Add brev find instance-type command with filtering and sorting

[devin-ai-integration]

Add brev find instance-type command with filtering and Root Disk column

Summary

Implements a new CLI command brev find instance-type that queries the Brev API (https://api.brev.dev/v1/instance/types) to help users find suitable instance types based on various criteria. The command supports filtering by GPU type, count, VRAM, memory, CPU, price, provider, and capabilities, then displays results in a sorted table format.

Key features:

• Comprehensive filtering: Support for --gpu, --min-gpu-count, --min-node-vram, --min-gpu-vram, --min-ram, --min-cpu, --max-hourly-price, --provider, --capabilities, and --min-disk
• Smart sorting: Results sorted by hourly price (ascending) to show most cost-effective options first
• Rich table display: Shows instance type, provider, GPU info, Node VRAM, memory, vCPUs, Root Disk, price, and capabilities
• Root Disk column: Displays disk size (fixed or range) and cheapest price per TB, handling different storage formats across providers (Crusoe: fixed sizes, AWS/GCP: ranges)

The implementation addresses PR feedback by renaming --min-total-vram to --min-node-vram and adding both "Node VRAM" and "Root Disk" columns to the output.

Review & Testing Checklist for Human

• Verify price calculations are accurate - Test that the GB-to-TB conversion (×1000) and price parsing logic produces correct $/TB/hr values for different providers
• Test storage format handling edge cases - Verify Root Disk column correctly handles Crusoe's fixed sizes vs AWS/GCP ranges, and gracefully handles missing/malformed storage data
• Validate memory parsing across units - Test filtering with various memory formats (GB, GiB, TB, etc.) to ensure parseMemoryToGB correctly converts all units
• Test filtering logic comprehensively - Verify edge cases like instances with no GPUs, mixed GPU types, and boundary conditions for numeric filters
• Confirm API error handling - Test behavior when API is unreachable, returns errors, or has malformed responses

Recommended test plan:

# Test different provider storage formats
brev find instance-type --provider crusoe
brev find instance-type --provider aws  
brev find instance-type --provider gcp

# Test edge cases and combinations
brev find instance-type --min-node-vram 100 --max-hourly-price 1.0
brev find instance-type --gpu nonexistent-gpu
brev find instance-type --capabilities invalid-capability

---

Diagram

%%{ init : { "theme" : "default" }}%%
graph TD
    cmd["pkg/cmd/cmd.go<br/>(Register find command)"]:::minor-edit
    find["pkg/cmd/find/find.go<br/>(Main implementation)"]:::major-edit
    http["pkg/store/http.go<br/>(API client & structs)"]:::major-edit
    api["api.brev.dev/v1/instance/types<br/>(External API)"]:::context
    
    cmd -->|"imports & registers"| find
    find -->|"calls GetInstanceTypes()"| http
    http -->|"HTTP request"| api
    api -->|"JSON response"| http
    http -->|"parsed structs"| find
    find -->|"filtered & sorted table"| user["CLI Output"]:::context

    subgraph Legend
        L1["Major Edit"]:::major-edit
        L2["Minor Edit"]:::minor-edit  
        L3["Context/No Edit"]:::context
    end

    classDef major-edit fill:#90EE90
    classDef minor-edit fill:#87CEEB
    classDef context fill:#FFFFFF

Loading

Notes

• Storage complexity: The Root Disk implementation handles 3 different formats: Crusoe (fixed size + count), AWS/GCP (min/max ranges), and fallback cases. This logic is complex and prone to edge cases.
• Memory unit parsing: The parseMemoryToGB function handles various units (B, KB, MB, GB, TB, with/without 'i' suffix). Unit conversion bugs are common in this type of logic.
• API dependency: Command relies on external Brev API which could change format or become unavailable. Error handling exists but may not cover all scenarios.
• Price calculations: Converting from price-per-GB-per-hour to price-per-TB-per-hour involves floating point arithmetic that should be verified for accuracy.

Session details:

• Link to Devin run: https://app.devin.ai/sessions/03a119873a0c4850b935e3c77dd7e82c
• Requested by: @theFong (Alec Fong, [email protected])

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[theFong]

Lets change --min-node-vram to mean the number of gpus * gpu vram. --min-gpu-vram is the per gpu vram. I think Node vram deserves its own column too.

[theFong]

Lets also add a column to the list to show the root disk. If it is static, then should be a scalar value but if it can be a range lets show the min and max it can be. You can choose the cheapest disk price. Lets also show the disk price per TB if available.