Add token tracking, API key auth, and tool calling to Ollama client

[Copilot]

Enhances Ollama integration with production-ready features: full token usage metrics, optional API key authentication, and tool/function calling for web search and URL reading.

Token Usage Tracking

• Captures prompt_eval_count and eval_count from API responses
• Exposes per-request and cumulative session metrics via LastPromptTokens, TotalTokensUsed, etc.
• Handles multi-request flows (tool calls) correctly by accumulating tokens across both initial and follow-up requests

var response = await client.ChatAsync("What is 2+2?");
Console.WriteLine(client.GetTokenUsageSummary());
// Output: Last Request: 45 prompt + 12 completion = 57 total
//         Total Session: 45 prompt + 12 completion = 57 total

API Key Authentication

• ApiKey property setter automatically updates Authorization: Bearer header
• Persisted in AppSettings with password-masked UI input
• Empty key removes authentication header

Web Search & URL Reading

• Implements Ollama's tool/function calling API for DuckDuckGo search and URL content fetching
• Tools dynamically filtered by enabled features to minimize request payload
• Two-phase flow: model requests tool → execute locally → inject results → final response
• URL content: HTML stripped, decoded, truncated to 2000 chars (MAX_URL_CONTENT_LENGTH)

client.EnableWebSearch = true;
client.EnableUrlReading = true;
var response = await client.ChatAsync("What's on example.com?");
// Model can call read_url tool, response incorporates webpage content

UI controls added to Settings > Ollama tab for all features. Backward compatible—all features opt-in.

Original prompt

Make the Ollama implementation better

• full token usage
• able to use API key
• Websearching and URL reading

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.