zeroclaw/src/tools/web_search_tool.rs

use super::traits::{Tool, ToolResult};
use crate::security::SecurityPolicy;
use async_trait::async_trait;
use regex::Regex;
use serde_json::json;
use std::sync::Arc;
use std::time::Duration;

/// Web search tool for searching the internet.
/// Supports providers: DuckDuckGo (free), Brave, Firecrawl.
pub struct WebSearchTool {
    security: Arc<SecurityPolicy>,
    provider: String,
    api_key: Option<String>,
    api_url: Option<String>,
    max_results: usize,
    timeout_secs: u64,
}

impl WebSearchTool {
    pub fn new(
        security: Arc<SecurityPolicy>,
        provider: String,
        api_key: Option<String>,
        api_url: Option<String>,
        max_results: usize,
        timeout_secs: u64,
    ) -> Self {
        Self {
            security,
            provider: provider.trim().to_lowercase(),
            api_key,
            api_url,
            max_results: max_results.clamp(1, 10),
            timeout_secs: timeout_secs.max(1),
        }
    }

    async fn search_duckduckgo(&self, query: &str) -> anyhow::Result<String> {
        let encoded_query = urlencoding::encode(query);
        let search_url = format!("https://html.duckduckgo.com/html/?q={}", encoded_query);

        let client = reqwest::Client::builder()
            .timeout(Duration::from_secs(self.timeout_secs))
            .user_agent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36")
            .build()?;

        let response = client.get(&search_url).send().await?;

        if !response.status().is_success() {
            anyhow::bail!(
                "DuckDuckGo search failed with status: {}",
                response.status()
            );
        }

        let html = response.text().await?;
        self.parse_duckduckgo_results(&html, query)
    }

    fn parse_duckduckgo_results(&self, html: &str, query: &str) -> anyhow::Result<String> {
        // Extract result links: <a class="result__a" href="...">Title</a>
        let link_regex = Regex::new(
            r#"<a[^>]*class="[^"]*result__a[^"]*"[^>]*href="([^"]+)"[^>]*>([\s\S]*?)</a>"#,
        )?;

        // Extract snippets: <a class="result__snippet">...</a>
        let snippet_regex = Regex::new(r#"<a class="result__snippet[^"]*"[^>]*>([\s\S]*?)</a>"#)?;

        let link_matches: Vec<_> = link_regex
            .captures_iter(html)
            .take(self.max_results + 2)
            .collect();

        let snippet_matches: Vec<_> = snippet_regex
            .captures_iter(html)
            .take(self.max_results + 2)
            .collect();

        if link_matches.is_empty() {
            return Ok(format!("No results found for: {}", query));
        }

        let mut lines = vec![format!("Search results for: {} (via DuckDuckGo)", query)];

        let count = link_matches.len().min(self.max_results);

        for i in 0..count {
            let caps = &link_matches[i];
            let url_str = decode_ddg_redirect_url(&caps[1]);
            let title = strip_tags(&caps[2]);

            lines.push(format!("{}. {}", i + 1, title.trim()));
            lines.push(format!("   {}", url_str.trim()));

            // Add snippet if available
            if i < snippet_matches.len() {
                let snippet = strip_tags(&snippet_matches[i][1]);
                let snippet = snippet.trim();
                if !snippet.is_empty() {
                    lines.push(format!("   {}", snippet));
                }
            }
        }

        Ok(lines.join("\n"))
    }

    async fn search_brave(&self, query: &str) -> anyhow::Result<String> {
        let auth_token = match self.api_key.as_ref() {
            Some(raw) if !raw.trim().is_empty() => raw.trim(),
            _ => anyhow::bail!("Brave API key not configured"),
        };

        let encoded_query = urlencoding::encode(query);
        let search_url = format!(
            "https://api.search.brave.com/res/v1/web/search?q={}&count={}",
            encoded_query, self.max_results
        );

        let client = reqwest::Client::builder()
            .timeout(Duration::from_secs(self.timeout_secs))
            .build()?;

        let response = client
            .get(&search_url)
            .header("Accept", "application/json")
            .header("X-Subscription-Token", auth_token)
            .send()
            .await?;

        if !response.status().is_success() {
            anyhow::bail!("Brave search failed with status: {}", response.status());
        }

        let json: serde_json::Value = response.json().await?;
        self.parse_brave_results(&json, query)
    }

    fn parse_brave_results(&self, json: &serde_json::Value, query: &str) -> anyhow::Result<String> {
        let results = json
            .get("web")
            .and_then(|w| w.get("results"))
            .and_then(|r| r.as_array())
            .ok_or_else(|| anyhow::anyhow!("Invalid Brave API response"))?;

        if results.is_empty() {
            return Ok(format!("No results found for: {}", query));
        }

        let mut lines = vec![format!("Search results for: {} (via Brave)", query)];

        for (i, result) in results.iter().take(self.max_results).enumerate() {
            let title = result
                .get("title")
                .and_then(|t| t.as_str())
                .unwrap_or("No title");
            let url = result.get("url").and_then(|u| u.as_str()).unwrap_or("");
            let description = result
                .get("description")
                .and_then(|d| d.as_str())
                .unwrap_or("");

            lines.push(format!("{}. {}", i + 1, title));
            lines.push(format!("   {}", url));
            if !description.is_empty() {
                lines.push(format!("   {}", description));
            }
        }

        Ok(lines.join("\n"))
    }

    #[cfg(feature = "firecrawl")]
    async fn search_firecrawl(&self, query: &str) -> anyhow::Result<String> {
        let auth_token = match self.api_key.as_ref() {
            Some(raw) if !raw.trim().is_empty() => raw.trim(),
            _ => {
                anyhow::bail!(
                    "web_search provider 'firecrawl' requires [web_search].api_key in config.toml"
                );
            }
        };

        let api_url = self
            .api_url
            .as_deref()
            .map(str::trim)
            .filter(|s| !s.is_empty())
            .unwrap_or("https://api.firecrawl.dev");
        let endpoint = format!("{}/v1/search", api_url.trim_end_matches('/'));
        let client = reqwest::Client::builder()
            .timeout(Duration::from_secs(self.timeout_secs))
            .build()?;

        let response = client
            .post(endpoint)
            .header(
                reqwest::header::AUTHORIZATION,
                format!("Bearer {auth_token}"),
            )
            .json(&json!({
                "query": query,
                "limit": self.max_results,
                "timeout": (self.timeout_secs * 1000) as u64,
            }))
            .send()
            .await
            .map_err(|e| anyhow::anyhow!("Firecrawl search failed: {e}"))?;
        let status = response.status();
        let body = response.text().await?;

        if !status.is_success() {
            anyhow::bail!(
                "Firecrawl search failed with status {}: {}",
                status.as_u16(),
                body
            );
        }

        let parsed: serde_json::Value = serde_json::from_str(&body)
            .map_err(|e| anyhow::anyhow!("Invalid Firecrawl response JSON: {e}"))?;
        if !parsed
            .get("success")
            .and_then(serde_json::Value::as_bool)
            .unwrap_or(false)
        {
            let error = parsed
                .get("error")
                .and_then(serde_json::Value::as_str)
                .unwrap_or("unknown error");
            anyhow::bail!("Firecrawl search failed: {error}");
        }

        let results = parsed
            .get("data")
            .and_then(serde_json::Value::as_array)
            .ok_or_else(|| anyhow::anyhow!("Firecrawl response missing data array"))?;

        if results.is_empty() {
            return Ok(format!("No results found for: {}", query));
        }

        let mut lines = vec![format!("Search results for: {} (via Firecrawl)", query)];

        for (i, result) in results.iter().take(self.max_results).enumerate() {
            let title = result
                .get("title")
                .and_then(serde_json::Value::as_str)
                .unwrap_or("No title");
            let url = result
                .get("url")
                .and_then(serde_json::Value::as_str)
                .unwrap_or("");
            let description = result
                .get("description")
                .and_then(serde_json::Value::as_str)
                .unwrap_or("");

            lines.push(format!("{}. {}", i + 1, title));
            lines.push(format!("   {}", url));
            if !description.trim().is_empty() {
                lines.push(format!("   {}", description.trim()));
            }
        }

        Ok(lines.join("\n"))
    }

    #[cfg(not(feature = "firecrawl"))]
    #[allow(clippy::unused_async)]
    async fn search_firecrawl(&self, _query: &str) -> anyhow::Result<String> {
        anyhow::bail!("web_search provider 'firecrawl' requires Cargo feature 'firecrawl'")
    }
}

fn decode_ddg_redirect_url(raw_url: &str) -> String {
    if let Some(index) = raw_url.find("uddg=") {
        let encoded = &raw_url[index + 5..];
        let encoded = encoded.split('&').next().unwrap_or(encoded);
        if let Ok(decoded) = urlencoding::decode(encoded) {
            return decoded.into_owned();
        }
    }

    raw_url.to_string()
}

fn strip_tags(content: &str) -> String {
    let re = Regex::new(r"<[^>]+>").unwrap();
    re.replace_all(content, "").to_string()
}

#[async_trait]
impl Tool for WebSearchTool {
    fn name(&self) -> &str {
        "web_search_tool"
    }

    fn description(&self) -> &str {
        "Search the web for information. Returns relevant search results with titles, URLs, and descriptions. Use this to find current information, news, or research topics."
    }

    fn parameters_schema(&self) -> serde_json::Value {
        json!({
            "type": "object",
            "properties": {
                "query": {
                    "type": "string",
                    "description": "The search query. Be specific for better results."
                }
            },
            "required": ["query"]
        })
    }

    async fn execute(&self, args: serde_json::Value) -> anyhow::Result<ToolResult> {
        if !self.security.can_act() {
            return Ok(ToolResult {
                success: false,
                output: String::new(),
                error: Some("Action blocked: autonomy is read-only".into()),
            });
        }

        if !self.security.record_action() {
            return Ok(ToolResult {
                success: false,
                output: String::new(),
                error: Some("Action blocked: rate limit exceeded".into()),
            });
        }

        let query = args
            .get("query")
            .and_then(|q| q.as_str())
            .ok_or_else(|| anyhow::anyhow!("Missing required parameter: query"))?;

        if query.trim().is_empty() {
            anyhow::bail!("Search query cannot be empty");
        }

        tracing::info!("Searching web for: {}", query);

        let result = match self.provider.as_str() {
            "duckduckgo" | "ddg" => self.search_duckduckgo(query).await?,
            "brave" => self.search_brave(query).await?,
            "firecrawl" => self.search_firecrawl(query).await?,
            _ => anyhow::bail!(
                "Unknown search provider: '{}'. Set tools.web_search.provider to 'duckduckgo', 'brave', or 'firecrawl' in config.toml",
                self.provider
            ),
        };

        Ok(ToolResult {
            success: true,
            output: result,
            error: None,
        })
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::security::{AutonomyLevel, SecurityPolicy};

    fn test_security() -> Arc<SecurityPolicy> {
        Arc::new(SecurityPolicy {
            autonomy: AutonomyLevel::Supervised,
            ..SecurityPolicy::default()
        })
    }

    #[test]
    fn test_tool_name() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        assert_eq!(tool.name(), "web_search_tool");
    }

    #[test]
    fn test_tool_description() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        assert!(tool.description().contains("Search the web"));
    }

    #[test]
    fn test_parameters_schema() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let schema = tool.parameters_schema();
        assert_eq!(schema["type"], "object");
        assert!(schema["properties"]["query"].is_object());
    }

    #[test]
    fn test_strip_tags() {
        let html = "<b>Hello</b> <i>World</i>";
        assert_eq!(strip_tags(html), "Hello World");
    }

    #[test]
    fn test_parse_duckduckgo_results_empty() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let result = tool
            .parse_duckduckgo_results("<html>No results here</html>", "test")
            .unwrap();
        assert!(result.contains("No results found"));
    }

    #[test]
    fn test_parse_duckduckgo_results_with_data() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let html = r#"
            <a class="result__a" href="https://example.com">Example Title</a>
            <a class="result__snippet">This is a description</a>
        "#;
        let result = tool.parse_duckduckgo_results(html, "test").unwrap();
        assert!(result.contains("Example Title"));
        assert!(result.contains("https://example.com"));
    }

    #[test]
    fn test_parse_duckduckgo_results_decodes_redirect_url() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let html = r#"
            <a class="result__a" href="https://duckduckgo.com/l/?uddg=https%3A%2F%2Fexample.com%2Fpath%3Fa%3D1&amp;rut=test">Example Title</a>
            <a class="result__snippet">This is a description</a>
        "#;
        let result = tool.parse_duckduckgo_results(html, "test").unwrap();
        assert!(result.contains("https://example.com/path?a=1"));
        assert!(!result.contains("rut=test"));
    }

    #[test]
    fn test_constructor_clamps_web_search_limits() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 0, 0);
        let html = r#"
            <a class="result__a" href="https://example.com">Example Title</a>
            <a class="result__snippet">This is a description</a>
        "#;
        let result = tool.parse_duckduckgo_results(html, "test").unwrap();
        assert!(result.contains("Example Title"));
    }

    #[tokio::test]
    async fn test_execute_missing_query() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let result = tool.execute(json!({})).await;
        assert!(result.is_err());
    }

    #[tokio::test]
    async fn test_execute_empty_query() {
        let tool = WebSearchTool::new(test_security(), "duckduckgo".to_string(), None, None, 5, 15);
        let result = tool.execute(json!({"query": ""})).await;
        assert!(result.is_err());
    }

    #[tokio::test]
    async fn test_execute_brave_without_api_key() {
        let tool = WebSearchTool::new(test_security(), "brave".to_string(), None, None, 5, 15);
        let result = tool.execute(json!({"query": "test"})).await;
        assert!(result.is_err());
        assert!(result.unwrap_err().to_string().contains("API key"));
    }

    #[tokio::test]
    async fn test_execute_firecrawl_without_api_key() {
        let tool = WebSearchTool::new(test_security(), "firecrawl".to_string(), None, None, 5, 15);
        let result = tool.execute(json!({"query": "test"})).await;
        assert!(result.is_err());
        let error = result.unwrap_err().to_string();
        if cfg!(feature = "firecrawl") {
            assert!(error.contains("api_key"));
        } else {
            assert!(error.contains("requires Cargo feature 'firecrawl'"));
        }
    }

    #[tokio::test]
    async fn test_execute_blocked_in_read_only_mode() {
        let security = Arc::new(SecurityPolicy {
            autonomy: AutonomyLevel::ReadOnly,
            ..SecurityPolicy::default()
        });
        let tool = WebSearchTool::new(security, "duckduckgo".to_string(), None, None, 5, 15);
        let result = tool.execute(json!({"query": "rust"})).await.unwrap();
        assert!(!result.success);
        assert!(result.error.unwrap().contains("read-only"));
    }
}