Add SSE disable header, data-URI format inference, broader response text
parsing, HTTP contract e2e, pipeline routing tests, and ASR model selection
guidance in bailian-gen.
Reviewed against the official text-tuning data rules; fixes two confirmed
mismatches and fills enforcement gaps:
- thinking: exempt assistant messages carrying tool_calls from the
THINK_TAG_NOT_LAST check — the spec's tool+thinking combo example puts
<think> on a non-last assistant and was previously false-flagged
- DPO support matrix: reject image/video content items, tools, tool_calls
and role:tool (DPO_UNSUPPORTED_ELEMENT); also scan chosen/rejected
- DPO: messages not ending with user upgraded warning -> error
- OpenAI migration: name/weight upgraded warning -> error (spec: must not
carry); drop dead record-level name branch
- tool_call_id: unmatched tool response upgraded warning -> error
(one-to-one per spec); new TOOL_CALL_NO_RESPONSE warning for orphan calls
- loss_weight: validate range at message level too; warn when placed on
anything but the last assistant message (LOSS_WEIGHT_PLACEMENT)
- video params: fps/sample_fps must be within [0.1, 10]
(INVALID_VIDEO_FPS); mode-mismatched params warned
(VIDEO_PARAM_MODE_MISMATCH); video_start/video_end type-checked
- zip: skip macOS packaging metadata (__MACOSX/, .DS_Store, ._*) in
filename constraints and image counting to stop false failures on
Finder-created archives
Tests 45 -> 59 covering every new/changed rule, including a replica of the
spec's official tool+thinking example.
- Support content array format [{text/image/video}] alongside legacy string
- Add tool role support with tool_calls structure and tool_call_id validation
- Add thinking tag placement check (only in last assistant message)
- Add OpenAI migration guards: warn on unsupported name/weight fields
- Add loss_weight range validation (0.0–1.0)
- Fix size limits: SFT/DPO 200MB, CPT 300MB, media ZIP 2GB
- Enforce data.jsonl at ZIP root (reject nested wrapping folders)
- Add ZIP filename constraints: charset [a-zA-Z0-9_-], length ≤120, uniqueness
- Add .tif to accepted image extensions
- Add DPO_LAST_MSG_NOT_USER warning when messages don't end with user role
- CPT profile now uses dedicated 300MB cap instead of shared default
- Expand unit tests from 19 to 45 covering all new validation paths
Previously, fan-out skipped any existing real directory not recorded in
the lock file, treating it as user content. This left stale skill copies
installed by other tools (e.g. npx skills add) permanently out of date.
Now: if the directory contains a SKILL.md, it is recognized as a skill
artifact and replaced with a symlink to the canonical dir. Directories
without SKILL.md are still preserved (user content safety boundary).
Stop inferring size/prompt_extend from sync vs async; use per-family sizeProfile. wanx*-imageedit uses function+base_image_url; bare qwen-image uses the fixed resolution table.
- convert local images to Base64 for Token Plan image and video commands
- preserve the existing OSS upload flow for standard API Key profiles
- use wan2.7-image as the default image model with the sync endpoint
- hide full Base64 image content in dry-run output
- add Token Plan compatibility tests and update related docs
- switch the default text model to qwen3.8-max-preview
- add dedicated T2V, I2V, and R2V model defaults
- persist and consume per-mode video model settings
- enable thinking when validating the qwen3.8 preview model
- persist the active profile in config.json
- resolve config with --config > active_config > default
- add config list and config use commands
- make auth and config writes target the selected profile
- reset activation to default when deleting the active profile
- update config UI with profile activation controls
- keep token refresh and pipeline execution profile-aware
- add loader, UI, auth, and CLI interaction coverage
- add the built-in Token Plan profile preset
- validate and persist model API keys atomically
- materialize the default Base URL and models on login
- preserve flag > env > config precedence
故璃