enable_auto_lang`.
+- TTS docs now describe the TTS 2.0 default model as
+  `seed-tts-2.0-standard`, expand ICL2.0 explicit-language guidance, and
+  clarify `context_texts`, `section_id`, and `use_tag_parser`. The existing SDK
+  fields cover those parameters.
+- TTS voice-list doc `1257544` is now tracked because the API docs delegate
+  speaker, model/resource compatibility, voice ability, and language support
+  details to that page. The SDK still keeps speaker fields open as `str`.
 
 Snapshot format guidance:
 
@@ -91,21 +95,21 @@ Snapshot format guidance:
   diffs stay readable.
 - Keep metadata such as `updated_time`, `source_url`, `api_url`, and the
   content hash in `manifest.json`.
+- Review `linked_document_ids` and `untracked_linked_document_ids` in
+  `manifest.json` after each sync. Add linked Volcengine docs to
+  `scripts/sync_volcengine_docs.py` when they carry SDK-relevant enum values,
+  field descriptions, resource/model compatibility, event IDs, or response
+  shapes. Do not add credential FAQ, console operation, or setup-only links
+  unless their content becomes part of SDK behavior.
 
 ## 4) Quality and Static Checks
 
 Run what is relevant for the change:
 
 ```bash
-# from package root
 uv run ruff check src tests
 uv run ruff format src tests
 uv run pytest
-
-# from repo root
-uv run ruff check packages/volcengine-audio/src packages/volcengine-audio/tests
-uv run ruff format packages/volcengine-audio/src packages/volcengine-audio/tests
-uv run pytest packages/volcengine-audio/tests
 ```
 
 ## 5) Testing
@@ -116,13 +120,13 @@ Useful commands:
 
 ```bash
 # full package suite
-uv run pytest packages/volcengine-audio/tests
+uv run pytest tests
 
 # targeted file
-uv run pytest packages/volcengine-audio/tests/test_realtime.py
+uv run pytest tests/test_realtime.py
 
 # targeted test selection
-uv run pytest packages/volcengine-audio/tests -k tts
+uv run pytest tests -k tts
 ```
 
 Testing expectations:
@@ -132,8 +136,6 @@ Testing expectations:
 - Prefer exact payload assertions over loose partial checks.
 - Cover both request generation and response parsing when protocol behavior
   changes.
-- Do not use the dispatcher Docker test wrapper for this package unless the
-  change truly depends on repo-level integration.
 
 ## 6) Repo Map
 
@@ -144,7 +146,7 @@ Testing expectations:
 - `src/volcengine_audio/__init__.py`: public exports
 - `tests/`: package-level regression coverage
 - `doc_sync/volcengine/`: tracked upstream snapshots and manifest
-- `scripts/sync_volcengine_docs.py`: Playwright-based doc sync utility
+- `scripts/sync_volcengine_docs.py`: direct API doc sync utility
 - `README.md`: English package guide and sync metadata
 - `README.zh-CN.md`: Chinese package guide and sync metadata
 
@@ -153,30 +155,29 @@ Testing expectations:
 Refresh docs with:
 
 ```bash
-uvx --with playwright python packages/volcengine-audio/scripts/sync_volcengine_docs.py
+python scripts/sync_volcengine_docs.py
 ```
 
 What the script does:
 
-1. Opens each public Volcengine docs page with Playwright.
-2. Captures the backing `api/doc/getDocDetail` JSON response.
-3. Extracts only `Result.Content`, removes span tags, and writes a `.md`
+1. Requests each public `api/doc/getDocDetail` JSON response directly.
+2. Extracts only `Result.Content`, removes span tags, and writes a `.md`
    snapshot for each tracked doc.
-4. Writes `manifest.json` with source metadata and content hashes.
+3. Writes `manifest.json` with source metadata, content hashes, and linked
+   Volcengine doc IDs.
 
 Important notes:
 
-- The public docs pages are JS-rendered.
-- Direct CLI requests to the JSON endpoint may return unauthorized.
-- Prefer Playwright response interception over raw `curl` scraping.
+- Fetch all docs successfully before clearing old snapshots so a failed sync
+  does not leave the snapshot directory empty.
 - If the task is only to sync `AGENTS.md` or `README.md` with already-tracked
-  snapshots, do not rerun Playwright. Read `manifest.json`, the README files,
-  and `git diff -- doc_sync/volcengine` instead.
+  snapshots, do not rerun the sync script. Read `manifest.json`, the README
+  files, and `git diff -- doc_sync/volcengine` instead.
 
 ## 8) Coding Standards (Package-Specific)
 
-- Keep the package standalone; do not import dispatcher app code into
-  `src/volcengine_audio`.
+- Keep the package standalone; do not import code from outside this package
+  into `src/volcengine_audio`.
 - Use `orjson` for JSON operations in package code when serialization is
   needed.
 - Prefer modern typing syntax such as `list[str]` and `str | None`.
@@ -240,9 +241,9 @@ Formatting from repo config:
 Useful commands:
 
 ```bash
-git diff -- packages/volcengine-audio/doc_sync/volcengine
-git diff -- packages/volcengine-audio/src/volcengine_audio
-rg -n "keep_alive|push_to_talk|concurr|UpdatedTime" packages/volcengine-audio
+git diff -- doc_sync/volcengine
+git diff -- src/volcengine_audio
+rg -n "keep_alive|push_to_talk|concurr|UpdatedTime" .
 ```
 
 High-signal fields to watch:
@@ -260,8 +261,8 @@ High-signal fields to watch:
 Run at least:
 
 ```bash
-uv run pytest packages/volcengine-audio/tests
-uv run ruff check packages/volcengine-audio/src packages/volcengine-audio/tests
+uv run pytest tests
+uv run ruff check src tests
 ```
 
 Also confirm:
@@ -277,8 +278,8 @@ Before finishing code changes, run what is relevant:
 1. Refresh doc snapshots if the task is an upstream sync.
 2. Run targeted package tests via `uv run pytest ...`.
 3. Run `uv run ruff check ...`.
-4. Confirm no dispatcher-only dependency leaked into this standalone package.
-5. Confirm no secrets or temporary scrape artifacts were added to tracked files.
-6. Add docstrings and type annotations for new best-effort public code.
-7. Update nearby docs in the same change if behavior or maintenance workflow
+
+4. Confirm no secrets or temporary scrape artifacts were added to tracked files.
+5. Add docstrings and type annotations for new best-effort public code.
+6. Update nearby docs in the same change if behavior or maintenance workflow
    changed materially.

{volcengine_audio-0.2.2 → volcengine_audio-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: volcengine-audio
-Version: 0.2.2
+Version: 0.2.4
 Summary: Python SDK for Volcengine Audio Services (TTS, STT, and Realtime Dialogue)
 Project-URL: Homepage, https://github.com/aiyou178/volcengine-audio
 Project-URL: Issues, https://github.com/aiyou178/volcengine-audio/issues
@@ -22,7 +22,7 @@ Requires-Python: >=3.10
 Requires-Dist: orjson
 Requires-Dist: pydantic<3.0,>=2.0
 Provides-Extra: dev
-Requires-Dist: pytest-asyncio>=1.0.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=1.4.0; extra == 'dev'
 Requires-Dist: pytest-randomly>=3.16.0; extra == 'dev'
 Requires-Dist: pytest>=8.0; extra == 'dev'
 Requires-Dist: ruff>=0.11.0; extra == 'dev'
@@ -42,41 +42,26 @@ Python SDK for Volcengine (ByteDance) Audio Services, providing comprehensive su
 - **Protocol Support**: Low-level protocol utilities for custom implementations
 - **Type Safety**: Full Pydantic model validation for all requests and responses
 
-## Documentation Sync
+## Documentation
 
-### Last SDK/doc sync
+Last SDK/doc sync: `2026-06-06`.
 
-* Local sync date: `2026-05-01`
-* Package maintenance guide: [`AGENTS.md`](AGENTS.md)
-* Snapshot manifest: [`doc_sync/volcengine/manifest.json`](doc_sync/volcengine/manifest.json)
-* Refresh command: `uvx --with playwright python packages/volcengine-audio/scripts/sync_volcengine_docs.py`
-* These Volcengine docs are JS-rendered. The sync script opens the public docs pages with Playwright, captures the backing `api/doc/getDocDetail` JSON response, and writes cleaned `Result.Content` markdown snapshots to `doc_sync/volcengine/`.
-* The tracked snapshot files store only doc content text with span tags removed, while source metadata stays in `manifest.json`.
-* The snapshot files are tracked in git for future diffs, but they are not packed into wheels because this package only ships `src/volcengine_audio`.
-* Latest upstream review: May 2026 doc updates changed realtime dialogue protocol coverage and auth/documentation wording. The SDK now includes the refreshed realtime control events, conversation truncation helpers, realtime `tts.extra` metadata, STT `X-Api-Key` guidance in snapshots, and TTS billing wording updates.
+### Current Tracked Sources
 
-### Tracked upstream sources
-
-* Realtime dialogue: `2026-04-29T07:34:45Z` - <https://www.volcengine.com/docs/6561/1594356?lang=zh>
-* TTS WebSocket bidirectional V3: `2026-04-24T03:45:24Z` - <https://www.volcengine.com/docs/6561/1329505?lang=zh>
-* TTS WebSocket unidirectional V3: `2026-04-24T03:44:56Z` - <https://www.volcengine.com/docs/6561/1719100?lang=zh>
-* TTS HTTP Chunked/SSE V3: `2026-04-24T03:44:50Z` - <https://www.volcengine.com/docs/6561/1598757?lang=zh>
-* STT streaming bigmodel: `2026-04-27T03:14:54Z` - <https://www.volcengine.com/docs/6561/1354869?lang=zh>
-
-### Sync checklist
-
-1. Refresh the tracked content snapshots with `uvx --with playwright python packages/volcengine-audio/scripts/sync_volcengine_docs.py`.
-2. Diff `doc_sync/volcengine/*.md` to see which request fields, enums, events, or examples changed upstream.
-3. Update `src/volcengine_audio/` schemas and helper functions as needed.
-4. Update or add tests under `tests/`.
-5. Update the local sync date in this README and in [`README.zh-CN.md`](README.zh-CN.md).
+- Realtime dialogue: `2026-06-04T10:15:01Z` - <https://www.volcengine.com/docs/6561/1594356?lang=zh>
+- TTS WebSocket bidirectional V3: `2026-05-25T08:51:30Z` - <https://www.volcengine.com/docs/6561/1329505?lang=zh>
+- TTS WebSocket unidirectional V3: `2026-05-25T08:49:18Z` - <https://www.volcengine.com/docs/6561/1719100?lang=zh>
+- TTS HTTP Chunked/SSE V3: `2026-05-25T09:03:36Z` - <https://www.volcengine.com/docs/6561/1598757?lang=zh>
+- STT streaming bigmodel: `2026-05-29T02:49:48Z` - <https://www.volcengine.com/docs/6561/1354869?lang=zh>
+- TTS voice list: `2026-05-26T05:41:00Z` - <https://www.volcengine.com/docs/6561/1257544?lang=zh>
 
 ## Installation
 
+Requires Python 3.10 or newer.
+
 ### Install from PyPI
 
 ```bash
-# From PyPI (when published)
 pip install volcengine-audio
 ```
 
@@ -88,6 +73,19 @@ cd volcengine-audio
 pip install -e .
 ```
 
+## Development
+
+```bash
+# from this repository root
+uv sync --frozen --group dev
+uv run pytest tests
+uv run ruff check src tests
+uv run ruff format src tests
+```
+
+This package is a standalone SDK. Keep source under `src/volcengine_audio` and
+tests under `tests`.
+
 ## Quick Start
 
 ### Speech-to-Text (STT)
@@ -228,13 +226,21 @@ finish = RealtimeDialogueFunctions.finish_session_payload("session-123")
 
 Core protocol definitions and utilities.
 
+**Package Metadata:**
+- `__version__`: Installed package version
+
 **Classes:**
 - `ProtocolVersion`: Protocol version enumeration (V1)
+- `HeaderSize`: Protocol header size enumeration
 - `MessageType`: Message types for bidirectional communication
+- `MessageTypeSpecificFlag`: Message flags for sequencing and event framing
+- `AsrMessageType`: ASR-specific message types
+- `AsrMessageTypeSpecificFlag`: ASR-specific message flags
 - `EventSend`: Events sent from client to server
 - `EventReceive`: Events received from server
 - `SerializationMethod`: Payload serialization methods (JSON, RAW, PROTOBUF)
 - `CompressionMethod`: Payload compression methods (NONE, GZIP)
+- `AudioCodec`: Audio codec values used by STT request schemas
 
 **Constants:**
 - `HOST`: `'openspeech.bytedance.com'` - Volcengine audio service host
@@ -249,6 +255,7 @@ Speech-to-Text (ASR) models and utilities.
 
 **Request Models:**
 - `VolcengineAsrRequestV3`: ASR V3 API request
+  - `request.enable_auto_lang`: Auto-detect language for non-streaming STT
 - `VolcengineAsrRequestV2`: ASR V2 API request
 
 **Response Models:**
@@ -258,13 +265,17 @@ Speech-to-Text (ASR) models and utilities.
 **Enums:**
 - `STTResource`: STT resource types for billing
 - `STTAudioFormatV3`: Audio formats (pcm, wav, mp3, ogg)
+- `AudioFormatV2`: Audio formats for the V2 request schema
 - `STTResultType`: Result types (full, single)
 - `STTBigmodelNoStreamLanguage`: Supported languages for bigmodel
 
 **Helper Classes:**
 - `VolcengineAsrFunctionsV3`: V3 API helper functions
+  - `generate_asr_header()`: Generate V3 ASR request headers
+  - `generate_asr_before_payload()`: Generate V3 ASR sequence metadata
   - `generate_asr_full_client_request()`: Generate full client request
   - `generate_asr_audio_only_request()`: Generate audio-only request
+  - `parse_request()`: Parse generated request bytes for inspection
   - `parse_response()`: Parse server response
 - `VolcengineAsrFunctionsV2`: V2 API helper functions
   - `full_client_request()`: Generate full client request
@@ -283,19 +294,35 @@ Text-to-Speech models and utilities.
 - `TTSSentenceStartResponse`: Sentence start notification
 - `TTSSentenceEndResponse`: Sentence end notification
 - `TTSEndResponse`: TTS ended notification
+- `TTSSentenceEndPayload`: Sentence-end payload typed dict
+- `TTSSubtitlePayload`: Subtitle payload typed dict
+- `TTSTimedWord`: Timed word typed dict
 
 **Enums:**
 - `TTSBigmodelResourceType`: TTS resource IDs (`seed-tts-1.0`, `seed-tts-2.0`, etc.)
 - `TTSBigmodelModelType`: Optional `req_params.model` values (`seed-tts-1.1`, `seed-tts-2.0-standard`, etc.)
 - `TTSAudioFormat`: Audio formats (wav, pcm, mp3, ogg_opus)
+- `OperationEnum`: HTTP TTS operation values
+
+**Configuration Models:**
+- `AppConfig`: HTTP TTS app credentials and cluster
+- `UserConfig`: User identifier for request metadata
+- `AudioConfig`: HTTP TTS audio options
+- `RequestConfig`: HTTP TTS request options
 
 **Helper Classes:**
 - `VolcengineTTSFunctions`: TTS API helper functions
+  - `prepare_request()`: Prepare HTTP TTS request payload
+  - `task_request_payload()`: Generate bidirectional task payload
   - `start_connection_payload()`: Start connection
   - `start_session_payload()`: Start TTS session
+  - `cancel_session_payload()`: Cancel TTS session
   - `finish_session_payload()`: Finish TTS session
+  - `finish_connection_payload()`: Finish connection
   - `extract_response_payload()`: Extract and parse response
   - `calculate_payload()`: Calculate request payload
+- `validate_tts_resource_model_mapping()`: Validate compatible resource/model
+  pairs for TTS 1.x and 2.x resources
 
 #### `volcengine_audio.realtime`
 
@@ -304,7 +331,8 @@ Realtime dialogue (combined TTS+STT) models and utilities.
 **Configuration:**
 - `RealtimeDialogueConfig`: Complete dialogue session configuration
   - `DialogConfig`: Bot persona, speaking style, location
-  - `TTSConfig`: Voice type and audio settings
+  - `TTSConfig`: Voice type, audio settings, and `tts_2_0_model` wire alias
+    `tts_2.0_model`
   - `Asr`: ASR-specific settings
 
 **Request Models:**
@@ -323,12 +351,16 @@ Realtime dialogue (combined TTS+STT) models and utilities.
 - `ChatTextQueryConfirmedResponse`: Text query acknowledgement
 - `ConversationCreatedResponse`, `ConversationUpdatedResponse`, `ConversationRetrievedResponse`, `ConversationTruncatedResponse`, `ConversationDeletedResponse`: Context management acknowledgements
 - `ConfigUpdatedResponse`: Runtime config update acknowledgement
+- `ConnectionFailedResponse`: Connection-level failure payload
 - `SessionStartedResponse`: Session started
 - `SessionFailedResponse`: Session failed
+- `RealtimeDialogueErrorResponse`: Generic realtime error payload
+- `RealtimeDialogueUsage`: Usage typed dict
 
 **Helper Classes:**
 - `RealtimeDialogueFunctions`: Realtime dialogue API helpers
   - `start_connection_payload()`: Start connection
+  - `finish_connection_payload()`: Finish connection
   - `start_session_payload()`: Start dialogue session
   - `task_request_payload()`: Send audio for recognition
   - `update_config_payload()`: Update runtime TTS/dialog config
@@ -538,9 +570,11 @@ config = RealtimeDialogueConfig(
 ## Error Handling
 
 ```python
-from volcengine_audio import EventReceive
+from volcengine_audio import EventReceive, VolcengineTTSFunctions
 
-event, session_id, payload = VolcengineTTSFunctions.extract_response_payload(response)
+event, session_id, payload = VolcengineTTSFunctions.extract_response_payload(
+    response
+)
 
 if event == EventReceive.SessionFailed:
     print(f"Session failed: {payload.get('error')}")
@@ -550,23 +584,6 @@ elif event == EventReceive.SERVER_PROCESSING_ERROR:
     print("Server processing error")
 ```
 
-## Development
-
-### Running Tests
-
-```bash
-pytest tests/
-```
-
-### Code Style
-
-This package uses Ruff for linting and formatting:
-
-```bash
-ruff check src/ tests/
-ruff format src/ tests/
-```
-
 ## License
 
 MIT