* fix: only pass dimensions param when explicitly configured
Models like bge-m3 don't support the dimensions parameter in the
embedding API, causing HTTP 400 errors. Previously dimensions was
always sent with a default value of 1024, even when the user never
configured it. Now dimensions is only included in the request when
embedding_dimensions is explicitly set in provider config.
Closes#6421
Signed-off-by: JiangNan <1394485448@qq.com>
* fix: handle invalid dimensions config and align get_dim return
- Add try-except around int() conversion in _embedding_kwargs to
gracefully handle invalid embedding_dimensions config values
- Update get_dim() to return 0 when embedding_dimensions is not
explicitly configured, so callers know dimensions weren't specified
and can handle it accordingly
- Both methods now share consistent logic for reading the config
Signed-off-by: JiangNan <1394485448@qq.com>
* fix: improve logging for invalid embedding_dimensions configuration
---------
Signed-off-by: JiangNan <1394485448@qq.com>
Co-authored-by: Soulter <905617992@qq.com>
* Fix CreateSkillPayloadTool array schema missing items field
The payload parameter's anyOf array variant lacked an items field,
causing Gemini API to reject the tool declaration with 400 Bad Request:
'parameters.properties[payload].any_of[1].items: missing field.'
Add items: {type: object} to the array variant to satisfy the Gemini
API requirement for array type schemas.
Fixes#6279
* Fix TypeError when OpenAI-compatible API returns null choices
Some providers (e.g. OpenRouter) may return a completion where
choices is None rather than an empty list — for instance on rate
limiting, content filtering, or transient errors. The existing code
used len(completion.choices) which throws TypeError on None.
Replace all len(...choices) == 0 checks with 'not ... .choices' which
handles both None and empty list. Affects _query_stream, _parse_openai_completion,
and _extract_reasoning_content.
Fixes#6252
In _handle_api_error(), when a 429 rate-limit is encountered, the code
calls available_api_keys.remove(chosen_key). If the same key was already
removed in a previous retry iteration (e.g. the key rotated back to the
same value), this raises ValueError which crashes the entire LLM request
with an opaque error instead of a proper retry/fallback.
Add a membership check before calling remove() to prevent the crash.
Co-authored-by: easonysliu <easonysliu@tencent.com>
- Add audio data validation in MiniMax TTS get_audio() method to detect empty responses
- Validate generated audio file size in TTSProvider.test() to ensure valid output
- Provide detailed error messages guiding users to check group_id configuration
- Auto-cleanup test audio files after validation
- Fixes issue where 0KB audio files would pass TTS detection when group_id is not configured
* fix(provider): handle MiniMax ThinkingBlock when max_tokens reached
Fixes#5912
Problem: MiniMax API returns ThinkingBlock when stop_reason='max_tokens',
but AstrBot throws 'completion 无法解析' exception because both
completion_text and tools_call_args are empty.
Root cause: The validation logic didn't consider ThinkingBlock
(reasoning_content) as valid content.
Fix: When completion_text and tools_call_args are empty but
reasoning_content is present, treat it as valid instead of throwing
exception. This happens when the model thinks but runs out of tokens
before generating the actual response.
Impact: MiniMax models now work correctly when responses are truncated
due to max_tokens limit.
* refactor: address review feedback
1. Use getattr for safe stop_reason access (prevent AttributeError)
2. Use ValueError instead of generic Exception for better error handling
Thanks @gemini-code-assist and @sourcery-ai for the review!
* refactor: flatten nested if/else with guard clause
Address Gemini Code Assist feedback:
- Use guard clause for early return
- Flattened nested conditional for better readability
Logic unchanged, just cleaner code structure.
* fix(provider): improve logging for ThinkingBlock completions in ProviderAnthropic
---------
Co-authored-by: Soulter <905617992@qq.com>
* fix(dashboard): 强化 API Key 复制临时节点清理逻辑
* fix(embedding): 自动检测改为探测 OpenAI embedding 最大可用维度
* fix: normalize openai embedding base url and add hint key
* i18n: add embedding_api_base hint translations
* i18n: localize provider embedding/proxy metadata hints
* fix: show provider-specific embedding API Base URL hint as field subtitle
* fix(embedding): cap OpenAI detect_dim probes with early short-circuit
* fix(dashboard): return generic error on provider adapter import failure
* 回退检测逻辑
* feat: add Anthropic Claude Code OAuth provider and adaptive thinking support
* fix: add defensive guard for metadata overrides and align budget condition with docs
* refactor: adopt sourcery-ai suggestions for OAuth provider
- Use use_api_key=False in OAuth subclass to avoid redundant
API-key client construction before replacing with auth_token client
- Generalize metadata override helper to merge all dict keys
instead of only handling 'limit', improving extensibility
* ci: remove Electron desktop build from release pipeline
* chore: remove electron desktop and switch to tauri release trigger
* ci: remove desktop workflow dispatch trigger
* refactor: migrate data paths to astrbot_path helpers
* fix: point desktop update prompt to AstrBot-desktop releases
* fix: handle list format content from OpenAI-compatible APIs
Some LLM providers (e.g., GLM-4.5V via SiliconFlow) return content as
list[dict] format like [{'type': 'text', 'text': '...'}] instead of
plain string. This causes the raw list representation to be displayed
to users.
Changes:
- Add _normalize_content() helper to extract text from various content formats
- Use json.loads instead of ast.literal_eval for safer parsing
- Add size limit check (8KB) before attempting JSON parsing
- Only convert lists that match OpenAI content-part schema (has 'type': 'text')
to avoid collapsing legitimate list-literal replies like ['foo', 'bar']
- Add strip parameter to preserve whitespace in streaming chunks
- Clean up orphan </think> tags that may leak from some models
Fixes#5124
* fix: improve content normalization safety
- Try json.loads first, fallback to ast.literal_eval for single-quoted
Python literals to avoid corrupting apostrophes (e.g., "don't")
- Coerce text values to str to handle null or non-string text fields
* feat: temporary file handling and introduce TempDirCleaner
- Updated various modules to use `get_astrbot_temp_path()` instead of `get_astrbot_data_path()` for temporary file storage.
- Renamed temporary files for better identification and organization.
- Introduced `TempDirCleaner` to manage the size of the temporary directory, ensuring it does not exceed a specified limit by deleting the oldest files.
- Added configuration option for maximum temporary directory size in the dashboard.
- Implemented tests for `TempDirCleaner` to verify cleanup functionality and size management.
* ruff
* 修复gemini toolcall 的名称导致的循环调用
* Apply suggestions from code review
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* Refactor function response creation for tool role
Refactor function response handling for tool role to ensure proper ID injection.
---------
Co-authored-by: Soulter <37870767+Soulter@users.noreply.github.com>
Co-authored-by: sourcery-ai[bot] <58596630+sourcery-ai[bot]@users.noreply.github.com>
* feat: astr live
* chore: remove
* feat: metrics
* feat: enhance audio processing and metrics display in live mode
* feat: genie tts
* feat: enhance live mode audio processing and text handling
* feat: add metrics
* feat: eyes
* feat: nervous
* chore: update readme
Added '自动压缩对话' feature and updated features list.
* feat: skip saving head system messages in history (#4538)
* feat: skip saving the first system message in history
* fix: rename variable for clarity in system message handling
* fix: update logic to skip all system messages until the first non-system message
* fix: clarify logic for skipping initial system messages in conversation
* chore: bump version to 4.12.2
* docs: update 4.12.2 changelog
* refactor: update event types for LLM tool usage and response
* chore: bump version to 4.12.3
* fix: ensure embedding dimensions are returned as integers in providers (#4547)
* fix: ensure embedding dimensions are returned as integers in providers
* chore: ruff format
* perf: T2I template editor preview (#4574)
* feat: add file drag upload feature for ChatUI (#4583)
* feat(chat): add drag-drop upload and fix batch file upload
* style(chat): adjust drop overlay to only cover input container
* fix: streaming response for DingTalk (#4590)
closes: #4384
* #4384 钉钉消息回复卡片模板
* chore: ruff format
* chore: ruff format
---------
Co-authored-by: ManJiang <man.jiang@jg-robust.com>
Co-authored-by: Soulter <905617992@qq.com>
* feat: implement persona folder for advanced persona management (#4443)
* feat(db): add persona folder management for hierarchical organization
Implement hierarchical folder structure for organizing personas:
- Add PersonaFolder model with recursive parent-child relationships
- Add folder_id and sort_order fields to Persona model
- Implement CRUD operations for persona folders in database layer
- Add migration support for existing databases
- Extend PersonaManager with folder management methods
- Add dashboard API routes for folder operations
* feat(persona): add batch sort order update endpoint for personas and folders
Add new API endpoint POST /persona/reorder to batch update sort_order
for both personas and folders. This enables drag-and-drop reordering
in the dashboard UI.
Changes:
- Add abstract batch_update_sort_order method to BaseDatabase
- Implement batch_update_sort_order in SQLiteDatabase
- Add batch_update_sort_order to PersonaManager with cache refresh
- Add reorder_items route handler with input validation
* feat(persona): add folder_id and sort_order params to persona creation
Extend persona creation flow to support folder placement and ordering:
- Add folder_id and sort_order parameters to insert_persona in db layer
- Update PersonaManager.create_persona to accept and pass folder params
- Add get_folder_detail API endpoint for retrieving folder information
- Include folder_id and sort_order in persona creation response
- Add session flush/refresh to return complete persona object
* feat(dashboard): implement persona folder management UI
- Add folder management system with tree view and breadcrumbs
- Implement create, rename, delete, and move operations for folders
- Add drag-and-drop support for organizing personas and folders
- Create new PersonaManager component and Pinia store for state management
- Refactor PersonaPage to support hierarchical structure
- Update locale files with folder-related translations
- Handle empty parent_id correctly in backend route
* feat(dashboard): centralize folder expansion state in persona store
Move folder expansion logic from local component state to global Pinia
store to persist expansion state.
- Add `expandedFolderIds` state and toggle actions to `personaStore`
- Update `FolderTreeNode` to use store state instead of local data
- Automatically navigate to target folder after moving a persona
* feat(dashboard): add reusable folder management component library
Extract folder management UI into reusable base components and create
persona-specific wrapper components that integrate with personaStore.
- Add base folder components (tree, breadcrumb, card, dialogs) with
customizable labels for i18n support
- Create useFolderManager composable for folder state management
- Implement drag-and-drop support for moving personas between folders
- Add persona-specific wrapper components connecting to personaStore
- Reorganize PersonaManager into views/persona directory structure
- Include comprehensive README documentation for component usage
* refactor(dashboard): remove legacy persona folder management components
Remove deprecated persona folder management Vue components that have been
superseded by the new reusable folder management component library.
Deleted components:
- CreateFolderDialog.vue
- FolderBreadcrumb.vue
- FolderCard.vue
- FolderTree.vue
- FolderTreeNode.vue
- MoveTargetNode.vue
- MoveToFolderDialog.vue
- PersonaCard.vue
- PersonaManager.vue
These components are replaced by the centralized folder management
implementation introduced in commit 3fbb3db2.
* fix(dashboard): add delayed skeleton loading to prevent UI flicker
Implement a 150ms delay before showing the skeleton loader in
PersonaManager to prevent visual flicker during fast loading operations.
- Add showSkeleton state with timer-based delay mechanism
- Use v-fade-transition for smooth skeleton visibility transitions
- Clean up timer on component unmount to prevent memory leaks
- Only display skeleton when loading exceeds threshold duration
* feat(dashboard): add generic folder item selector component for persona selection
Introduce BaseFolderItemSelector.vue as a reusable component for selecting
items within folder hierarchies. Refactor PersonaSelector to use this new
base component instead of its previous flat list implementation.
Changes:
- Add BaseFolderItemSelector with folder tree navigation and item selection
- Extend folder types with SelectableItem and FolderItemSelectorLabels
- Refactor PersonaSelector to leverage the new base component
- Add i18n translations for rootFolder and emptyFolder labels
* feat(persona): add tree-view display for persona list command
Add hierarchical folder tree output for the persona list command,
showing personas organized by folders with visual tree connectors.
- Add _build_tree_output method for recursive tree structure rendering
- Display folders with 📁 icon and personas with 👤 icon
- Show root-level personas separately from folder contents
- Include total persona count in output
* refactor(persona): simplify tree-view output with shorter indentation lines
Replace complex tree connector logic with simpler depth-based indentation
using "│ " prefix. Remove unnecessary parameters (prefix, is_last) and
computed variables (has_content, total_items, item_idx) in favor of a
cleaner depth-based approach.
* feat(dashboard): add duplicate persona ID validation in create form
Add frontend validation to prevent creating personas with duplicate IDs.
Load existing persona IDs when opening the create form and validate
against them in real-time.
- Add existingPersonaIds array and loadExistingPersonaIds method
- Add validation rule to check for duplicate persona IDs
- Add i18n messages for duplicate ID error (en-US and zh-CN)
- Fix minLength validation to require at least 1 character
* i18n(persona): add createButton translation key for folder dialog
Move create button label to folder-specific translation path
instead of using generic buttons.create key.
* feat(persona): show target folder name in persona creation dialog
Add visual feedback showing which folder a new persona will be created in.
- Add info alert in PersonaForm displaying the target folder name
- Pass currentFolderName prop from PersonaManager and PersonaSelector
- Add recursive findFolderName helper to resolve folder ID to name
- Add i18n translations for createInFolder and rootFolder labels
* style:format code
* fix: remove 'persistent' attribute from dialog components
---------
Co-authored-by: Soulter <905617992@qq.com>
* perf: live mode entry
* chore: remove japanese prompt
---------
Co-authored-by: Anima-IGCenter <cacheigcrystal2@gmail.com>
Co-authored-by: Clhikari <Clhikari@qq.com>
Co-authored-by: jiangman202506 <jiangman202506@163.com>
Co-authored-by: ManJiang <man.jiang@jg-robust.com>
Co-authored-by: Ruochen Pan <67079377+RC-CHN@users.noreply.github.com>
* perf: support extended thinking for Anthropic, DeepSeek reasoning mode, and Gemini text part thought signatures to improve multi-turn reasoning performance.
* chore: remove verbose
* perf
* refactor: remove special tools handling for deepseek-reasoner model in openai source
* fix: improve error handling and logging in InternalAgentSubStage processing
* refactor: remove unused reasoning content from Gemini source processing
* refactor: enhance modality determination logic in useProviderSources
Co-authored-by: kawayiYokami <289104862@qq.com>
* fix: handle None values in _extract_usage to prevent TypeError
Some LLM providers (especially API proxies) may return None for
prompt_tokens and completion_tokens in the usage response. This
causes a TypeError when attempting arithmetic operations.
Added null checks with fallback to 0 for both prompt_tokens and
completion_tokens before performing calculations.
* refactor: use explicit None check and reuse cached variable
- Use `is None` instead of `or 0` to avoid masking unexpected falsy values
- Reuse `cached` variable for `input_cached` to avoid redundant calculation
* ruff format
---------
Co-authored-by: Soulter <905617992@qq.com>
- Expanded model name checks to include specific Gemini 2.5 and 3 variants, ensuring correct configuration for thinking parameters based on the model used.
* feat: supports thinking level of google gemini
- Updated google-genai version to >=1.56.0 in pyproject.toml and requirements.txt.
- Changed model configuration from "gemini-1.5-flash" to "gemini-3-flash-preview" in default.py.
- Enhanced thinking configuration handling in gemini_source.py to support new parameters for Gemini 3 models.
* fix: standardize thinking level configuration in default.py and gemini_source.py
- Updated the thinking level values in default.py to uppercase for consistency.
- Enhanced gemini_source.py to validate the thinking level and default to "HIGH" if an invalid value is provided.
* feat: enhance tool call handling and UI integration for tool calls render
- Added support for tool call messages in the agent runner and webchat event handling.
- Implemented JSON message component for structured tool call data.
- Updated chat route to save tool call information in message history.
- Enhanced frontend to display tool call details in a collapsible format, including status and results.
- Introduced elapsed time tracking for ongoing tool calls in the chat interface.
* fix: improve message handling in agent run utility and tool loop runner
- Refactored message sending logic in `astr_agent_run_util.py` to use `msg_chain` directly for better clarity.
- Added a check in `tool_loop_agent_runner.py` to ensure `tool_call_result_blocks` is not empty before yielding the last tool call result, preventing potential errors.
* refactor: enhance message structure and UI for chat components
- Updated message handling in `MessageList.vue` to support structured message parts, including plain text, images, audio, and files.
- Improved the `Chat.vue` component styles for better visual consistency.
- Refactored message parsing logic in `useMessages.ts` to accommodate new message formats and ensure proper rendering of embedded content.
- Removed deprecated tool call handling from the message structure, streamlining the message display process.
* chore: ruff format
* feat: implement agent statistics tracking and display in chat
- Added `AgentStats` and `TokenUsage` data classes to track agent performance metrics.
- Enhanced `ToolLoopAgentRunner` to collect and update agent statistics during execution.
- Integrated agent statistics sending to webchat for real-time updates.
- Updated chat route to save and display agent statistics in message history.
- Improved frontend components to visualize agent statistics, including token usage and duration metrics.
* fix: improve message handling in Telegram event and agent run utility
- Updated message sending logic in `astr_agent_run_util.py` to send the correct message chain for tool calls.
- Enhanced `tg_event.py` to edit messages during streaming breaks, improving message management and user experience.
- Added error handling for message editing failures to ensure robustness.
* chore: ruff format
LIghtJUNction
jnMetaCode
Futureppo
Stable Genius
Yufeng He
Rin
邹永赫
eason
ChuwuYo
lustresixx
Jason
whatevertogo
exynos
氕氙
Soulter
Minidoracat
NayukiMeko
エイカク
Soulter
Dt8333
搁浅
xunxiing
時壹
tjc66666666
NieiR
simplify123
Yokami
大饼鸡蛋