ppt-task-orchestrator
Orchestration layer for customer PPT workflows:
- 1. Parse
pptx into per-page tasks (structured fields first). - Fallback to OCR for pages missing required fields.
- Convert page tasks into
main-image-editor batch tasks. - Execute with transaction rollback by reusing
main-image-editor and psd-automator. - Aggregate delivery artifacts: result directory +
all-images.zip.
Usage
CODEBLOCK0
Dry-run:
CODEBLOCK1
Request payload
INLINECODE5 JSON supports:
- -
pptPath: absolute or ~/ path to .pptx file (required) - INLINECODE9 : optional tracking id
- INLINECODE10 : low-confidence gate threshold (default
0.8) - INLINECODE12 :
structured_only or structured_first_with_ocr (default) - INLINECODE15 :
dryRun, force, indexPath, INLINECODE19 - INLINECODE20 :
outputDir, zipName, INLINECODE23
Failure policy
- - Default policy is
rollback_all inherited from main-image-editor. - Any single PSD failure restores all touched PSD files from transaction backups.
- Delivery artifacts are produced only for success/dry-run outputs.
ppt-task-orchestrator
客户PPT工作流的编排层：
- 1. 将pptx解析为逐页任务（优先处理结构化字段）。
- 对缺少必填字段的页面回退至OCR识别。
- 将页面任务转换为main-image-editor的批量任务。
- 通过复用main-image-editor和psd-automator执行事务回滚。
- 聚合交付产物：结果目录 + all-images.zip。
使用方法
bash
node skills/ppt-task-orchestrator/scripts/run-ppt-task-orchestrator.js \
--request skills/ppt-task-orchestrator/examples/request.sample.json \
--index ~/.openclaw/psd-index.json
试运行：
bash
node skills/ppt-task-orchestrator/scripts/run-ppt-task-orchestrator.js \
--request skills/ppt-task-orchestrator/examples/request.sample.json \
--dry-run
请求负载
request JSON支持以下参数：
- - pptPath：.pptx文件的绝对路径或~/路径（必填）
- requestId：可选追踪ID
- confidenceThreshold：低置信度阈值（默认0.8）
- fallbackPolicy：structuredonly或structuredfirstwithocr（默认）
- execution：dryRun、force、indexPath、bundleZip
- delivery：outputDir、zipName、copySelectedOnly
失败策略
- - 默认策略为继承自main-image-editor的rollback_all。
- 任一PSD失败将从事务备份中恢复所有受影响的PSD文件。
- 仅对成功/试运行的输出生成交付产物。
