embedding-ft

分类: 数据与AI | 上传者: tiangong-lca | 下载: 0 | 版本: v1.0（最新）

Execute and troubleshoot Supabase edge function `embedding_ft` that consumes PGMQ embedding jobs, calls AWS SageMaker embeddings, and writes vectors back to Postgres. Use when validating job payload handling, investigating failed embeddings, tuning ack semantics, or adjusting worker auth/environment.

更新日志: Source: GitHub https://github.com/tiangong-lca/skills

目录结构

当前层级: embedding-ft/

📁 agents/
- 📄 openai.yaml 212 B
📁 assets/
- 📄 example-jobs.json 237 B
📁 references/
- 📄 env.md 289 B
- 📄 job-contract.md 978 B
- 📄 testing.md 666 B
📁 scripts/
- 📄 run-embedding-ft.sh 2.2 KB
📄 SKILL.md 1.1 KB

SKILL.md

---
name: embedding-ft
description: Execute and troubleshoot Supabase edge function `embedding_ft` that consumes PGMQ embedding jobs, calls AWS SageMaker embeddings, and writes vectors back to Postgres. Use when validating job payload handling, investigating failed embeddings, tuning ack semantics, or adjusting worker auth/environment.
---

# Embedding FT

## Run Workflow
1. Set auth token (`TOKEN`) or pass `--token`.
2. Execute `scripts/run-embedding-ft.sh` with a request file.
3. Inspect `completedJobs` and `failedJobs`, then triage via references.

## Commands
```bash
scripts/run-embedding-ft.sh --dry-run
scripts/run-embedding-ft.sh --token "$TOKEN"
scripts/run-embedding-ft.sh --data ./assets/example-jobs.json --token "$TOKEN"
```

## Fast Triage
- `400`: request body is not a valid job array.
- `500`: SageMaker request/response parsing failure.
- `completedJobs < submitted`: inspect queue payload, row version, and content function output.

## Load References On Demand
- `references/env.md`: auth and caller environment.
- `references/job-contract.md`: queue semantics and DB side effects.
- `references/testing.md`: smoke-test and debug checklist.

登录后下载/点赞/收藏 ❤ 3 | ★ 0

embedding-ft

目录结构

SKILL.md

举报内容

提示