Implement filtering of OpenCode tool-call JSON lines in output

[beettlle]

Summary

• Introduces a new function filterOpencodeToolCallLines to clean up the output from the OpenCode CLI by removing JSON lines that represent tool calls, which are not needed for the review process.
• Adds a helper function isOpencodeToolCallLine to identify these JSON lines.

Changes

• Updated the Review method to utilize the new filtering function on the command output before returning the result.

Test plan

• Verify that the output from OpenCodeAgent.Review no longer includes tool-call JSON lines.

Generated with Cursor

[wesm]

Thank you. I'll push some more commits here -- I'm also going to add some configuration options for opencode (e.g. model selection in .roborev.toml or config.toml)

[beettlle]

I'm getting other errors with OpenCode now. Will keep looking a this later.

My use case is, I use Cursor and want to use roborev with a local Ollama model for reviews to save some money. My MacPro can run qwen2.5-coder:32b with 32k context window without breaking a sweat.

[wesm]

Makes sense. I have a couple inference servers on my tailnet I can use for testing so I can work on ollama support after I get #126 merged and cut the 0.17.0 release

[wesm]

This was included in #126

[beettlle]

In case it's useful, this is the opencode.json that I have that got it to, mostly, work.

{
  "$schema": "https://opencode.ai/config.json",
  "model": "ollama/roborev-coder",
  "permission": {
    "*": "deny",
    "read": "allow",
    "edit": "allow",
    "bash": "allow",
    "grep": "allow",
    "glob": "allow",
    "list": "allow",
    "webfetch": "deny",
    "task": "deny",
    "todoread": "deny",
    "todowrite": "deny"
  },
  "provider": {
    "ollama": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ollama",
      "options": {
        "baseURL": "http://localhost:11434/v1"
      },
      "models": {
        "roborev-coder": {
          "name": "roborev-coder (local)"
        }
      }
    }
  }
}