ceres-search 0.4.0

use ceres_core::PortalType;
use clap::{Parser, Subcommand, ValueEnum};
use std::path::PathBuf;
use std::sync::LazyLock;

static VERSION_INFO: LazyLock<String> = LazyLock::new(|| {
    let version = env!("CARGO_PKG_VERSION");

    // Use VERGEN_GIT_SHA for the commit hash (with safe slicing)
    let commit = option_env!("VERGEN_GIT_SHA")
        .map(|s| s.chars().take(7).collect::<String>())
        .unwrap_or_else(|| "unknown".to_string());

    let built = option_env!("VERGEN_BUILD_DATE").unwrap_or("unknown"); // YYYY-MM-DD
    let target = option_env!("VERGEN_CARGO_TARGET_TRIPLE").unwrap_or("unknown");
    let rustc = option_env!("VERGEN_RUSTC_SEMVER").unwrap_or("unknown");

    format!("{version}\ncommit: {commit}\nbuilt: {built}\ntarget: {target}\nrustc: {rustc}")
});

pub fn version_info() -> &'static str {
    &VERSION_INFO
}

/// CLI configuration parsed from command line arguments and environment variables
#[derive(Parser, Debug)]
#[command(name = "ceres")]
#[command(
    author,
    version = version_info(),
    about = "Semantic search engine for open data portals"
)]
#[command(after_help = "Examples:
  ceres harvest https://dati.comune.milano.it
  ceres search \"air quality monitoring\" --limit 5
  ceres export --format jsonl > datasets.jsonl
  ceres stats

Embedding providers:
  EMBEDDING_PROVIDER=gemini (default) - Google Gemini (768 dimensions)
  EMBEDDING_PROVIDER=openai           - OpenAI (1536 or 3072 dimensions)
  EMBEDDING_PROVIDER=ollama           - Ollama local (768 dimensions, default: nomic-embed-text)")]
pub struct Config {
    /// PostgreSQL database connection URL
    #[arg(long, env = "DATABASE_URL")]
    pub database_url: String,

    /// Embedding provider to use: gemini (default), openai, or ollama
    #[arg(long, env = "EMBEDDING_PROVIDER", default_value = "gemini")]
    pub embedding_provider: String,

    /// Google Gemini API key (required when embedding_provider=gemini)
    #[arg(long, env = "GEMINI_API_KEY")]
    pub gemini_api_key: Option<String>,

    /// OpenAI API key (required when embedding_provider=openai)
    #[arg(long, env = "OPENAI_API_KEY")]
    pub openai_api_key: Option<String>,

    /// Embedding model name (provider-specific, uses default if not set)
    #[arg(long, env = "EMBEDDING_MODEL")]
    pub embedding_model: Option<String>,

    /// Ollama API endpoint (default: http://localhost:11434)
    #[arg(long, env = "OLLAMA_ENDPOINT")]
    pub ollama_endpoint: Option<String>,

    #[command(subcommand)]
    pub command: Command,
}

/// Available CLI commands
#[derive(Subcommand, Debug)]
pub enum Command {
    /// Harvest datasets from open data portals
    #[command(after_help = "Examples:
  ceres harvest                                       # Harvest all enabled portals from config
  ceres harvest https://dati.comune.milano.it         # Harvest single CKAN URL (default type)
  ceres harvest https://data.public.lu --type dcat    # Harvest DCAT portal by URL
  ceres harvest https://data.europa.eu --type dcat --profile sparql  # Harvest SPARQL DCAT endpoint
  ceres harvest --portal milano                       # Harvest portal by name from config
  ceres harvest --config ~/custom.toml                # Use custom config file
  ceres harvest --full-sync                           # Force full sync even if incremental is available")]
    Harvest {
        /// URL of a portal to harvest (ad-hoc, not from config)
        #[arg(value_name = "URL")]
        portal_url: Option<String>,

        /// Portal type when harvesting an ad-hoc URL (ignored when using --portal or batch mode)
        #[arg(
            long,
            value_name = "TYPE",
            default_value = "ckan",
            requires = "portal_url"
        )]
        r#type: PortalType,

        /// DCAT profile when harvesting an ad-hoc DCAT URL (e.g., "sparql")
        #[arg(long, value_name = "PROFILE", requires = "portal_url")]
        profile: Option<String>,

        /// Harvest a specific portal by name from config file
        #[arg(short, long, value_name = "NAME", conflicts_with = "portal_url")]
        portal: Option<String>,

        /// Custom path to portals.toml configuration file
        #[arg(short, long, value_name = "PATH")]
        config: Option<PathBuf>,

        /// Force full sync even if incremental sync is available
        #[arg(long)]
        full_sync: bool,

        /// Preview what would be harvested without writing to DB or calling embedding API
        #[arg(long)]
        dry_run: bool,

        /// Only harvest metadata (no embedding). Does not require an API key.
        #[arg(long)]
        metadata_only: bool,
    },
    /// Generate embeddings for datasets that don't have them yet
    #[command(after_help = "Examples:
  ceres embed                                     # Embed all pending datasets
  ceres embed --portal https://dati.comune.milano.it  # Embed pending from one portal")]
    Embed {
        /// Filter to a specific portal URL
        #[arg(short, long)]
        portal: Option<String>,
    },
    /// Search indexed datasets using semantic similarity
    #[command(after_help = "Example: ceres search \"trasporto pubblico\" --limit 10")]
    Search {
        /// Search query text
        query: String,
        /// Maximum number of results to return
        #[arg(short, long, default_value = "10")]
        limit: usize,
    },
    /// Export indexed datasets to various formats
    #[command(after_help = "Examples:
  ceres export --format jsonl > datasets.jsonl
  ceres export --format json --portal https://dati.gov.it
  ceres export --format parquet --output ./ceres-export")]
    Export {
        /// Output format for exported data
        #[arg(short, long, default_value = "jsonl")]
        format: ExportFormat,
        /// Filter by source portal URL
        #[arg(short, long)]
        portal: Option<String>,
        /// Maximum number of datasets to export
        #[arg(short, long)]
        limit: Option<usize>,
        /// Output directory (required for parquet format)
        #[arg(short, long, value_name = "DIR")]
        output: Option<std::path::PathBuf>,
        /// Custom path to portals.toml (for portal name resolution in parquet export)
        #[arg(short, long, value_name = "PATH")]
        config: Option<std::path::PathBuf>,
    },
    /// Show database statistics
    Stats,
}

/// Supported export formats
#[derive(Debug, Clone, ValueEnum)]
pub enum ExportFormat {
    /// JSON Lines format (one JSON object per line)
    Jsonl,
    /// Standard JSON array format
    Json,
    /// CSV format (comma-separated values)
    Csv,
    /// Parquet format (curated, flattened for HuggingFace)
    Parquet,
}

#[cfg(test)]
mod tests {
    use super::version_info;

    #[test]
    fn test_version_info_contains_expected_fields() {
        let info = version_info();
        assert!(info.contains("commit:"));
        assert!(info.contains("built:"));
        assert!(info.contains("target:"));
        assert!(info.contains("rustc:"));
    }
}