Mtauraso robot instructions

[mtauraso]

No description provided.

[Copilot]

Pull request overview

This pull request consolidates AI coding assistant instructions into a canonical reference document. The PR creates a new HYRAX_GUIDE.md that serves as the single source of truth for project conventions, architecture, and development workflows, while updating both CLAUDE.md and .github/copilot-instructions.md to reference it rather than duplicating content.

Changes:

• Creates comprehensive HYRAX_GUIDE.md covering project architecture, design principles, development setup, repository structure, and common workflows
• Refactors CLAUDE.md to be Claude-specific, deferring to HYRAX_GUIDE.md for general information
• Refactors .github/copilot-instructions.md to be Copilot-specific, deferring to HYRAX_GUIDE.md for general information
• Fixes Python version requirement from 3.10 to 3.11 (matching pyproject.toml)

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
HYRAX_GUIDE.md	New canonical reference guide for AI assistants; documents design principles, architecture, plugin registries, configuration system, and development workflows
CLAUDE.md	Simplified to Claude Code-specific guidance; references HYRAX_GUIDE.md for comprehensive project information
.github/copilot-instructions.md	Simplified to GitHub Copilot-specific guidance; emphasizes long-running command timeouts and validation workflow; references HYRAX_GUIDE.md

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 62.96%. Comparing base (a800e9c) to head (a371be9).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #679   +/-   ##
=======================================
  Coverage   62.96%   62.96%           
=======================================
  Files          59       59           
  Lines        5730     5730           
=======================================
  Hits         3608     3608           
  Misses       2122     2122           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[mtauraso]

@copilot Please review the robot instructions on this branch, and comment on any issues that would be confusing or unclear to an agent

[Copilot]

@mtauraso I've opened a new pull request, #680, to work on those changes. Once the pull request is ready, I'll request review from you.

[github-actions]

Before [a800e9c]	After [963d2dc]	Ratio	Benchmark (Parameter)
8.89±0.1ms	9.33±0.1ms	1.05	vector_db_benchmarks.VectorDBSearchBenchmarks.time_search_by_vector_many_shards(64, 'chromadb')
9.20±0.1ms	9.43±0.2ms	1.03	vector_db_benchmarks.VectorDBSearchBenchmarks.time_search_by_vector_many_shards(128, 'chromadb')
114±1μs	115±1μs	1.01	data_request_benchmarks.DatasetRequestBenchmarks.time_request_all_data
6.85±0.05s	6.89±0.02s	1.01	vector_db_benchmarks.VectorDBInsertBenchmarks.time_load_vector_db(2048, 'qdrant')
194±1ms	193±2ms	1	benchmarks.time_import
36.6±0.3ms	36.8±0.8ms	1	benchmarks.time_nb_obj_construct
7.61±0.01s	7.58±0.03s	1	data_cache_benchmarks.DataCacheBenchmarks.time_preload_cache_hsc1k
1.63G	1.63G	1	vector_db_benchmarks.VectorDBInsertBenchmarks.peakmem_load_vector_db(16384, 'chromadb')
3.87G	3.88G	1	vector_db_benchmarks.VectorDBInsertBenchmarks.peakmem_load_vector_db(16384, 'qdrant')
1.03G	1.03G	1	vector_db_benchmarks.VectorDBInsertBenchmarks.peakmem_load_vector_db(2048, 'chromadb')

Click here to view all benchmarks.

[mtauraso]

General feedback

High level Assertions (not true today, but want to be):

• Support Pytorch now, but we want to support other frameworks down the line: (List specific disciplines we're holding to leave room for this.)
• Moving toward a ecosystem model where datasets and models are easily selectable by the user/ want to support open/community contributed models and datasets a-la-carte
• General principle of future support for iterator datasets and continous running; however today there's a strict rule that all data is of finite size and map-style random access. Access by index!!
• Want most datsets and models to be external libraries; however, today we don't have enough examples for most users to do this!

Missing Principles:

• Everything is Numpy so much as it possibly can be. Up until the last moment where data is fed to the model (in specific function in pytorch ignite). Motive is framework independence.
• Push against perfusion of verbs and configs. We want to avoid adding new ones.

Note about boyscouting/ leaving it better than you found it/ Making it so the current assumptions of a change appear to have always been true.

"Matplotlib for ML": Understand the LEVEL of our users. Our users write code, but they can't really handle classes or multiple files so you need example-driven documentation showing the right way to use these primitives. Go for compact inspirational examples which demonstrate the breadth of the framework.

[mtauraso]

Last point: Common customizations ought to happen at the notebook level! Should be easy for researchers to get started in a notebook, they can deal with publishing code as an external plugin later. (mostly for documentation, but also for notebook interface design)

[mtauraso]

(Note of don't run)

[mtauraso]

String = "" also a reasonable convention.