Feat/native web admin - #192
Open
qingan123 wants to merge 71 commits into
Open
Conversation
…al input for Qwen AI
Add protocol-level Anthropic heartbeats on both LiteLLM streaming adapters and resume interrupted Qwen responses by response id without resubmitting prompts.
Drop the fixed CHAT2API_QWEN_AI_STREAM_PREFLIGHT_MAX_HOLD_MS default so the upstream HTTP status stays mutable until the first client-visible SSE frame or a terminal failure; the variable is now an optional deployment override that releases a still-quiet stream after N ms (0 = immediately). Trim the tool-workflow recovery heuristics, runtime-plan fields, and the tests that only existed to support the bounded-hold path, and sync the docker and litellm docs to the new preflight model. Co-Authored-By: Claude <noreply@anthropic.com>
Apply one configurable request deadline across Qwen admission, uploads, retries, streams, and recovery while preserving structured timeout and cancellation errors. Harden managed tool output boundaries across built-in streaming providers and cover the behavior with integration tests.
- Add Qwen Hermes protocol adapter for managed tool workflows - Implement session repair mechanism for Qwen AI connections - Add feature config and managed prompt metadata support - Enhance tool calling engine with workflow completion tracking - Add comprehensive test coverage for new protocols and features - Update documentation and deployment configurations Co-Authored-By: Claude <noreply@anthropic.com>
- Update .gitignore to exclude __pycache__/, *.pyc, *.pyo - Update .gitignore to exclude tests/manual/*-live.mjs - Remove tracked pyc file and manual live test scripts Co-Authored-By: Claude <noreply@anthropic.com>
Translate image parts inside Anthropic tool_result blocks into Responses input_image items so clients such as Claude Code can return screenshots through the compat layer. Guard the Responses transformation patches with marker checks so re-applying the build patch is idempotent, and extend the integration and compat tests to cover image-bearing tool output. Co-Authored-By: Claude <noreply@anthropic.com>
Map Qwen3.8-Max_Fast/_Auto/_Thinking (plus raw _TeX_AtX flags) to qwen3.8-max with per-mode thinking_enabled/auto_thinking feature flags, split auto_thinking from thinking_enabled in the Qwen feature config, replace the adapter _forceThinking hack with resolveQwenAiFeatureMode, append the aliases after qwen-ai catalogue refreshes, and document the modes with updated tests. Co-Authored-By: Claude <noreply@anthropic.com>
… continuation - Add qwenAiAccountPolicy: narrow account-fault classification to prevent pool drain - Add qwenAiSessionBridge and qwenAiToolCallSessionStore for cross-request tool continuation - Token refresh: detect unregistered accounts and persist inactive status - Forwarder: Responses session continuation with stale-session fallback - Governor: restrict cooldown to capacity-limit 429 only, not ordinary rate limits - Routes: tool-call session claim lifecycle management in chat and responses endpoints Co-Authored-By: Claude <noreply@anthropic.com>
…scripts Co-Authored-By: Claude <noreply@anthropic.com>
Also ignore local .tmp-live-*.ndjson debug logs. Co-Authored-By: Claude <noreply@anthropic.com>
Merge pyf-feifei/Chat2API as the maintained application base while preserving the existing deployment scripts and prepend-only README deployment block.
- add administrator username and password login for the web admin - deploy fork and Py upstream as isolated Docker instances - add conservative multi-instance updates with backups and rollback - document official/Py upstream comparison and maintenance workflow
Avoid overriding persisted UI settings on restart and make updates follow the branch recorded by each deployment.
Remove the custom administrator username layer and restore the Py upstream WebUI and management API authentication contract exactly.
- use the complete Py edition with native Management Secret login - keep API keys and storage encryption secrets separate - add versioned multi-instance metadata, backups, and rollback - document official/Py divergence and known upstream test baselines
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.