Adapt SWEAgentTranslator for local model use via Ollama and vLLM

[ikhillan]

No description provided.

[Copilot]

Pull Request Overview

This PR adapts the SWEAgentTranslator class to support local LLM models via Ollama, in addition to existing cloud-based models. The changes enable users to run translations using local models with appropriate configurations.

Key changes:

• Adds automatic detection and setup for Ollama-based models (prefixed with "ollama/")
• Implements automatic Ollama server launch with health checks
• Introduces new configuration parameters for parser type, max input tokens, and custom config files

Reviewed Changes

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

File	Description
targets/microXOR/cuda/repo/translation_task.md	Removes translation task documentation (entire file deleted)
src/translate/swe_agent/swe_agent_translator.py	Adds Ollama support with server management, new configuration options, and model detection logic

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Undefined variable 'env'. The Popen call references 'env' parameter but it is not defined in this function. Either remove the env parameter or define it before use.

Suggested change

	start_new_session=True,
	env=env)
	start_new_session=True)

[Copilot]

Inconsistent indentation in Popen arguments. The arguments should be consistently indented, and trailing whitespace on lines 108-109 should be removed.

Suggested change

	subprocess.Popen(ollama_command,
	stdout=subprocess.DEVNULL,
	stderr=subprocess.STDOUT,
	stdin=subprocess.DEVNULL,
	start_new_session=True,
	env=env)
	subprocess.Popen(
	ollama_command,
	stdout=subprocess.DEVNULL,
	stderr=subprocess.STDOUT,
	stdin=subprocess.DEVNULL,
	start_new_session=True,
	env=env
	)