Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Principle:Infiniflow Ragflow Progress Tracking

From Leeroopedia
Knowledge Sources
Domains RAG, Task_Queue, Observability
Last Updated 2026-02-12 06:00 GMT

Overview

A task lifecycle management pattern that tracks processing progress, accumulates statistics, and finalizes document state after pipeline completion.

Description

Progress Tracking maintains real-time visibility into document processing status. Each task's progress (0.0-1.0, or -1 for error) is updated atomically with monotonic progression (progress only increases or goes to -1). Progress messages are appended and trimmed to 3000 lines. Upon completion, document-level statistics (chunk_num, token_num, process_duration) are updated, and knowledge base totals are recalculated.

Usage

Used throughout the document processing pipeline to report status. Progress is visible in the UI document list.

Theoretical Basis

Progress tracking follows monotonic state machine semantics:

  • Progress only moves forward (new value must exceed current) unless transitioning to error state (-1)
  • Error state (-1) is terminal and locks the progress field
  • Process duration is computed from task creation to completion

Related Pages

Implemented By

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment