add tools

[sergei-bronnikov]

https://bugtracker.codiodev.com/issue/codio-17375/eCornell-still-having-issues-they-think-tools-are-being-blocked

Summary by CodeRabbit

• New Features
  • Code interpreter tool is now available as a fully supported feature
  • Containers support configurable memory limits (1GB, 4GB, 16GB, 64GB) with separate per-tier pricing models
  • Container creation costs are automatically estimated and included in overall API usage charges
  • Built-in validation ensures only supported memory configurations are accepted

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[AndreyNikitin]

@coderabbitai full review, summary

[coderabbitai]

@AndreyNikitin: I'll perform a full review of the PR and regenerate the summary.

✅ Actions performed

Summary regeneration triggered.

[coderabbitai]

Walkthrough

Introduces container-specific cost modeling for OpenAI code interpreter tools through a memory-limit-based cost map, API methods to estimate container creation costs, updated request types with memory limit support, and middleware validation for container configurations.

Changes

Cohort / File(s)	Summary
Cost model internal/provider/openai/cost.go	Adds OpenAiCodeInterpreterContainerCost map for memory-limit-to-cost pricing and new EstimateResponseApiToolCreateContainerCost() method on CostEstimator to compute container creation costs per tool container memory limit.
Request type structure internal/provider/openai/types.go	Introduces ResponseRequestToolContainer with MemoryLimit field and GetMemoryLimit() method; refactors ResponseRequestToolUnion with Container field and adds GetContainerAsResponseRequestToolContainer() method for runtime type conversion.
Middleware integration internal/server/web/proxy/middleware.go	Adds EstimateResponseApiToolCreateContainerCost() to estimator interface; implements container memory limit validation against allowed costs with 403 abort on invalid limits.
Response cost calculation internal/server/web/proxy/responses.go	Integrates container cost calculation into response paths; adds ginCtxSetResponsesRequest() and ginCtxGetResponsesRequest() helper functions for context-based request storage and retrieval.

Sequence Diagram

sequenceDiagram
    participant Client
    participant Middleware as Middleware
    participant Estimator as Cost Estimator
    participant Response as Response Handler
    
    Client->>Middleware: POST /responses with tool containers
    Middleware->>Middleware: Extract tool container memory limits
    Middleware->>Estimator: Validate memory limits against<br/>OpenAiCodeInterpreterContainerCost
    alt Memory limit not allowed
        Estimator-->>Middleware: Error
        Middleware-->>Client: 403 Forbidden
    else Memory limit valid
        Middleware->>Middleware: Store response request in context
        Middleware->>Response: Process response
        Response->>Estimator: EstimateResponseApiToolCreateContainerCost()
        Estimator-->>Response: Container cost (float64)
        Response->>Response: Add container cost to total cost
        Response-->>Client: Response with final cost
    end

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• PR #9 — Extends existing OpenAI responses and tool cost handling by adding container-specific cost modeling and request type updates to the same files and estimator interface introduced in that PR.

✨ Finishing touches

• 📝 Generate docstrings

---

---

📜 Recent review details

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between e1ea19b and 786c321.

📒 Files selected for processing (4)

• internal/provider/openai/cost.go
• internal/provider/openai/types.go
• internal/server/web/proxy/middleware.go
• internal/server/web/proxy/responses.go

_{✏️ Tip: You can disable this entire section by setting review_details to false in your review settings.}

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}