oxirs_core/parser/

mod.rs

//! RDF parsing utilities for various formats with high-performance streaming
//!
//! **Stability**: ✅ **Stable** - Core parser APIs are production-ready.
//!
//! This module provides parsers for all major RDF serialization formats:
//! - **Turtle** (.ttl) - A compact, human-readable format
//! - **N-Triples** (.nt) - Line-based triple format
//! - **TriG** (.trig) - Turtle with named graphs
//! - **N-Quads** (.nq) - Line-based quad format
//! - **RDF/XML** (.rdf, .xml) - XML-based format
//! - **JSON-LD** (.jsonld) - JSON-based linked data format
//!
//! ## Features
//!
//! - **Streaming parsers** - Process large files without loading into memory
//! - **Error recovery** - Continue parsing after encountering errors (optional)
//! - **Base IRI resolution** - Resolve relative IRIs against a base
//! - **Format detection** - Automatic format detection from file extensions or content
//! - **Async support** - Non-blocking I/O for high-throughput applications
//!
//! ## Examples
//!
//! ### Basic Parsing
//!
//! ```rust
//! use oxirs_core::parser::{Parser, RdfFormat};
//!
//! # fn main() -> Result<(), Box<dyn std::error::Error>> {
//! let turtle_data = r#"
//!     @prefix foaf: <http://xmlns.com/foaf/0.1/> .
//!
//!     <http://example.org/alice> foaf:name "Alice" ;
//!                                 foaf:knows <http://example.org/bob> .
//! "#;
//!
//! let parser = Parser::new(RdfFormat::Turtle);
//! let quads = parser.parse_str_to_quads(turtle_data)?;
//!
//! println!("Parsed {} quads", quads.len());
//! # Ok(())
//! # }
//! ```
//!
//! ### Parsing with Configuration
//!
//! ```rust,ignore
//! use oxirs_core::parser::{Parser, RdfFormat, ParserConfig};
//!
//! # fn main() -> Result<(), Box<dyn std::error::Error>> {
//! let config = ParserConfig {
//!     base_iri: Some("http://example.org/base/".to_string()),
//!     ignore_errors: true,
//!     max_errors: Some(10),
//! };
//!
//! let parser = Parser::new(RdfFormat::Turtle).with_config(config);
//! let quads = parser.parse_str_to_quads("<relative> <p> <o> .")?;
//! # Ok(())
//! # }
//! ```
//!
//! ### Format Detection
//!
//! ```rust,ignore
//! use oxirs_core::parser::RdfFormat;
//!
//! // Detect from file extension
//! let format = RdfFormat::from_extension("ttl");
//! assert_eq!(format, Some(RdfFormat::Turtle));
//!
//! // Check format capabilities
//! assert!(!RdfFormat::Turtle.supports_quads());
//! assert!(RdfFormat::TriG.supports_quads());
//! ```
//!
//! ### Streaming Large Files
//!
//! The high-level [`Parser`] in this module only exposes in-memory
//! string/byte entry points ([`Parser::parse_str_to_quads`],
//! [`Parser::parse_bytes_to_quads`], [`Parser::parse_str_with_handler`]) and
//! always materializes the whole document before parsing, for every format.
//! For genuine bounded-memory streaming from a [`std::io::Read`] source, use
//! the lower-level [`crate::format::RdfParser`] instead, which reads
//! incrementally for all six supported formats:
//!
//! ```rust,no_run
//! use oxirs_core::format::{RdfParser, RdfFormat};
//! use std::fs::File;
//! use std::io::BufReader;
//!
//! # fn main() -> Result<(), Box<dyn std::error::Error>> {
//! let file = File::open("large_dataset.nt")?;
//! let reader = BufReader::new(file);
//!
//! let parser = RdfParser::new(RdfFormat::NTriples);
//! for quad in parser.for_reader(reader) {
//!     let quad = quad?;
//!     // Process quad without loading the entire file into memory
//! }
//! # Ok(())
//! # }
//! ```
//!
//! ### Async Parsing (with `async` feature)
//!
//! ```rust,no_run
//! # #[cfg(feature = "async")]
//! use oxirs_core::parser::{AsyncStreamingParser, RdfFormat};
//!
//! # #[cfg(feature = "async")]
//! # async fn example() -> Result<(), Box<dyn std::error::Error>> {
//! let parser = AsyncStreamingParser::new(RdfFormat::Turtle);
//! let mut quads = Vec::new();
//! parser.parse_stream(tokio::io::stdin(), |quad| {
//!     quads.push(quad);
//!     async { Ok(()) }
//! }).await?;
//! # Ok(())
//! # }
//! ```
//!
//! ## Performance Tips
//!
//! 1. **Use streaming** - For large files, use [`crate::format::RdfParser::for_reader`]
//!    (not the [`Parser`] in this module, which always buffers the whole document)
//!    to avoid loading everything into memory
//! 2. **Choose the right format** - N-Triples/N-Quads are fastest to parse (line-based)
//! 3. **Enable async** - [`AsyncStreamingParser`] truly streams N-Triples/N-Quads
//!    incrementally; other formats are buffered up to a configurable limit
//!    (see [`AsyncStreamingParser::with_max_buffer_size`]) before parsing
//! 4. **Batch processing** - Process multiple files in parallel using rayon
//!
//! ## Error Handling
//!
//! Parsers can be configured to handle errors in different ways:
//!
//! - **Strict mode** (default) - Stop on first error
//! - **Error recovery** - Collect errors and continue parsing
//! - **Max errors** - Stop after a threshold of errors
//!
//! ## Format Support Matrix
//!
//! "Streaming" below means [`crate::format::RdfParser::for_reader`] (bounded
//! memory, reads incrementally). The [`Parser`] in *this* module always
//! buffers the whole input first regardless of format; [`AsyncStreamingParser`]
//! truly streams only N-Triples/N-Quads (see the Performance Tips above).
//!
//! | Format | Triples | Quads | Prefixes | Comments | Streaming |
//! |--------|---------|-------|----------|----------|-----------|
//! | Turtle | ✅ | ❌ | ✅ | ✅ | ✅ |
//! | N-Triples | ✅ | ❌ | ❌ | ✅ | ✅ |
//! | TriG | ✅ | ✅ | ✅ | ✅ | ✅ |
//! | N-Quads | ✅ | ✅ | ❌ | ✅ | ✅ |
//! | RDF/XML | ✅ | ❌ | ✅ | ✅ | ✅ |
//! | JSON-LD | ✅ | ✅ | ✅ | ❌ | ✅ |
//!
//! ## Related Modules
//!
//! - [`crate::serializer`] - Serialize RDF to various formats
//! - [`crate::model`] - RDF data model types
//! - [`crate::rdf_store`] - Store parsed RDF data

#[cfg(feature = "async")]
mod async_parser;

#[cfg(feature = "async")]
pub use async_parser::{AsyncRdfSink, AsyncStreamingParser, MemoryAsyncSink, ParseProgress};

// Native implementation - no external dependencies needed
use crate::model::{
    BlankNode, GraphName, Literal, NamedNode, Object, Predicate, Quad, Subject, Triple,
};
use crate::{OxirsError, Result};

/// RDF format enumeration
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
pub enum RdfFormat {
    /// Turtle format (TTL)
    Turtle,
    /// N-Triples format (NT)
    NTriples,
    /// TriG format (named graphs)
    TriG,
    /// N-Quads format
    NQuads,
    /// RDF/XML format
    RdfXml,
    /// JSON-LD format
    JsonLd,
}

impl RdfFormat {
    /// Detect format from file extension
    pub fn from_extension(ext: &str) -> Option<Self> {
        match ext.to_lowercase().as_str() {
            "ttl" | "turtle" => Some(RdfFormat::Turtle),
            "nt" | "ntriples" => Some(RdfFormat::NTriples),
            "trig" => Some(RdfFormat::TriG),
            "nq" | "nquads" => Some(RdfFormat::NQuads),
            "rdf" | "xml" | "rdfxml" => Some(RdfFormat::RdfXml),
            "jsonld" | "json-ld" => Some(RdfFormat::JsonLd),
            _ => None,
        }
    }

    /// Get the media type for this format
    pub fn media_type(&self) -> &'static str {
        match self {
            RdfFormat::Turtle => "text/turtle",
            RdfFormat::NTriples => "application/n-triples",
            RdfFormat::TriG => "application/trig",
            RdfFormat::NQuads => "application/n-quads",
            RdfFormat::RdfXml => "application/rdf+xml",
            RdfFormat::JsonLd => "application/ld+json",
        }
    }

    /// Get file extension for this format
    pub fn extension(&self) -> &'static str {
        match self {
            RdfFormat::Turtle => "ttl",
            RdfFormat::NTriples => "nt",
            RdfFormat::TriG => "trig",
            RdfFormat::NQuads => "nq",
            RdfFormat::RdfXml => "rdf",
            RdfFormat::JsonLd => "jsonld",
        }
    }

    /// Returns true if this format supports named graphs (quads)
    pub fn supports_quads(&self) -> bool {
        matches!(self, RdfFormat::TriG | RdfFormat::NQuads)
    }
}

/// Configuration for RDF parsing
#[derive(Debug, Clone, Default)]
pub struct ParserConfig {
    /// Base IRI for resolving relative IRIs
    pub base_iri: Option<String>,
    /// Whether to ignore parsing errors and continue
    pub ignore_errors: bool,
    /// Maximum number of errors to collect before stopping
    pub max_errors: Option<usize>,
}

/// RDF parser interface
#[derive(Debug, Clone)]
pub struct Parser {
    format: RdfFormat,
    config: ParserConfig,
}

impl Parser {
    /// Create a new parser for the specified format
    pub fn new(format: RdfFormat) -> Self {
        Parser {
            format,
            config: ParserConfig::default(),
        }
    }

    /// Create a parser with custom configuration
    pub fn with_config(format: RdfFormat, config: ParserConfig) -> Self {
        Parser { format, config }
    }

    /// Set the base IRI for resolving relative IRIs
    pub fn with_base_iri(mut self, base_iri: impl Into<String>) -> Self {
        self.config.base_iri = Some(base_iri.into());
        self
    }

    /// Enable or disable error tolerance
    pub fn with_error_tolerance(mut self, ignore_errors: bool) -> Self {
        self.config.ignore_errors = ignore_errors;
        self
    }

    /// Parse RDF data from a string into a vector of quads
    pub fn parse_str_to_quads(&self, data: &str) -> Result<Vec<Quad>> {
        let mut quads = Vec::new();
        self.parse_str_with_handler(data, |quad| {
            quads.push(quad);
            Ok(())
        })?;
        Ok(quads)
    }

    /// Parse RDF data from a string into a vector of triples (only default graph)
    pub fn parse_str_to_triples(&self, data: &str) -> Result<Vec<Triple>> {
        let quads = self.parse_str_to_quads(data)?;
        Ok(quads
            .into_iter()
            .filter(|quad| quad.is_default_graph())
            .map(|quad| quad.to_triple())
            .collect())
    }

    /// Parse RDF data with a custom handler for each quad
    pub fn parse_str_with_handler<F>(&self, data: &str, handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        match self.format {
            RdfFormat::Turtle => self.parse_turtle(data, handler),
            RdfFormat::NTriples => self.parse_ntriples(data, handler),
            RdfFormat::TriG => self.parse_trig(data, handler),
            RdfFormat::NQuads => self.parse_nquads(data, handler),
            RdfFormat::RdfXml => self.parse_rdfxml(data, handler),
            RdfFormat::JsonLd => self.parse_jsonld(data, handler),
        }
    }

    /// Parse RDF data from bytes
    pub fn parse_bytes_to_quads(&self, data: &[u8]) -> Result<Vec<Quad>> {
        let data_str = std::str::from_utf8(data)
            .map_err(|e| OxirsError::Parse(format!("Invalid UTF-8: {e}")))?;
        self.parse_str_to_quads(data_str)
    }

    fn parse_turtle<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        // Delegate to the real, oxttl-backed grammar (crate::format::turtle
        // via crate::format::RdfParser) instead of a hand-rolled line-based
        // state machine. This correctly handles semicolons/commas inside
        // quoted literals, comma object lists, collections `( ... )`,
        // blank-node property lists `[ ... ]`, and multi-line triple-quoted
        // string literals -- none of which a per-line splitter can support.
        let mut internal_parser = crate::format::RdfParser::new(crate::format::RdfFormat::Turtle);
        if let Some(base) = &self.config.base_iri {
            internal_parser = internal_parser.with_base_iri(base.clone());
        }

        for result in internal_parser.for_slice(data.as_bytes()) {
            match result {
                Ok(quad) => handler(quad)?,
                Err(e) => {
                    if self.config.ignore_errors {
                        tracing::warn!("Turtle parse error: {e}");
                        continue;
                    } else {
                        return Err(OxirsError::Parse(format!("Turtle parse error: {e}")));
                    }
                }
            }
        }

        Ok(())
    }

    fn parse_ntriples<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        for (line_num, line) in data.lines().enumerate() {
            let line = line.trim();

            // Skip empty lines and comments
            if line.is_empty() || line.starts_with('#') {
                continue;
            }

            // Parse the line into a triple
            match self.parse_ntriples_line(line) {
                Ok(Some(quad)) => {
                    handler(quad)?;
                }
                Ok(None) => {
                    // Skip this line (e.g., blank line)
                    continue;
                }
                Err(e) => {
                    if self.config.ignore_errors {
                        tracing::warn!("Parse error on line {}: {}", line_num + 1, e);
                        continue;
                    } else {
                        return Err(OxirsError::Parse(format!(
                            "Parse error on line {}: {}",
                            line_num + 1,
                            e
                        )));
                    }
                }
            }
        }

        Ok(())
    }

    pub fn parse_ntriples_line(&self, line: &str) -> Result<Option<Quad>> {
        // Simple N-Triples parser - parse line like: <s> <p> "o" .
        let line = line.trim();

        if line.is_empty() || line.starts_with('#') {
            return Ok(None);
        }

        // Find the final period
        if !line.ends_with('.') {
            return Err(OxirsError::Parse("Line must end with '.'".to_string()));
        }

        let line = &line[..line.len() - 1].trim(); // Remove trailing period and whitespace

        // Split into tokens respecting quoted strings
        let tokens = self.tokenize_ntriples_line(line)?;

        if tokens.len() != 3 {
            return Err(OxirsError::Parse(format!(
                "Expected 3 tokens (subject, predicate, object), found {}",
                tokens.len()
            )));
        }

        // Parse subject
        let subject = self.parse_subject(&tokens[0])?;

        // Parse predicate
        let predicate = self.parse_predicate(&tokens[1])?;

        // Parse object
        let object = self.parse_object(&tokens[2])?;

        let triple = Triple::new(subject, predicate, object);
        let quad = Quad::from_triple(triple);

        Ok(Some(quad))
    }

    fn tokenize_ntriples_line(&self, line: &str) -> Result<Vec<String>> {
        let mut tokens = Vec::new();
        let mut current_token = String::new();
        let mut in_quotes = false;
        let mut escaped = false;
        let mut chars = line.chars().peekable();

        while let Some(c) = chars.next() {
            if escaped {
                // Preserve escape sequences - don't unescape during tokenization
                current_token.push('\\');
                current_token.push(c);
                escaped = false;
            } else if c == '\\' && in_quotes {
                escaped = true;
            } else if c == '"' && !escaped {
                current_token.push(c);
                if in_quotes {
                    // Check for language tag or datatype after closing quote
                    if let Some(&'@') = chars.peek() {
                        // Language tag
                        current_token.push(chars.next().expect("peeked '@' should be available"));
                        while let Some(&next_char) = chars.peek() {
                            if next_char.is_alphanumeric() || next_char == '-' {
                                current_token
                                    .push(chars.next().expect("peeked char should be available"));
                            } else {
                                break;
                            }
                        }
                    } else if chars.peek() == Some(&'^') {
                        // Datatype
                        chars.next(); // first ^
                        if chars.peek() == Some(&'^') {
                            chars.next(); // second ^
                            current_token.push_str("^^");
                            if chars.peek() == Some(&'<') {
                                // IRI datatype
                                for next_char in chars.by_ref() {
                                    current_token.push(next_char);
                                    if next_char == '>' {
                                        break;
                                    }
                                }
                            }
                        }
                    }
                    in_quotes = false;
                } else {
                    in_quotes = true;
                }
            } else if c == '"' && escaped {
                // This is an escaped quote, add it to the token
                current_token.push(c);
                escaped = false;
            } else if c.is_whitespace() && !in_quotes {
                if !current_token.is_empty() {
                    tokens.push(current_token.clone());
                    current_token.clear();
                }
            } else {
                current_token.push(c);
            }
        }

        if !current_token.is_empty() {
            tokens.push(current_token);
        }

        Ok(tokens)
    }

    fn parse_subject(&self, token: &str) -> Result<Subject> {
        if token.starts_with('<') && token.ends_with('>') {
            let iri = &token[1..token.len() - 1];
            let named_node = NamedNode::new(iri)?;
            Ok(Subject::NamedNode(named_node))
        } else if token.starts_with("_:") {
            let blank_node = BlankNode::new(token)?;
            Ok(Subject::BlankNode(blank_node))
        } else {
            Err(OxirsError::Parse(format!(
                "Invalid subject: {token}. Must be IRI or blank node"
            )))
        }
    }

    fn parse_predicate(&self, token: &str) -> Result<Predicate> {
        if token.starts_with('<') && token.ends_with('>') {
            let iri = &token[1..token.len() - 1];
            let named_node = NamedNode::new(iri)?;
            Ok(Predicate::NamedNode(named_node))
        } else {
            Err(OxirsError::Parse(format!(
                "Invalid predicate: {token}. Must be IRI"
            )))
        }
    }

    fn parse_object(&self, token: &str) -> Result<Object> {
        if token.starts_with('<') && token.ends_with('>') {
            // IRI
            let iri = &token[1..token.len() - 1];
            let named_node = NamedNode::new(iri)?;
            Ok(Object::NamedNode(named_node))
        } else if token.starts_with("_:") {
            // Blank node
            let blank_node = BlankNode::new(token)?;
            Ok(Object::BlankNode(blank_node))
        } else if token.starts_with('"') {
            // Literal
            self.parse_literal(token)
        } else {
            Err(OxirsError::Parse(format!(
                "Invalid object: {token}. Must be IRI, blank node, or literal"
            )))
        }
    }

    fn parse_literal(&self, token: &str) -> Result<Object> {
        if !token.starts_with('"') {
            return Err(OxirsError::Parse(
                "Literal must start with quote".to_string(),
            ));
        }

        // Find the closing quote. We scan via `char_indices` so that
        // `end_quote_pos` is a *byte* offset into `token` from the start —
        // matching the byte-indexed slicing below. The previous
        // implementation collected into `Vec<char>` and used the resulting
        // *char* index to slice the original `&str`, which panics on any
        // literal containing multi-byte UTF-8 characters (e.g. Japanese
        // text, emoji, accented Latin) once the char index diverges from
        // the byte offset.
        let mut end_quote_pos = None;
        let mut escaped = false;

        for (i, ch) in token.char_indices().skip(1) {
            if escaped {
                escaped = false;
                continue;
            }

            if ch == '\\' {
                escaped = true;
            } else if ch == '"' {
                end_quote_pos = Some(i);
                break;
            }
        }

        let end_quote_pos =
            end_quote_pos.ok_or_else(|| OxirsError::Parse("Unterminated literal".to_string()))?;

        // Extract the literal value (without quotes) and unescape.
        // `1` and `end_quote_pos` are both valid char-boundary byte offsets
        // (the opening quote and closing quote are each a single ASCII
        // byte), so this slice is safe even when the value itself contains
        // multi-byte characters.
        let raw_value = &token[1..end_quote_pos];
        let literal_value = self.unescape_literal_value(raw_value)?;

        // Check for language tag or datatype
        let remaining = &token[end_quote_pos + 1..];

        if let Some(lang_tag) = remaining.strip_prefix('@') {
            // Language tag
            let literal = Literal::new_lang(literal_value, lang_tag)?;
            Ok(Object::Literal(literal))
        } else if remaining.starts_with("^^<") && remaining.ends_with('>') {
            // Datatype
            let datatype_iri = &remaining[3..remaining.len() - 1];
            let datatype = NamedNode::new(datatype_iri)?;
            let literal = Literal::new_typed(literal_value, datatype);
            Ok(Object::Literal(literal))
        } else if remaining.is_empty() {
            // Plain literal
            let literal = Literal::new(literal_value);
            Ok(Object::Literal(literal))
        } else {
            Err(OxirsError::Parse(format!(
                "Invalid literal syntax: {token}"
            )))
        }
    }

    fn parse_trig<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        // Delegate to the real, oxttl-backed TriG grammar (same rationale as
        // parse_turtle above: named-graph blocks, nested Turtle syntax, and
        // multi-line statements are not safely handled by line splitting).
        let mut internal_parser = crate::format::RdfParser::new(crate::format::RdfFormat::TriG);
        if let Some(base) = &self.config.base_iri {
            internal_parser = internal_parser.with_base_iri(base.clone());
        }

        for result in internal_parser.for_slice(data.as_bytes()) {
            match result {
                Ok(quad) => handler(quad)?,
                Err(e) => {
                    if self.config.ignore_errors {
                        tracing::warn!("TriG parse error: {e}");
                        continue;
                    } else {
                        return Err(OxirsError::Parse(format!("TriG parse error: {e}")));
                    }
                }
            }
        }

        Ok(())
    }

    fn parse_nquads<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        for (line_num, line) in data.lines().enumerate() {
            let line = line.trim();

            // Skip empty lines and comments
            if line.is_empty() || line.starts_with('#') {
                continue;
            }

            // Parse the line into a quad
            match self.parse_nquads_line(line) {
                Ok(Some(quad)) => {
                    handler(quad)?;
                }
                Ok(None) => {
                    // Skip this line (e.g., blank line)
                    continue;
                }
                Err(e) => {
                    if self.config.ignore_errors {
                        tracing::warn!("Parse error on line {}: {}", line_num + 1, e);
                        continue;
                    } else {
                        return Err(OxirsError::Parse(format!(
                            "Parse error on line {}: {}",
                            line_num + 1,
                            e
                        )));
                    }
                }
            }
        }

        Ok(())
    }

    pub fn parse_nquads_line(&self, line: &str) -> Result<Option<Quad>> {
        // N-Quads parser - parse line like: <s> <p> "o" <g> .
        let line = line.trim();

        if line.is_empty() || line.starts_with('#') {
            return Ok(None);
        }

        // Find the final period
        if !line.ends_with('.') {
            return Err(OxirsError::Parse("Line must end with '.'".to_string()));
        }

        let line = &line[..line.len() - 1].trim(); // Remove trailing period and whitespace

        // Split into tokens respecting quoted strings
        let tokens = self.tokenize_ntriples_line(line)?;

        if tokens.len() != 4 {
            return Err(OxirsError::Parse(format!(
                "Expected 4 tokens (subject, predicate, object, graph), found {}",
                tokens.len()
            )));
        }

        // Parse subject
        let subject = self.parse_subject(&tokens[0])?;

        // Parse predicate
        let predicate = self.parse_predicate(&tokens[1])?;

        // Parse object
        let object = self.parse_object(&tokens[2])?;

        // Parse graph name
        let graph_name = self.parse_graph_name(&tokens[3])?;

        let quad = Quad::new(subject, predicate, object, graph_name);

        Ok(Some(quad))
    }

    fn parse_graph_name(&self, token: &str) -> Result<GraphName> {
        if token.starts_with('<') && token.ends_with('>') {
            let iri = &token[1..token.len() - 1];
            let named_node = NamedNode::new(iri)?;
            Ok(GraphName::NamedNode(named_node))
        } else if token.starts_with("_:") {
            let blank_node = BlankNode::new(token)?;
            Ok(GraphName::BlankNode(blank_node))
        } else {
            Err(OxirsError::Parse(format!(
                "Invalid graph name: {token}. Must be IRI or blank node"
            )))
        }
    }

    fn parse_rdfxml<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        use crate::rdfxml::wrapper::parse_rdfxml;
        use std::io::Cursor;

        // Parse RDF/XML data using the wrapper
        let reader = Cursor::new(data.as_bytes());
        let base_iri = self.config.base_iri.as_deref();
        let quads = parse_rdfxml(reader, base_iri, self.config.ignore_errors)?;

        // Process each quad through the handler
        for quad in quads {
            handler(quad)?;
        }

        Ok(())
    }

    fn parse_jsonld<F>(&self, data: &str, mut handler: F) -> Result<()>
    where
        F: FnMut(Quad) -> Result<()>,
    {
        // Basic JSON-LD parser implementation using existing jsonld module
        use crate::jsonld::to_rdf::JsonLdParser;

        let parser = JsonLdParser::new();
        let parser = if let Some(base_iri) = &self.config.base_iri {
            parser
                .with_base_iri(base_iri.clone())
                .map_err(|e| OxirsError::Parse(format!("Invalid base IRI: {e}")))?
        } else {
            parser
        };

        // Parse JSON-LD data into quads
        for result in parser.for_slice(data.as_bytes()) {
            match result {
                Ok(quad) => handler(quad)?,
                Err(e) => {
                    if self.config.ignore_errors {
                        tracing::warn!("JSON-LD parse error: {}", e);
                        continue;
                    } else {
                        return Err(OxirsError::Parse(format!("JSON-LD parse error: {e}")));
                    }
                }
            }
        }

        Ok(())
    }

    /// Unescape special characters in literal values
    fn unescape_literal_value(&self, value: &str) -> Result<String> {
        let mut result = String::new();
        let mut chars = value.chars();

        while let Some(c) = chars.next() {
            if c == '\\' {
                match chars.next() {
                    Some('"') => result.push('"'),
                    Some('\\') => result.push('\\'),
                    Some('n') => result.push('\n'),
                    Some('r') => result.push('\r'),
                    Some('t') => result.push('\t'),
                    Some('u') => {
                        // Parse \uHHHH Unicode escape
                        let hex_chars: String = chars.by_ref().take(4).collect();
                        if hex_chars.len() != 4 {
                            return Err(OxirsError::Parse(
                                "Invalid Unicode escape sequence \\uHHHH - expected 4 hex digits"
                                    .to_string(),
                            ));
                        }
                        let code_point = u32::from_str_radix(&hex_chars, 16).map_err(|_| {
                            OxirsError::Parse(
                                "Invalid hex digits in Unicode escape sequence".to_string(),
                            )
                        })?;
                        let unicode_char = char::from_u32(code_point).ok_or_else(|| {
                            OxirsError::Parse("Invalid Unicode code point".to_string())
                        })?;
                        result.push(unicode_char);
                    }
                    Some('U') => {
                        // Parse \UHHHHHHHH Unicode escape
                        let hex_chars: String = chars.by_ref().take(8).collect();
                        if hex_chars.len() != 8 {
                            return Err(OxirsError::Parse(
                                "Invalid Unicode escape sequence \\UHHHHHHHH - expected 8 hex digits".to_string()
                            ));
                        }
                        let code_point = u32::from_str_radix(&hex_chars, 16).map_err(|_| {
                            OxirsError::Parse(
                                "Invalid hex digits in Unicode escape sequence".to_string(),
                            )
                        })?;
                        let unicode_char = char::from_u32(code_point).ok_or_else(|| {
                            OxirsError::Parse("Invalid Unicode code point".to_string())
                        })?;
                        result.push(unicode_char);
                    }
                    Some(other) => {
                        return Err(OxirsError::Parse(format!(
                            "Invalid escape sequence \\{other}"
                        )));
                    }
                    None => {
                        return Err(OxirsError::Parse(
                            "Incomplete escape sequence at end of literal".to_string(),
                        ));
                    }
                }
            } else {
                result.push(c);
            }
        }

        Ok(result)
    }

    // Native parsing implementation complete - no external dependencies needed
}

/// Convenience function to detect RDF format from content
pub fn detect_format_from_content(content: &str) -> Option<RdfFormat> {
    let content = content.trim();

    // Check for XML-like content (RDF/XML)
    if content.starts_with("<?xml")
        || content.starts_with("<rdf:RDF")
        || content.starts_with("<RDF")
    {
        return Some(RdfFormat::RdfXml);
    }

    // Check for JSON-LD
    if content.starts_with('{') && (content.contains("@context") || content.contains("@type")) {
        return Some(RdfFormat::JsonLd);
    }

    // Check for Turtle syntax elements first (has priority over N-Quads/N-Triples)
    if content.contains("@prefix") || content.contains("@base") || content.contains(';') {
        return Some(RdfFormat::Turtle);
    }

    // Check for TriG (named graphs syntax)
    if content.contains('{') && content.contains('}') {
        return Some(RdfFormat::TriG);
    }

    // Count tokens in first meaningful line to distinguish N-Quads vs N-Triples
    for line in content.lines() {
        let line = line.trim();
        if !line.is_empty() && !line.starts_with('#') {
            let parts: Vec<&str> = line.split_whitespace().collect();
            if parts.len() == 4 && parts[3] == "." {
                // Exactly 4 parts (s p o .) - N-Triples
                return Some(RdfFormat::NTriples);
            } else if parts.len() == 5 && parts[4] == "." {
                // Exactly 5 parts (s p o g .) - N-Quads
                return Some(RdfFormat::NQuads);
            } else if parts.len() >= 3 && parts[parts.len() - 1] == "." {
                // Fallback: assume N-Triples for basic triple pattern
                return Some(RdfFormat::NTriples);
            }
            break; // Only check first meaningful line
        }
    }

    None
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::model::graph::Graph;

    #[test]
    fn test_format_detection_from_extension() {
        assert_eq!(RdfFormat::from_extension("ttl"), Some(RdfFormat::Turtle));
        assert_eq!(RdfFormat::from_extension("turtle"), Some(RdfFormat::Turtle));
        assert_eq!(RdfFormat::from_extension("nt"), Some(RdfFormat::NTriples));
        assert_eq!(
            RdfFormat::from_extension("ntriples"),
            Some(RdfFormat::NTriples)
        );
        assert_eq!(RdfFormat::from_extension("trig"), Some(RdfFormat::TriG));
        assert_eq!(RdfFormat::from_extension("nq"), Some(RdfFormat::NQuads));
        assert_eq!(RdfFormat::from_extension("rdf"), Some(RdfFormat::RdfXml));
        assert_eq!(RdfFormat::from_extension("jsonld"), Some(RdfFormat::JsonLd));
        assert_eq!(RdfFormat::from_extension("unknown"), None);
    }

    #[test]
    fn test_format_properties() {
        assert_eq!(RdfFormat::Turtle.media_type(), "text/turtle");
        assert_eq!(RdfFormat::NTriples.extension(), "nt");
        assert!(RdfFormat::TriG.supports_quads());
        assert!(!RdfFormat::Turtle.supports_quads());
    }

    #[test]
    fn test_format_detection_from_content() {
        // XML content
        let xml_content = "<?xml version=\"1.0\"?>\n<rdf:RDF>";
        assert_eq!(
            detect_format_from_content(xml_content),
            Some(RdfFormat::RdfXml)
        );

        // JSON-LD content
        let jsonld_content = r#"{"@context": "http://example.org", "@type": "Person"}"#;
        assert_eq!(
            detect_format_from_content(jsonld_content),
            Some(RdfFormat::JsonLd)
        );

        // Turtle content
        let turtle_content = "@prefix foaf: <http://xmlns.com/foaf/0.1/> .";
        assert_eq!(
            detect_format_from_content(turtle_content),
            Some(RdfFormat::Turtle)
        );

        // N-Triples content
        let ntriples_content = "<http://example.org/s> <http://example.org/p> \"object\" .";
        assert_eq!(
            detect_format_from_content(ntriples_content),
            Some(RdfFormat::NTriples)
        );
    }

    #[test]
    fn test_ntriples_parsing_simple() {
        let ntriples_data = r#"<http://example.org/alice> <http://xmlns.com/foaf/0.1/name> "Alice Smith" .
<http://example.org/alice> <http://xmlns.com/foaf/0.1/age> "30"^^<http://www.w3.org/2001/XMLSchema#integer> .
_:person1 <http://xmlns.com/foaf/0.1/knows> <http://example.org/bob> ."#;

        let parser = Parser::new(RdfFormat::NTriples);
        let result = parser.parse_str_to_quads(ntriples_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 3);

        // Check that all quads are in the default graph
        for quad in &quads {
            assert!(quad.is_default_graph());
        }

        // Convert to triples for easier checking
        let triples: Vec<_> = quads.into_iter().map(|q| q.to_triple()).collect();

        // Check first triple
        let alice_iri = NamedNode::new("http://example.org/alice").expect("valid IRI");
        let name_pred = NamedNode::new("http://xmlns.com/foaf/0.1/name").expect("valid IRI");
        let name_literal = Literal::new("Alice Smith");
        let expected_triple1 = Triple::new(alice_iri.clone(), name_pred, name_literal);
        assert!(triples.contains(&expected_triple1));

        // Check typed literal triple
        let age_pred = NamedNode::new("http://xmlns.com/foaf/0.1/age").expect("valid IRI");
        let integer_type =
            NamedNode::new("http://www.w3.org/2001/XMLSchema#integer").expect("valid IRI");
        let age_literal = Literal::new_typed("30", integer_type);
        let expected_triple2 = Triple::new(alice_iri, age_pred, age_literal);
        assert!(triples.contains(&expected_triple2));

        // Check blank node triple
        let blank_node = BlankNode::new("_:person1").expect("valid blank node id");
        let knows_pred = NamedNode::new("http://xmlns.com/foaf/0.1/knows").expect("valid IRI");
        let bob_iri = NamedNode::new("http://example.org/bob").expect("valid IRI");
        let expected_triple3 = Triple::new(blank_node, knows_pred, bob_iri);
        assert!(triples.contains(&expected_triple3));
    }

    #[test]
    fn test_ntriples_parsing_language_tag() {
        let ntriples_data =
            r#"<http://example.org/alice> <http://example.org/description> "Une personne"@fr ."#;

        let parser = Parser::new(RdfFormat::NTriples);
        let result = parser.parse_str_to_quads(ntriples_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 1);

        let triple = quads[0].to_triple();
        if let Object::Literal(literal) = triple.object() {
            assert_eq!(literal.value(), "Une personne");
            assert_eq!(literal.language(), Some("fr"));
            assert!(literal.is_lang_string());
        } else {
            panic!("Expected literal object");
        }
    }

    #[test]
    fn test_ntriples_parsing_escaped_literals() {
        let ntriples_data = r#"<http://example.org/test> <http://example.org/desc> "Text with \"quotes\" and \n newlines" ."#;

        let parser = Parser::new(RdfFormat::NTriples);
        let result = parser.parse_str_to_quads(ntriples_data);

        if let Err(e) = &result {
            println!("Parse error: {e}");
        }
        assert!(result.is_ok(), "Parse failed: {result:?}");

        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 1);

        let triple = quads[0].to_triple();
        if let Object::Literal(literal) = triple.object() {
            assert!(literal.value().contains("\"quotes\""));
            assert!(literal.value().contains("\n"));
        } else {
            panic!("Expected literal object");
        }
    }

    #[test]
    fn test_ntriples_parsing_comments_and_empty_lines() {
        let ntriples_data = r#"
# This is a comment
<http://example.org/alice> <http://xmlns.com/foaf/0.1/name> "Alice Smith" .

# Another comment
<http://example.org/bob> <http://xmlns.com/foaf/0.1/name> "Bob Jones" .
"#;

        let parser = Parser::new(RdfFormat::NTriples);
        let result = parser.parse_str_to_quads(ntriples_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 2);
    }

    #[test]
    fn test_ntriples_parsing_error_handling() {
        // Test invalid syntax
        let invalid_data = "invalid ntriples data";
        let parser = Parser::new(RdfFormat::NTriples);
        let result = parser.parse_str_to_quads(invalid_data);
        assert!(result.is_err());

        // Test error tolerance
        let mixed_data = r#"<http://example.org/valid> <http://example.org/pred> "Valid triple" .
invalid line here
<http://example.org/valid2> <http://example.org/pred> "Another valid triple" ."#;

        let parser_strict = Parser::new(RdfFormat::NTriples);
        let result_strict = parser_strict.parse_str_to_quads(mixed_data);
        assert!(result_strict.is_err());

        let parser_tolerant = Parser::new(RdfFormat::NTriples).with_error_tolerance(true);
        let result_tolerant = parser_tolerant.parse_str_to_quads(mixed_data);
        assert!(result_tolerant.is_ok());
        let quads = result_tolerant.expect("tolerant parse should succeed");
        assert_eq!(quads.len(), 2); // Should parse the two valid triples
    }

    #[test]
    fn test_nquads_parsing() {
        let nquads_data = r#"<http://example.org/alice> <http://xmlns.com/foaf/0.1/name> "Alice Smith" <http://example.org/graph1> .
<http://example.org/alice> <http://xmlns.com/foaf/0.1/age> "30"^^<http://www.w3.org/2001/XMLSchema#integer> <http://example.org/graph2> .
_:person1 <http://xmlns.com/foaf/0.1/knows> <http://example.org/bob> _:graph1 ."#;

        let parser = Parser::new(RdfFormat::NQuads);
        let result = parser.parse_str_to_quads(nquads_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 3);

        // Check that quads have proper graph names
        let first_quad = &quads[0];
        assert!(!first_quad.is_default_graph());

        // Check that we can extract graph names
        if let GraphName::NamedNode(graph_name) = first_quad.graph_name() {
            assert!(graph_name.as_str().contains("example.org"));
        } else {
            panic!("Expected named graph");
        }
    }

    #[test]
    fn test_turtle_parsing_basic() {
        let turtle_data = r#"@prefix foaf: <http://xmlns.com/foaf/0.1/> .
@prefix ex: <http://example.org/> .

ex:alice foaf:name "Alice Smith" .
ex:alice foaf:age "30"^^<http://www.w3.org/2001/XMLSchema#integer> .
ex:alice foaf:knows ex:bob ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let result = parser.parse_str_to_quads(turtle_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 3);

        // All quads should be in default graph
        for quad in &quads {
            assert!(quad.is_default_graph());
        }
    }

    #[test]
    fn test_turtle_parsing_prefixes() {
        let turtle_data = r#"@prefix foaf: <http://xmlns.com/foaf/0.1/> .
foaf:Person a foaf:Person ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let result = parser.parse_str_to_quads(turtle_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 1);

        let triple = quads[0].to_triple();
        // Should expand foaf:Person to full IRI
        if let Subject::NamedNode(subj) = triple.subject() {
            assert!(subj.as_str().contains("xmlns.com/foaf"));
        } else {
            panic!("Expected named node subject");
        }

        // Predicate should be rdf:type (from 'a')
        if let Predicate::NamedNode(pred) = triple.predicate() {
            assert!(pred.as_str().contains("rdf-syntax-ns#type"));
        } else {
            panic!("Expected named node predicate");
        }
    }

    #[test]
    fn test_turtle_parsing_abbreviated_syntax() {
        let turtle_data = r#"@prefix ex: <http://example.org/> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .

ex:alice foaf:name "Alice" ;
         foaf:age "30" ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let result = parser.parse_str_to_quads(turtle_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 2);

        // Both triples should have the same subject
        let subjects: Vec<_> = quads
            .iter()
            .map(|q| q.to_triple().subject().clone())
            .collect();
        assert_eq!(subjects[0], subjects[1]);
    }

    /// Regression test for the P0 finding: semicolon-splitting used to be
    /// done on the raw accumulated statement *before* any quote-aware
    /// tokenization, so a literal value containing a `;` corrupted the
    /// triple (spurious extra/garbled triples or parse errors). The real
    /// oxttl-backed grammar must treat `;` inside a string literal as plain
    /// literal content, not a predicate-object-list separator.
    #[test]
    fn test_turtle_semicolon_inside_literal_not_split() {
        let turtle_data = r#"@prefix ex: <http://example.org/> .
ex:alice ex:bio "Loves cats; dogs; and turtles" ;
         ex:name "Alice" ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let quads = parser
            .parse_str_to_quads(turtle_data)
            .expect("semicolon inside a literal must not corrupt parsing");

        assert_eq!(quads.len(), 2, "expected exactly 2 triples, got {quads:?}");

        let bio_triple = quads
            .iter()
            .map(|q| q.to_triple())
            .find(|t| t.predicate().to_string().contains("bio"))
            .expect("bio triple should be present");
        if let Object::Literal(lit) = bio_triple.object() {
            assert_eq!(lit.value(), "Loves cats; dogs; and turtles");
        } else {
            panic!("Expected literal object for ex:bio");
        }
    }

    /// Regression test for the P0 finding: the hand-rolled state machine had
    /// no support for comma-separated object lists (`predicate obj1, obj2`).
    #[test]
    fn test_turtle_comma_object_list() {
        let turtle_data = r#"@prefix ex: <http://example.org/> .
ex:alice ex:knows ex:bob, ex:carol, ex:dave ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let quads = parser
            .parse_str_to_quads(turtle_data)
            .expect("comma object lists must parse");

        assert_eq!(quads.len(), 3, "expected 3 triples, got {quads:?}");
        let objects: std::collections::HashSet<String> = quads
            .iter()
            .map(|q| q.to_triple().object().to_string())
            .collect();
        assert!(objects.iter().any(|o| o.contains("bob")));
        assert!(objects.iter().any(|o| o.contains("carol")));
        assert!(objects.iter().any(|o| o.contains("dave")));
    }

    /// Regression test for the P0 finding: no support for blank-node
    /// property lists (`[ p o ]`) or RDF collections (`( a b c )`).
    #[test]
    fn test_turtle_blank_node_property_list_and_collection() {
        let turtle_data = r#"@prefix ex: <http://example.org/> .
ex:alice ex:address [ ex:city "Springfield" ; ex:zip "12345" ] ;
         ex:favorites ( ex:tea ex:coffee ex:cocoa ) ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let quads = parser
            .parse_str_to_quads(turtle_data)
            .expect("blank-node property lists and collections must parse");

        // ex:address triple + 2 property triples on the blank node = 3
        // ex:favorites triple + 3-element rdf:List (3 rdf:first + 3 rdf:rest, one being rdf:nil) = 4 triples
        // Just assert we got a reasonable number of triples and specific content is present.
        assert!(
            quads.len() >= 7,
            "expected at least 7 triples for property list + collection, got {} ({quads:?})",
            quads.len()
        );

        let has_city = quads.iter().any(|q| {
            let t = q.to_triple();
            t.predicate().to_string().contains("city")
                && matches!(t.object(), Object::Literal(l) if l.value() == "Springfield")
        });
        assert!(has_city, "blank-node property list content missing");

        let has_first = quads.iter().any(|q| {
            q.to_triple()
                .predicate()
                .to_string()
                .contains("rdf-syntax-ns#first")
        });
        assert!(
            has_first,
            "RDF collection must expand to rdf:first/rdf:rest"
        );
    }

    /// Regression test for the P0 finding: multi-line `data.lines()`
    /// processing used to collapse newlines inside triple-quoted string
    /// literals into spaces, corrupting their content.
    #[test]
    fn test_turtle_triple_quoted_multiline_literal() {
        let turtle_data = "@prefix ex: <http://example.org/> .\nex:alice ex:bio \"\"\"Line one\nLine two\nLine three\"\"\" .";

        let parser = Parser::new(RdfFormat::Turtle);
        let quads = parser
            .parse_str_to_quads(turtle_data)
            .expect("triple-quoted multi-line literals must parse");

        assert_eq!(quads.len(), 1);
        let triple = quads[0].to_triple();
        if let Object::Literal(lit) = triple.object() {
            assert_eq!(lit.value(), "Line one\nLine two\nLine three");
        } else {
            panic!("Expected literal object");
        }
    }

    /// Same triple-quoted multi-line literal regression, but through the
    /// TriG entry point (which delegates through the same real grammar).
    #[test]
    fn test_trig_triple_quoted_multiline_literal_in_named_graph() {
        let trig_data = "@prefix ex: <http://example.org/> .\nex:g1 { ex:alice ex:bio \"\"\"Line one\nLine two\"\"\" . }";

        let parser = Parser::new(RdfFormat::TriG);
        let quads = parser
            .parse_str_to_quads(trig_data)
            .expect("triple-quoted multi-line literals must parse in TriG too");

        assert_eq!(quads.len(), 1);
        assert!(!quads[0].is_default_graph());
        let triple = quads[0].to_triple();
        if let Object::Literal(lit) = triple.object() {
            assert_eq!(lit.value(), "Line one\nLine two");
        } else {
            panic!("Expected literal object");
        }
    }

    #[test]
    fn test_turtle_parsing_base_iri() {
        let turtle_data = r#"@base <http://example.org/> .
<alice> <knows> <bob> ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let result = parser.parse_str_to_quads(turtle_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 1);

        let triple = quads[0].to_triple();
        // IRIs should be resolved relative to base
        if let Subject::NamedNode(subj) = triple.subject() {
            assert!(subj.as_str().contains("example.org"));
        } else {
            panic!("Expected named node subject");
        }
    }

    #[test]
    fn test_turtle_parsing_literals() {
        let turtle_data = r#"@prefix ex: <http://example.org/> .
ex:alice ex:name "Alice"@en .
ex:alice ex:age "30"^^<http://www.w3.org/2001/XMLSchema#integer> ."#;

        let parser = Parser::new(RdfFormat::Turtle);
        let result = parser.parse_str_to_quads(turtle_data);

        assert!(result.is_ok());
        let quads = result.expect("should have value");
        assert_eq!(quads.len(), 2);

        // Check for language tag and datatype
        let triples: Vec<_> = quads.into_iter().map(|q| q.to_triple()).collect();

        let mut found_lang_literal = false;
        let mut found_typed_literal = false;

        for triple in triples {
            if let Object::Literal(literal) = triple.object() {
                if literal.language().is_some() {
                    found_lang_literal = true;
                    assert_eq!(literal.language(), Some("en"));
                } else {
                    let datatype = literal.datatype();
                    // Check for typed literal (not language-tagged and not plain string)
                    if datatype.as_str() != "http://www.w3.org/2001/XMLSchema#string"
                        && datatype.as_str()
                            != "http://www.w3.org/1999/02/22-rdf-syntax-ns#langString"
                    {
                        found_typed_literal = true;
                        assert!(
                            datatype.as_str().contains("integer"),
                            "Expected integer datatype but got: {}",
                            datatype.as_str()
                        );
                    }
                }
            }
        }

        assert!(found_lang_literal);
        assert!(found_typed_literal);
    }

    #[test]
    fn test_parser_round_trip() {
        use crate::serializer::Serializer;

        // Create a graph with various types of triples
        let mut original_graph = Graph::new();

        let alice = NamedNode::new("http://example.org/alice").expect("valid IRI");
        let name_pred = NamedNode::new("http://xmlns.com/foaf/0.1/name").expect("valid IRI");
        let name_literal = Literal::new("Alice Smith");
        original_graph.insert(Triple::new(alice.clone(), name_pred, name_literal));

        let age_pred = NamedNode::new("http://xmlns.com/foaf/0.1/age").expect("valid IRI");
        let age_literal = Literal::new_typed("30", crate::vocab::xsd::INTEGER.clone());
        original_graph.insert(Triple::new(alice.clone(), age_pred, age_literal));

        let desc_pred = NamedNode::new("http://example.org/description").expect("valid IRI");
        let desc_literal =
            Literal::new_lang("Une personne", "fr").expect("construction should succeed");
        original_graph.insert(Triple::new(alice, desc_pred, desc_literal));

        // Serialize to N-Triples
        let serializer = Serializer::new(RdfFormat::NTriples);
        let ntriples = serializer
            .serialize_graph(&original_graph)
            .expect("operation should succeed");

        // Parse back from N-Triples
        let parser = Parser::new(RdfFormat::NTriples);
        let quads = parser
            .parse_str_to_quads(&ntriples)
            .expect("operation should succeed");

        // Convert back to graph
        let parsed_graph = Graph::from_iter(quads.into_iter().map(|q| q.to_triple()));

        // Should have the same number of triples
        assert_eq!(original_graph.len(), parsed_graph.len());

        // All original triples should be present in parsed graph
        for triple in original_graph.iter() {
            assert!(
                parsed_graph.contains(triple),
                "Parsed graph missing triple: {triple}"
            );
        }
    }

    #[test]
    fn test_trig_parser() {
        let trig_data = r#"
@prefix ex: <http://example.org/> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

# Default graph
{
    ex:alice rdf:type ex:Person .
    ex:alice ex:name "Alice" .
}

# Named graph
ex:graph1 {
    ex:bob rdf:type ex:Person .
    ex:bob ex:name "Bob" .
    ex:bob ex:age "30" .
}
"#;

        let parser = Parser::new(RdfFormat::TriG);
        let quads = parser
            .parse_str_to_quads(trig_data)
            .expect("operation should succeed");

        // Should parse all statements
        assert!(
            quads.len() >= 5,
            "Should parse at least 5 quads, got {}",
            quads.len()
        );

        // Check that we have both default and named graph quads
        let default_graph_count = quads.iter().filter(|q| q.is_default_graph()).count();
        let named_graph_count = quads.len() - default_graph_count;

        assert!(
            default_graph_count >= 2,
            "Should have at least 2 default graph quads, got {default_graph_count}"
        );
        assert!(
            named_graph_count >= 3,
            "Should have at least 3 named graph quads, got {named_graph_count}"
        );

        // Verify specific content
        let alice_uri = "http://example.org/alice";
        let bob_uri = "http://example.org/bob";
        let person_uri = "http://example.org/Person";

        // Check for Alice in default graph
        let alice_type_found = quads.iter().any(|q| {
            q.is_default_graph()
                && q.subject().to_string().contains(alice_uri)
                && q.object().to_string().contains(person_uri)
        });
        assert!(
            alice_type_found,
            "Should find Alice type assertion in default graph"
        );

        // Check for Bob in named graph
        let bob_in_named_graph = quads
            .iter()
            .any(|q| !q.is_default_graph() && q.subject().to_string().contains(bob_uri));
        assert!(
            bob_in_named_graph,
            "Should find Bob statements in named graph"
        );
    }

    #[test]
    fn test_trig_parser_prefixes() {
        let trig_data = r#"
@prefix ex: <http://example.org/> .
@prefix foaf: <http://xmlns.com/foaf/0.1/> .

ex:person1 foaf:name "John Doe" .
"#;

        let parser = Parser::new(RdfFormat::TriG);
        let quads = parser
            .parse_str_to_quads(trig_data)
            .expect("operation should succeed");

        assert!(!quads.is_empty(), "Should parse prefixed statements");

        // Verify prefix expansion worked
        let expanded_found = quads.iter().any(|q| {
            q.subject()
                .to_string()
                .contains("http://example.org/person1")
                && q.predicate()
                    .to_string()
                    .contains("http://xmlns.com/foaf/0.1/name")
        });
        assert!(expanded_found, "Should expand prefixes correctly");
    }

    #[test]
    fn test_jsonld_parser() {
        let jsonld_data = r#"{
    "@context": {
        "name": "http://xmlns.com/foaf/0.1/name",
        "Person": "http://schema.org/Person"
    },
    "@type": "Person",
    "@id": "http://example.org/john",
    "name": "John Doe"
}"#;

        let parser = Parser::new(RdfFormat::JsonLd);
        let result = parser.parse_str_to_quads(jsonld_data);

        match result {
            Ok(quads) => {
                println!("JSON-LD parsed {} quads:", quads.len());
                for quad in &quads {
                    println!("  {quad}");
                }
                assert!(!quads.is_empty(), "Should parse some quads from JSON-LD");
            }
            Err(e) => {
                // For now, just verify that the parser attempts to parse
                println!("JSON-LD parsing error (expected during development): {e}");
                // Don't fail the test yet as the implementation might need more work
            }
        }
    }

    #[test]
    fn test_jsonld_parser_simple() {
        let jsonld_data = r#"{
    "@context": "http://schema.org/",
    "@type": "Person",
    "name": "Alice"
}"#;

        let parser = Parser::new(RdfFormat::JsonLd);
        let result = parser.parse_str_to_quads(jsonld_data);

        // For now, just verify the parser doesn't crash
        match result {
            Ok(quads) => {
                println!("Simple JSON-LD parsed {} quads", quads.len());
            }
            Err(e) => {
                println!("Simple JSON-LD parsing error: {e}");
                // Don't fail during development
            }
        }
    }
}