Fan-out Map Gather Showcase

Split a list into batches, score every batch concurrently with one `horus_map`, gather the results back into one summary.

by Temple Compute

Run Fan-out Map Gather Showcase on Horus

W-06 · Fan-out / Map / Gather Showcase

Split a list into batches, score every batch concurrently with one horus_map, gather the results back into one summary.

Domain: Engine Showcases

Overview

Shows horus-runtime's horus_map: split chunks a JSON list into batch folders, score fans out one clone per batch (word count and uppercase), and gather reads the map's single folder output and merges the results. The workflow has 3 tasks.

Pipeline

split              items.json (8 strings) ──► batches/batch_0..batch_3/data.txt
score[0..3]        horus_map over batches/, max_concurrency 4 ──► scored/<slot>/{count,upper}.txt
gather             scored/ (4 slots) ──► results/summary/

Inputs / Outputs

• Input: examples/items.json, a JSON array of strings.
• scored/<slot>/count.txt: word count of one batch (2 for the bundled example).
• scored/<slot>/upper.txt: the batch text, uppercased.
• results/summary/summary.json: {"total_words": 8, "batches": [...]} for the bundled example.
• results/summary/combined_upper.txt: all uppercased text in batch order.

Parameters

| id | Default | Meaning | |---|---|---| | items | examples/items.json | Items to batch and score | | batch_size | 2 | Items per batch; the batch count is the number of score clones |