monothink 0.3.1

/*
* This file is part of monothink.
*
* monothink is free software: you can redistribute it and/or modify it under the terms
* of the GNU General Public License as published by the Free Software Foundation,
* either version 3 of the License, or (at your option) any later version.
*
* monothink is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
* without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
* See the GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along with Wavething.
* If not, see <https://www.gnu.org/licenses/>.

*/

use rig::{
    message::ToolResultContent,
    tool::{Tool, ToolContext},
};
use serde::{Deserialize, Serialize};
use serde_json::json;
use std::{
    fs::{self, read_to_string},
    io::ErrorKind,
    path::{Path, PathBuf},
};

use crate::llm::compactors::SharedObservations;

#[derive(Debug, Clone, PartialEq, Deserialize, Serialize)]
pub enum DirectoryEntryKind {
    File,
    Directory,
}

impl std::fmt::Display for DirectoryEntryKind {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            DirectoryEntryKind::File => {
                write!(f, "File")
            }
            DirectoryEntryKind::Directory => {
                write!(f, "Directory")
            }
        }
    }
}

#[derive(Clone, Debug, Deserialize, Serialize)]
struct DirectoryEntry {
    path: String,
    kind: DirectoryEntryKind,
}

#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct DirectoryListing {
    entries: Vec<DirectoryEntry>,
    total_entries: usize,
    truncated: bool,
}

#[derive(Clone, Debug, Deserialize, Serialize)]
struct Location {
    path: PathBuf,
    line: usize,
    //column: u32,
}

#[derive(Clone, Debug, Deserialize, Serialize)]
struct Range {
    path: PathBuf,
    start: usize,
    end: usize,
}

#[derive(Clone, Debug, Deserialize, Serialize)]
struct SearchMatch {
    location: Location,
    preview: String,
}

#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct SearchResults {
    total_entries: usize,
    returned_entries: usize,
    truncated: bool,
    entries: Vec<SearchMatch>,
    // @todo list entries that could not be accessed
    // failed_entries: Vec<String>
}

#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct FileContent {
    range: Range,
    content: String,
}

#[derive(Deserialize, Serialize)]
pub struct ListDirArgs {
    path: PathBuf,
    // recursive: Option<bool>,
    start: Option<usize>,
}

#[derive(Deserialize, Serialize)]
pub struct ListDir {
    base_dir: PathBuf,
    max_results: usize,
}

// @todo check access rights (i.e. calculated path must be below base_dir)
impl ListDir {
    pub fn new(base_dir: &Path, max_results: usize) -> Self {
        Self {
            base_dir: base_dir.to_path_buf(),
            max_results,
        }
    }
}
impl Tool for ListDir {
    const NAME: &'static str = "list_dir";

    type Error = ToolError;
    type Args = ListDirArgs;
    type Output = DirectoryListing;

    async fn call(
        &self,
        _context: &mut ToolContext,
        args: Self::Args,
    ) -> Result<Self::Output, Self::Error> {
        let path = allowed_canonical_path(&args.path, &self.base_dir)?;
        // list_dir(&path, args.recursive.unwrap_or(false))
        let mut result = list_dir(&path, false)?;
        // @todo clip long lists but preserve total entries (add actual entries as info)
        // @todo this is brickle, because it also includes permission denied errors, we should log the error and go on
        result.total_entries = result.entries.len();
        if result.entries.len() > self.max_results {
            result.entries.truncate(self.max_results);
            result.truncated = true;
        }

        Ok(result)
    }

    fn description(&self) -> String {
        // "List the files and directory of the given directory (relative to project root) and optionally its subdirectories"
        //
        format!(
            "List the files and directories in the given directory (relative to project root), \
      returning at most {} entries. Use an empty string \"\" for the project root. \
      Results include both files and subdirectories with their types.",
            self.max_results
        )
    }

    fn parameters(&self) -> serde_json::Value {
        json!({
            "type": "object",
            "properties": {
                "path": {
                    "type": "string",
                    "description": "Path to the directory relative to project root directory. \
                     Use an empty string for the root directory."
                },
                "start": {
                    "type": "integer",
                    "description": "Optional, zero-based index of the first result to return. \
                    Use this to paginate through results if a previous call was truncated."
                },
            },
            "required": ["path"]
        })
    }

    fn map_error(&self, error: Self::Error) -> rig::tool::ToolExecutionError {
        rig::tool::ToolExecutionError::from_error(error)
    }
}

#[derive(Deserialize, Serialize)]
pub struct ReadFileArgs {
    path: PathBuf,
    start: Option<usize>,
    end: Option<usize>,
}

#[derive(Deserialize, Serialize)]
pub struct ReadFile {
    base_dir: PathBuf,
}

impl ReadFile {
    pub fn new(base_dir: &Path) -> Self {
        Self {
            base_dir: base_dir.to_path_buf(),
        }
    }
}

// @todo check access rights (i.e. calculated path must be below base_dir)
impl Tool for ReadFile {
    const NAME: &'static str = "read_file";

    type Error = ToolError;
    type Args = ReadFileArgs;
    type Output = FileContent;

    async fn call(
        &self,
        _context: &mut ToolContext,
        args: Self::Args,
    ) -> Result<Self::Output, Self::Error> {
        let path = allowed_canonical_path(&args.path, &self.base_dir)?;

        if path.is_dir() {
            return Err(ToolError::NotAFile {
                path: args.path.to_string_lossy().to_string(),
                suggestion: "Use list_dir to explore this directory instead.".to_string(),
            });
        }

        let content: Vec<String> = fs::read_to_string(path)
            .map_err(|err| io_error_to_tool_error(err, &args.path))?
            .lines()
            .map(|line| line.to_string())
            .collect();

        let mut start = match args.start {
            Some(v) => v.saturating_sub(1),
            None => 0,
        };
        let mut end = match args.end {
            Some(v) => {
                if v > 0 {
                    v // we do not substract to convert from one-based to zero-based because the line shall be included which it isn't for slicing
                } else {
                    1 // LLMs like to provide 0 or -1 for end of file, we do not support -1 but can support 0
                }
            }
            None => content.len(),
        };
        if end > content.len() {
            end = content.len();
        }
        if start > end {
            start = end;
        };

        let extract = &content[start..end].join("\n");

        // @todo clip large files
        Ok(FileContent {
            range: Range {
                path: args.path,
                start: start + 1,
                end: end + 1,
            },
            content: extract.to_owned(),
        })
    }

    fn description(&self) -> String {
        "Read the content of a file, optionally specifying a line range. \
            Lines are 1-indexed and inclusive. \
            If no range is specified, the entire file is returned."
            .to_string()
    }

    fn parameters(&self) -> serde_json::Value {
        json!({
            "type": "object",
            "properties": {
                "path": {
                    "type": "string",
                    "description": "Path to the file relative to project root directory"
                },
                "start": {
                    "type": "integer",
                    "description": "First line number to read (1-indexed, inclusive). \
                    If not provided, reads from the beginning of the file."
                },
                "end": {
                    "type": "integer",
                    "description": "Last line number to read (1-indexed, inclusive). \
                    If not provided, reads to the end of the file. \
                      If end exceeds the file length, reads until the end."
                }
            },
            "required": ["path"]
        })
    }

    fn map_error(&self, error: Self::Error) -> rig::tool::ToolExecutionError {
        rig::tool::ToolExecutionError::from_error(error)
    }
}

#[derive(Deserialize, Serialize)]
pub struct SearchTextArgs {
    path: PathBuf,
    text: String,
    start: Option<usize>,
}

#[derive(Deserialize, Serialize)]
pub struct SearchText {
    base_dir: PathBuf,
    max_results: usize,
}

impl SearchText {
    pub fn new(base_dir: &Path, max_results: usize) -> Self {
        Self {
            base_dir: base_dir.to_path_buf(),
            max_results,
        }
    }
}

impl Tool for SearchText {
    const NAME: &'static str = "search_text";

    type Error = ToolError;
    type Args = SearchTextArgs;
    type Output = SearchResults;

    async fn call(
        &self,
        _context: &mut ToolContext,
        args: Self::Args,
    ) -> Result<Self::Output, Self::Error> {
        let path = allowed_canonical_path(&args.path, &self.base_dir)?;
        let entries = list_dir(&path, true)?;
        let matches: Vec<SearchMatch> = entries
            .entries
            .iter()
            // filter files
            .filter_map(|e| match e.kind {
                DirectoryEntryKind::File => {
                    read_to_string(&e.path).ok().map(|v| (e.path.clone(), v))
                }
                DirectoryEntryKind::Directory => None,
            })
            // map to lines that contain the searched text
            .map(|(path, content)| {
                (
                    path,
                    content
                        .lines()
                        .enumerate()
                        .filter(|(_nr, line)| line.contains(&args.text))
                        .map(|(nr, line)| (nr, line.to_owned()))
                        .collect::<Vec<(usize, String)>>(),
                )
            })
            // filter files containing the match
            .filter(|(_path, content)| !content.is_empty())
            // map to search matches
            .flat_map(|(path, content)| {
                content
                    .iter()
                    .map(move |(nr, line)| SearchMatch {
                        location: Location {
                            path: PathBuf::from(path.clone()),
                            line: *nr,
                        },
                        preview: line.clone(),
                    })
                    .collect::<Vec<_>>()
            })
            .collect();

        // clip
        let start = args.start.unwrap_or_default();
        let requested_matches = &matches[start..];
        let final_matches = if requested_matches.len() > 100 {
            &requested_matches[..100]
        } else {
            requested_matches
        };
        Ok(SearchResults {
            total_entries: matches.len(),
            returned_entries: final_matches.len(),
            truncated: final_matches.len() != matches.len(),
            entries: final_matches.to_vec(),
        })
    }

    fn description(&self) -> String {
        format!(
            "Search for text in all files within the given directory and its subdirectories. \
      Returns matches with file paths, line numbers, and content previews. \
      The search is case-sensitive. Results are clipped at {} entries.",
            self.max_results
        )
    }

    fn parameters(&self) -> serde_json::Value {
        json!({
            "type": "object",
            "properties": {
                "path": {
                    "type": "string",
                    "description": "Path to the directory to search in, \
                    relative to project root directory. \
                    Use an empty string for the root directory. \
                    Search is recursive."
                },
                "text": {
                    "type": "string",
                    "description": "The exact text string to search for. \
                    The search is case-sensitive."
                },
                "start": {
                    "type": "integer",
                    "description": "Optional, zero-based index of the first result to return. \
                    Use this to paginate through results if a previous call was truncated."
                },
            },
            "required": ["path", "text"]
        })
    }

    fn map_error(&self, error: Self::Error) -> rig::tool::ToolExecutionError {
        rig::tool::ToolExecutionError::from_error(error)
    }
}

#[derive(Deserialize, Serialize)]
pub struct RetrieveObservationArgs {
    id: u64,
}

#[derive(Debug)]
pub struct RetrieveObservation {
    pub observations: SharedObservations,
}

impl RetrieveObservation {
    pub fn new(observations: SharedObservations) -> Self {
        Self { observations }
    }
}

impl Tool for RetrieveObservation {
    const NAME: &'static str = "retrieve_observation";

    type Error = ToolError;
    type Args = RetrieveObservationArgs;
    type Output = Vec<ToolResultContent>;

    async fn call(
        &self,
        _context: &mut ToolContext,
        args: Self::Args,
    ) -> Result<Self::Output, Self::Error> {
        self.observations
            .lock()
            .unwrap()
            .get(&args.id)
            .ok_or_else(|| ToolError::ResourceNotFound {
                resource_type: ResourceType::Observation,
                identifier: args.id.to_string(),
            })
            .cloned()
    }

    fn description(&self) -> String {
        "Retrieve the full content of a masked tool result from conversation history. \
     When conversation history is compacted to save context space, older tool results \
     are replaced with placeholders like '[Observation:3 - masked]'. \
     Use this tool when you need to see the full content of a masked result. \
     The number inside the brackets (e.g., 3 in '[Observation:3 - masked]') is the \
     id you should pass to this tool. Each tool result in the history is assigned a \
     unique observation number starting from 1."
            .to_string()
    }

    fn parameters(&self) -> serde_json::Value {
        let id_description = "The numeric identifier of the observation to retrieve. \
        When a tool result is masked in the conversation history, it appears as \
        '[Observation:N - masked]' where N is this identifier. For example, \
        if you see '[Observation:3 - masked]', pass 3 as this parameter. \
        Observation IDs start at 1 and increment sequentially.";

        json!({
            "type": "object",
            "properties": {
                "id": {
                    "type": "integer",
                    "description": id_description
                 },
            },
            "required": ["id"]
        })
    }

    fn map_error(&self, error: Self::Error) -> rig::tool::ToolExecutionError {
        rig::tool::ToolExecutionError::from_error(error)
    }
}

// path is absolute
fn list_dir(path: &Path, recursive: bool) -> Result<DirectoryListing, ToolError> {
    let mut result = DirectoryListing {
        entries: Vec::new(),
        total_entries: 0,
        truncated: false,
    };
    let mut backlog = vec![path.to_path_buf()];

    while let Some(dir) = backlog.pop() {
        if !dir.is_dir() {
            return Result::Err(ToolError::DirectoryEntryNotFound {
                path: path.to_string_lossy().to_string(),
                kind: DirectoryEntryKind::Directory,
                suggestion: "Provide a directory to the list_dir tool.".to_string(),
            });
        }

        let paths = match fs::read_dir(dir) {
            Ok(paths) => paths,
            Err(err) if err.kind() == ErrorKind::PermissionDenied => {
                // @todo Log warning and skip this directory
                continue;
            }
            Err(err) => return Err(io_error_to_tool_error(err, path)),
        };

        for entry in paths.flatten() {
            if entry.path().is_dir() {
                if recursive {
                    backlog.push(entry.path());
                }
                result.entries.push(DirectoryEntry {
                    path: entry.path().to_string_lossy().to_string(),
                    kind: DirectoryEntryKind::Directory,
                });
            } else {
                result.entries.push(DirectoryEntry {
                    path: entry.path().to_string_lossy().to_string(),
                    kind: DirectoryEntryKind::File,
                });
            }
        }
    }

    result.total_entries = result.entries.len();

    Ok(result)
}

fn allowed_canonical_path(path_string: &Path, base_dir: &Path) -> Result<PathBuf, ToolError> {
    // let path_string = if path_string.is_empty() {
    //     "."
    // } else {
    //     path_string
    // };

    let canonical_base = base_dir
        .canonicalize()
        .map_err(|err| io_error_to_tool_error(err, path_string))?;
    let path = base_dir
        .join(path_string)
        .canonicalize()
        .map_err(|err| io_error_to_tool_error(err, path_string))?;

    if path.starts_with(canonical_base) {
        Ok(path)
    } else {
        Err(ToolError::PathTraversal {
            requested: path_string.to_string_lossy().to_string(),
            base: base_dir.to_string_lossy().to_string(),
        })
    }
}

fn io_error_to_tool_error(err: std::io::Error, path: &Path) -> ToolError {
    let path_string = path.to_string_lossy().to_string();
    match err.kind() {
        ErrorKind::NotFound => ToolError::DirectoryEntryNotFound {
            path: path_string,
            kind: DirectoryEntryKind::File,
            suggestion: "Use list_dir tool to explore available files.".to_string(),
        },
        ErrorKind::PermissionDenied => ToolError::PermissionDenied { path: path_string },
        _ => ToolError::Io {
            path: Some(path_string),
            source: err,
        },
    }
}

#[derive(Debug)]
pub enum ToolError {
    /// File or directory does not exist
    DirectoryEntryNotFound {
        path: String,
        kind: DirectoryEntryKind,
        suggestion: String,
    },
    /// Invalid arguments provided to the tool
    InvalidArgs {
        argument: &'static str,
        reason: String,
    },
    /// Entry is not a File
    NotAFile { path: String, suggestion: String },
    /// Path is outside the allowed base directory
    PathTraversal { requested: String, base: String },
    /// Permission denied accessing path
    PermissionDenied { path: String },
    /// Resource not found (e.g., observation ID)
    ResourceNotFound {
        resource_type: ResourceType,
        identifier: String,
    },
    /// Other IO error
    Io {
        path: Option<String>,
        source: std::io::Error,
    },
}

impl std::fmt::Display for ToolError {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            ToolError::DirectoryEntryNotFound {
                path,
                kind,
                suggestion,
            } => {
                write!(f, "Error: {} '{}' not found. {}", kind, path, suggestion)
            }
            ToolError::InvalidArgs { argument, reason } => {
                write!(
                    f,
                    "Error: Invalid value for argument '{}': {}",
                    argument, reason
                )
            }
            ToolError::NotAFile { path, suggestion } => {
                write!(f, "Error: '{}' is not a file. {}", path, suggestion)
            }
            ToolError::PathTraversal { requested, base } => {
                write!(
                    f,
                    "Error: Path '{}' is outside the allowed directory '{}'. \
                     Only paths within the project root are accessible.",
                    requested, base
                )
            }
            ToolError::PermissionDenied { path } => {
                write!(
                    f,
                    "Error: Permission denied when accessing '{}'. \
                     This file/directory requires elevated permissions. \
                     You cannot access this resource. Continue with available data \
                     or ask the user to check permissions.",
                    path
                )
            }
            ToolError::ResourceNotFound {
                resource_type,
                identifier,
            } => {
                write!(
                    f,
                    "Error: {} with identifier '{}' not found. \
                     Check the conversation history for valid {} IDs.",
                    resource_type, identifier, resource_type
                )
            }
            ToolError::Io { path, source } => {
                if let Some(p) = path {
                    write!(f, "Error reading '{}': {}", p, source)
                } else {
                    write!(f, "Error: {}", source)
                }
            }
        }
    }
}
impl std::error::Error for ToolError {}

#[derive(Debug, Clone)]
pub enum ResourceType {
    Observation,
}

impl std::fmt::Display for ResourceType {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            ResourceType::Observation => {
                write!(f, "Observation")
            }
        }
    }
}

#[tokio::test]
async fn test_list_dir() {
    // fake test to just read the current dir
    let base_dir = std::env::current_dir().expect("Cannot get current directory");
    let ld = ListDir::new(&base_dir, 100);
    let result = ld
        .call(
            &mut ToolContext::default(),
            ListDirArgs {
                path: PathBuf::from(""),
                // recursive: Some(true),
                start: None,
            },
        )
        .await
        .unwrap();
    println!("Result: {:?}", result.total_entries);
}

#[tokio::test]
async fn test_read_file() {
    // fake test to just see the output
    // Should output the GPLv3 license header of main.rs
    let base_dir = std::env::current_dir().expect("Cannot get current directory");
    let ld = ReadFile::new(&base_dir);
    let result = ld
        .call(
            &mut ToolContext::default(),
            ReadFileArgs {
                path: PathBuf::from("src/main.rs"),
                start: Some(1),
                end: Some(14),
            },
        )
        .await
        .unwrap();
    println!("Result: {:?}", result);
}

#[tokio::test]
async fn test_search_text() {
    // fake test to just see the serch results
    let base_dir = std::env::current_dir().expect("Cannot get current directory");
    let ld = SearchText::new(&base_dir, 100);
    let result = ld
        .call(
            &mut ToolContext::default(),
            SearchTextArgs {
                path: PathBuf::from("src"),
                text: "unwrap()".to_string(),
                start: None,
            },
        )
        .await
        .unwrap();
    println!("Result: {:?}", result);
}