agentscope-ai / agentscope-ai/agentscope
feat(pipeline): support pipeline module in agentscope
- Langage dominant
- Python
- Étoiles
- 31.6k
- Forks
- 3.5k
- Merge moyen
- 1 j 16 h
- PR mergées (30 j)
- 103
Description
**Background**
While implementing SOP support in AgentScope, we identified strong structural similarities between pipelines and SOPs — both require unified input interfaces, human-in-the-loop (HITL) handling, and orchestration of multi-agent operations. The key distinction is that pipelines operate on **fixed, predefined logic**, making them a natural and solid foundation for building SOPs.
**Changes**
- Introduce a `pipeline` module in AgentScope
- Support a variety of pipeline types
**Note**
Unlike AgentScope 1.0, the goal of this new pipeline module is to provide **more mature, well-encapsulated functionality**, rather than limiting it to simple primitives like `fanout` or `sequential`. We aim to cover richer, more practical orchestration patterns.
If you have proposals for new pipeline types or designs, feel free to open a discussion and submit a proposal — contributions and ideas are welcome!
Guide de contribution
Ouvrir le guide de contribution
Piste de recherche
No concrete files or tests are listed in the issue. Start by reading the existing SOP and orchestration-related codepaths and docs to infer how unified inputs, HITL, and multi-agent execution are currently modeled and exposed. Next, identify the module entry points where public functionality is registered so a new `pipeline` namespace can be added consistently, then run the relevant orchestration tests (or existing pipeline-equivalent tests) after any initial type implementation. Consider it done when a first cut of the pipeline module/API is integrated and documented enough to show clear behavior for one or more pipeline types.
Rédigé par le modèle d'indexation à partir du texte de l'issue.
Évaluation
- Stack technique
- python
- Domaine
- ai-infra-agents, backend
- Type d'issue
- Fonctionnalité
- Difficulté
- 5/5
- Temps estimé
- Plus d'une semaine
- Activité
- Active
- Clarté
- À clarifier
- Accessibilité débutants
- 30/100