feat: map Ollama options to OpenAI format in HTTP handler

[ilopezluna]

This pull request refactors how Ollama API options are mapped to the OpenAI-compatible request format in the HTTPHandler. The main improvement is the introduction of a dedicated helper function to centralize and streamline option mapping.

[sourcery-ai]

Hey there - I've reviewed your changes - here's some feedback:

• Consider making the backendSpecificOptions list a package-level var or const (or at least a shared slice) so it’s not reallocated on every call to mapOllamaOptionsToOpenAI, especially if this path is hot.
• The debug log inside the loop over backendSpecificOptions will emit a line per option per request; if this handler is high-traffic, you might want to either aggregate this logging, gate it behind a higher log level, or remove it to avoid log noise.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- Consider making the `backendSpecificOptions` list a package-level `var` or `const` (or at least a shared slice) so it’s not reallocated on every call to `mapOllamaOptionsToOpenAI`, especially if this path is hot.
- The debug log inside the loop over `backendSpecificOptions` will emit a line per option per request; if this handler is high-traffic, you might want to either aggregate this logging, gate it behind a higher log level, or remove it to avoid log noise.

## Individual Comments

### Comment 1
<location> `pkg/ollama/http_handler.go:691-700` </location>
<code_context>
+
+	// Backend-specific options that may not be supported by all OpenAI-compatible backends
+	// We'll pass these through and let the backend decide whether to use them
+	backendSpecificOptions := []string{
+		"repeat_last_n",
+		"typical_p",
+		"min_p",
+		"num_keep",
+		"num_batch",
+		"num_gpu",
+		"main_gpu",
+		"use_mmap",
+		"num_thread",
+	}
+
</code_context>

<issue_to_address>
**issue (bug_risk):** Passing backend-specific options through as top-level fields may break stricter OpenAI-compatible backends

Forwarding `repeat_last_n`, `num_gpu`, `num_thread`, etc. as top-level `openAIReq` fields assumes all OpenAI-compatible backends ignore unknown keys. Some may instead reject the request. Consider either scoping this to specific backends, namespacing under a dedicated options field, or making the passthrough configurable so stricter backends can disable it.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}