1import re
 2from typing import List, Dict
 3
 4class MarkdownLinkChecker:
 5    """Markdown 連結檢查器"""
 6
 7    # Markdown 連結正則表達式
 8    # 匹配 [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/target) 格式，排除圖片 ![alt](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/src)
 9    INLINE_LINK_PATTERN = re.compile(
10        r'(?<!!)\[([^\]]+)\]\(([^)]+)\)'
11    )
12
13    # 引用式連結定義 [ref]: target
14    REFERENCE_DEF_PATTERN = re.compile(
15        r'^\s*\[([^\]]+)\]:\s*(.+)$',
16        re.MULTILINE
17    )
18
19    # 引用式連結使用 [text][ref]
20    REFERENCE_USE_PATTERN = re.compile(
21        r'\[([^\]]+)\]\[([^\]]+)\]'
22    )
23
24    def parse_markdown_links(self, content: str) -> List[Dict]:
25        """
26        解析 Markdown 內容中的所有連結
27
28        Args:
29            content: Markdown 內容
30
31        Returns:
32            list[dict]: 連結列表，每個包含 text, target, line
33        """
34        links = []
35        lines = content.split('\n')
36
37        # 首先收集引用式連結定義
38        reference_defs = {}
39        for match in self.REFERENCE_DEF_PATTERN.finditer(content):
40            ref_name = match.group(1).lower()
41            ref_target = match.group(2).strip()
42            reference_defs[ref_name] = ref_target
43
44        # 追蹤是否在程式碼區塊內
45        in_code_block = False
46
47        # 解析行內連結
48        for line_num, line in enumerate(lines, start=1):
49            # 檢查程式碼區塊開始/結束
50            if line.strip().startswith("```"):
51                in_code_block = not in_code_block
52                continue
53
54            # 跳過程式碼區塊內的連結
55            if in_code_block:
56                continue
57
58            # 行內連結 [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/target)
59            for match in self.INLINE_LINK_PATTERN.finditer(line):
60                links.append({
61                    "text": match.group(1),
62                    "target": match.group(2),
63                    "line": line_num
64                })
65
66            # 引用式連結 [text][ref]
67            for match in self.REFERENCE_USE_PATTERN.finditer(line):
68                ref_name = match.group(2).lower()
69                if ref_name in reference_defs:
70                    links.append({
71                        "text": match.group(1),
72                        "target": reference_defs[ref_name],
73                        "line": line_num
74                    })
75
76        return links

 1# 安裝 maturin（如果尚未安裝）
 2pip install maturin
 3
 4# 建立新專案
 5mkdir markdown_parser_rs
 6cd markdown_parser_rs
 7
 8# 初始化 maturin 專案（選擇 pyo3 作為綁定）
 9maturin init --bindings pyo3
10
11# 專案結構如下：
12# markdown_parser_rs/
13# ├── Cargo.toml
14# ├── pyproject.toml
15# └── src/
16#     └── lib.rs

 1[package]
 2name = "markdown_parser_rs"
 3version = "0.1.0"
 4edition = "2021"
 5
 6[lib]
 7name = "markdown_parser_rs"
 8crate-type = ["cdylib"]
 9
10[dependencies]
11pyo3 = { version = "0.22", features = ["extension-module"] }
12regex = "1.11"
13once_cell = "1.20"

 1// src/lib.rs
 2use once_cell::sync::Lazy;
 3use regex::Regex;
 4use std::collections::HashMap;
 5
 6/// Represents a parsed markdown link
 7#[derive(Debug, Clone)]
 8pub struct MarkdownLink {
 9    pub text: String,
10    pub target: String,
11    pub line: usize,
12}
13
14// Pre-compiled regex patterns (compile once, use many times)
15static INLINE_LINK_PATTERN: Lazy<Regex> = Lazy::new(|| {
16    // Match [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/target), excluding images ![alt](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/src)
17    Regex::new(r"(?<!!)\[([^\]]+)\]\(([^)]+)\)").unwrap()
18});
19
20static REFERENCE_DEF_PATTERN: Lazy<Regex> = Lazy::new(|| {
21    // Match [ref]: target
22    Regex::new(r"(?m)^\s*\[([^\]]+)\]:\s*(.+)$").unwrap()
23});
24
25static REFERENCE_USE_PATTERN: Lazy<Regex> = Lazy::new(|| {
26    // Match [text][ref]
27    Regex::new(r"\[([^\]]+)\]\[([^\]]+)\]").unwrap()
28});
29
30/// Parse markdown content and extract all links
31pub fn parse_links(content: &str) -> Vec<MarkdownLink> {
32    let mut links = Vec::new();
33
34    // First, collect reference definitions
35    let mut reference_defs: HashMap<String, String> = HashMap::new();
36    for cap in REFERENCE_DEF_PATTERN.captures_iter(content) {
37        let ref_name = cap[1].to_lowercase();
38        let ref_target = cap[2].trim().to_string();
39        reference_defs.insert(ref_name, ref_target);
40    }
41
42    // Track code block state
43    let mut in_code_block = false;
44
45    // Parse line by line
46    for (line_num, line) in content.lines().enumerate() {
47        let line_number = line_num + 1; // 1-indexed
48
49        // Check for code block markers
50        if line.trim_start().starts_with("```") {
51            in_code_block = !in_code_block;
52            continue;
53        }
54
55        // Skip content inside code blocks
56        if in_code_block {
57            continue;
58        }
59
60        // Parse inline links [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/target)
61        for cap in INLINE_LINK_PATTERN.captures_iter(line) {
62            links.push(MarkdownLink {
63                text: cap[1].to_string(),
64                target: cap[2].to_string(),
65                line: line_number,
66            });
67        }
68
69        // Parse reference links [text][ref]
70        for cap in REFERENCE_USE_PATTERN.captures_iter(line) {
71            let ref_name = cap[2].to_lowercase();
72            if let Some(target) = reference_defs.get(&ref_name) {
73                links.push(MarkdownLink {
74                    text: cap[1].to_string(),
75                    target: target.clone(),
76                    line: line_number,
77                });
78            }
79        }
80    }
81
82    links
83}

  1use pyo3::prelude::*;
  2use pyo3::types::PyDict;
  3
  4/// Python-visible link structure
  5#[pyclass]
  6#[derive(Clone)]
  7pub struct PyMarkdownLink {
  8    #[pyo3(get)]
  9    pub text: String,
 10    #[pyo3(get)]
 11    pub target: String,
 12    #[pyo3(get)]
 13    pub line: usize,
 14}
 15
 16#[pymethods]
 17impl PyMarkdownLink {
 18    fn __repr__(&self) -> String {
 19        format!(
 20            "MarkdownLink(text='{}', target='{}', line={})",
 21            self.text, self.target, self.line
 22        )
 23    }
 24
 25    /// Convert to Python dict for compatibility
 26    fn to_dict<'py>(&self, py: Python<'py>) -> Bound<'py, PyDict> {
 27        let dict = PyDict::new(py);
 28        dict.set_item("text", &self.text).unwrap();
 29        dict.set_item("target", &self.target).unwrap();
 30        dict.set_item("line", self.line).unwrap();
 31        dict
 32    }
 33}
 34
 35impl From<MarkdownLink> for PyMarkdownLink {
 36    fn from(link: MarkdownLink) -> Self {
 37        PyMarkdownLink {
 38            text: link.text,
 39            target: link.target,
 40            line: link.line,
 41        }
 42    }
 43}
 44
 45/// Parse markdown content and return list of links as objects
 46#[pyfunction]
 47fn parse_markdown_links(content: &str) -> Vec<PyMarkdownLink> {
 48    parse_links(content)
 49        .into_iter()
 50        .map(PyMarkdownLink::from)
 51        .collect()
 52}
 53
 54/// Parse markdown content and return list of links as dicts
 55/// (for drop-in compatibility with existing Python code)
 56#[pyfunction]
 57fn parse_markdown_links_as_dicts<'py>(
 58    py: Python<'py>,
 59    content: &str,
 60) -> Vec<Bound<'py, PyDict>> {
 61    parse_links(content)
 62        .into_iter()
 63        .map(|link| {
 64            let dict = PyDict::new(py);
 65            dict.set_item("text", link.text).unwrap();
 66            dict.set_item("target", link.target).unwrap();
 67            dict.set_item("line", link.line).unwrap();
 68            dict
 69        })
 70        .collect()
 71}
 72
 73/// Filter out external links, keeping only internal links
 74#[pyfunction]
 75fn filter_internal_links(links: Vec<PyMarkdownLink>) -> Vec<PyMarkdownLink> {
 76    links
 77        .into_iter()
 78        .filter(|link| {
 79            let target = &link.target;
 80            // Skip pure anchor links
 81            if target.starts_with('#') {
 82                return false;
 83            }
 84            // Skip external links
 85            if target.starts_with("http://")
 86                || target.starts_with("https://")
 87                || target.starts_with("mailto:")
 88                || target.starts_with("tel:")
 89                || target.starts_with("ftp://")
 90            {
 91                return false;
 92            }
 93            true
 94        })
 95        .collect()
 96}
 97
 98/// Python module definition
 99#[pymodule]
100fn markdown_parser_rs(m: &Bound<'_, PyModule>) -> PyResult<()> {
101    m.add_class::<PyMarkdownLink>()?;
102    m.add_function(wrap_pyfunction!(parse_markdown_links, m)?)?;
103    m.add_function(wrap_pyfunction!(parse_markdown_links_as_dicts, m)?)?;
104    m.add_function(wrap_pyfunction!(filter_internal_links, m)?)?;
105    Ok(())
106}

 1# 開發模式建置（快速，用於測試）
 2maturin develop
 3
 4# 或者以 release 模式建置（優化效能）
 5maturin develop --release
 6
 7# 建置 wheel 套件
 8maturin build --release
 9
10# 安裝到當前環境
11pip install target/wheels/markdown_parser_rs-*.whl

  1//! Markdown Link Parser - A high-performance parser written in Rust
  2//!
  3//! This module provides fast markdown link parsing capabilities
  4//! using Rust's regex crate and PyO3 for Python bindings.
  5
  6use once_cell::sync::Lazy;
  7use pyo3::prelude::*;
  8use pyo3::types::PyDict;
  9use regex::Regex;
 10use std::collections::HashMap;
 11
 12// ============================================================================
 13// Core Data Structures
 14// ============================================================================
 15
 16/// Internal link representation
 17#[derive(Debug, Clone)]
 18struct MarkdownLink {
 19    text: String,
 20    target: String,
 21    line: usize,
 22}
 23
 24/// Python-visible link structure with getter methods
 25#[pyclass]
 26#[derive(Clone)]
 27pub struct PyMarkdownLink {
 28    #[pyo3(get)]
 29    pub text: String,
 30    #[pyo3(get)]
 31    pub target: String,
 32    #[pyo3(get)]
 33    pub line: usize,
 34}
 35
 36#[pymethods]
 37impl PyMarkdownLink {
 38    /// String representation for debugging
 39    fn __repr__(&self) -> String {
 40        format!(
 41            "MarkdownLink(text='{}', target='{}', line={})",
 42            self.text, self.target, self.line
 43        )
 44    }
 45
 46    /// Convert to Python dict for compatibility with existing code
 47    fn to_dict<'py>(&self, py: Python<'py>) -> Bound<'py, PyDict> {
 48        let dict = PyDict::new(py);
 49        dict.set_item("text", &self.text).unwrap();
 50        dict.set_item("target", &self.target).unwrap();
 51        dict.set_item("line", self.line).unwrap();
 52        dict
 53    }
 54}
 55
 56impl From<MarkdownLink> for PyMarkdownLink {
 57    fn from(link: MarkdownLink) -> Self {
 58        PyMarkdownLink {
 59            text: link.text,
 60            target: link.target,
 61            line: link.line,
 62        }
 63    }
 64}
 65
 66// ============================================================================
 67// Pre-compiled Regex Patterns
 68// ============================================================================
 69
 70// Inline link: [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/target), excluding images ![alt](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/src)
 71static INLINE_LINK_PATTERN: Lazy<Regex> = Lazy::new(|| {
 72    Regex::new(r"(?<!!)\[([^\]]+)\]\(([^)]+)\)").unwrap()
 73});
 74
 75// Reference definition: [ref]: target
 76static REFERENCE_DEF_PATTERN: Lazy<Regex> = Lazy::new(|| {
 77    Regex::new(r"(?m)^\s*\[([^\]]+)\]:\s*(.+)$").unwrap()
 78});
 79
 80// Reference usage: [text][ref]
 81static REFERENCE_USE_PATTERN: Lazy<Regex> = Lazy::new(|| {
 82    Regex::new(r"\[([^\]]+)\]\[([^\]]+)\]").unwrap()
 83});
 84
 85// ============================================================================
 86// Core Parsing Logic
 87// ============================================================================
 88
 89/// Parse markdown content and extract all links
 90///
 91/// This function handles:
 92/// - Inline links: [text](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/url)
 93/// - Reference links: [text][ref] with [ref]: url definitions
 94/// - Code block detection (skips links inside ```)
 95fn parse_links(content: &str) -> Vec<MarkdownLink> {
 96    let mut links = Vec::new();
 97
 98    // Phase 1: Collect all reference definitions
 99    let mut reference_defs: HashMap<String, String> = HashMap::new();
100    for cap in REFERENCE_DEF_PATTERN.captures_iter(content) {
101        let ref_name = cap[1].to_lowercase();
102        let ref_target = cap[2].trim().to_string();
103        reference_defs.insert(ref_name, ref_target);
104    }
105
106    // Phase 2: Parse links line by line
107    let mut in_code_block = false;
108
109    for (line_num, line) in content.lines().enumerate() {
110        let line_number = line_num + 1; // Convert to 1-indexed
111
112        // Toggle code block state
113        if line.trim_start().starts_with("```") {
114            in_code_block = !in_code_block;
115            continue;
116        }
117
118        // Skip content inside code blocks
119        if in_code_block {
120            continue;
121        }
122
123        // Extract inline links
124        for cap in INLINE_LINK_PATTERN.captures_iter(line) {
125            links.push(MarkdownLink {
126                text: cap[1].to_string(),
127                target: cap[2].to_string(),
128                line: line_number,
129            });
130        }
131
132        // Extract reference-style links
133        for cap in REFERENCE_USE_PATTERN.captures_iter(line) {
134            let ref_name = cap[2].to_lowercase();
135            if let Some(target) = reference_defs.get(&ref_name) {
136                links.push(MarkdownLink {
137                    text: cap[1].to_string(),
138                    target: target.clone(),
139                    line: line_number,
140                });
141            }
142        }
143    }
144
145    links
146}
147
148/// Check if a link target is external
149fn is_external_link(target: &str) -> bool {
150    target.starts_with("http://")
151        || target.starts_with("https://")
152        || target.starts_with("mailto:")
153        || target.starts_with("tel:")
154        || target.starts_with("ftp://")
155}
156
157// ============================================================================
158// Python Interface Functions
159// ============================================================================
160
161/// Parse markdown content and return a list of MarkdownLink objects
162///
163/// Args:
164///     content: The markdown content to parse
165///
166/// Returns:
167///     List of MarkdownLink objects
168///
169/// Example:
170///     >>> links = parse_markdown_links("Check [docs](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/README.md)")
171///     >>> links[0].text
172///     'docs'
173///     >>> links[0].target
174///     './README.md'
175#[pyfunction]
176fn parse_markdown_links(content: &str) -> Vec<PyMarkdownLink> {
177    parse_links(content)
178        .into_iter()
179        .map(PyMarkdownLink::from)
180        .collect()
181}
182
183/// Parse markdown content and return a list of dicts
184///
185/// This function provides drop-in compatibility with the original
186/// Python implementation that returns dicts.
187///
188/// Args:
189///     content: The markdown content to parse
190///
191/// Returns:
192///     List of dicts with keys: text, target, line
193#[pyfunction]
194fn parse_markdown_links_as_dicts<'py>(
195    py: Python<'py>,
196    content: &str,
197) -> Vec<Bound<'py, PyDict>> {
198    parse_links(content)
199        .into_iter()
200        .map(|link| {
201            let dict = PyDict::new(py);
202            dict.set_item("text", link.text).unwrap();
203            dict.set_item("target", link.target).unwrap();
204            dict.set_item("line", link.line).unwrap();
205            dict
206        })
207        .collect()
208}
209
210/// Filter links to keep only internal ones
211///
212/// Removes:
213/// - External links (http://, https://, mailto:, etc.)
214/// - Pure anchor links (#section)
215///
216/// Args:
217///     links: List of MarkdownLink objects
218///
219/// Returns:
220///     Filtered list of internal links
221#[pyfunction]
222fn filter_internal_links(links: Vec<PyMarkdownLink>) -> Vec<PyMarkdownLink> {
223    links
224        .into_iter()
225        .filter(|link| {
226            let target = &link.target;
227            // Skip pure anchor links
228            if target.starts_with('#') {
229                return false;
230            }
231            // Skip external links
232            !is_external_link(target)
233        })
234        .collect()
235}
236
237/// Count total links in content (fast path, no object creation)
238///
239/// Args:
240///     content: The markdown content to parse
241///
242/// Returns:
243///     Number of links found
244#[pyfunction]
245fn count_links(content: &str) -> usize {
246    parse_links(content).len()
247}
248
249/// Parse and filter in one pass (most efficient for link checking)
250///
251/// Args:
252///     content: The markdown content to parse
253///
254/// Returns:
255///     List of internal MarkdownLink objects
256#[pyfunction]
257fn parse_internal_links(content: &str) -> Vec<PyMarkdownLink> {
258    parse_links(content)
259        .into_iter()
260        .filter(|link| {
261            !link.target.starts_with('#') && !is_external_link(&link.target)
262        })
263        .map(PyMarkdownLink::from)
264        .collect()
265}
266
267// ============================================================================
268// Module Definition
269// ============================================================================
270
271/// High-performance Markdown link parser
272///
273/// This module provides Rust-powered functions for parsing
274/// and filtering markdown links.
275///
276/// Functions:
277///     parse_markdown_links: Parse content, return MarkdownLink objects
278///     parse_markdown_links_as_dicts: Parse content, return dicts
279///     parse_internal_links: Parse and filter to internal links only
280///     filter_internal_links: Filter existing links
281///     count_links: Fast link counting
282///
283/// Classes:
284///     MarkdownLink: Represents a parsed link
285#[pymodule]
286fn markdown_parser_rs(m: &Bound<'_, PyModule>) -> PyResult<()> {
287    m.add_class::<PyMarkdownLink>()?;
288    m.add_function(wrap_pyfunction!(parse_markdown_links, m)?)?;
289    m.add_function(wrap_pyfunction!(parse_markdown_links_as_dicts, m)?)?;
290    m.add_function(wrap_pyfunction!(filter_internal_links, m)?)?;
291    m.add_function(wrap_pyfunction!(count_links, m)?)?;
292    m.add_function(wrap_pyfunction!(parse_internal_links, m)?)?;
293    Ok(())
294}

  1"""
  2使用 Rust 加速的 Markdown 連結檢查器
  3
  4這個範例展示如何用 Rust 模組替換原有的 Python 解析邏輯，
  5同時保持 API 相容性。
  6"""
  7
  8from pathlib import Path
  9from typing import List, Dict, Optional
 10
 11# Try to import Rust module, fallback to pure Python
 12try:
 13    import markdown_parser_rs as parser_rs
 14    USE_RUST = True
 15    print("Using Rust-powered parser")
 16except ImportError:
 17    USE_RUST = False
 18    print("Rust module not available, using pure Python")
 19
 20class MarkdownLinkChecker:
 21    """Markdown link checker with optional Rust acceleration"""
 22
 23    def __init__(self, use_rust: bool = True):
 24        """
 25        Initialize the checker
 26
 27        Args:
 28            use_rust: Whether to use Rust module if available
 29        """
 30        self.use_rust = use_rust and USE_RUST
 31
 32    def parse_markdown_links(self, content: str) -> List[Dict]:
 33        """
 34        Parse markdown content and extract all links
 35
 36        Args:
 37            content: Markdown content
 38
 39        Returns:
 40            List of dicts with keys: text, target, line
 41        """
 42        if self.use_rust:
 43            # Use Rust implementation
 44            return parser_rs.parse_markdown_links_as_dicts(content)
 45        else:
 46            # Fallback to pure Python (original implementation)
 47            return self._parse_python(content)
 48
 49    def parse_internal_links(self, content: str) -> List[Dict]:
 50        """
 51        Parse and filter to internal links only
 52
 53        Args:
 54            content: Markdown content
 55
 56        Returns:
 57            List of internal link dicts
 58        """
 59        if self.use_rust:
 60            # Use optimized Rust function that parses and filters in one pass
 61            links = parser_rs.parse_internal_links(content)
 62            return [link.to_dict() for link in links]
 63        else:
 64            all_links = self._parse_python(content)
 65            return self._filter_internal(all_links)
 66
 67    def _parse_python(self, content: str) -> List[Dict]:
 68        """Pure Python implementation (fallback)"""
 69        import re
 70
 71        INLINE_LINK = re.compile(r'(?<!!)\[([^\]]+)\]\(([^)]+)\)')
 72        REFERENCE_DEF = re.compile(r'(?m)^\s*\[([^\]]+)\]:\s*(.+)$')
 73        REFERENCE_USE = re.compile(r'\[([^\]]+)\]\[([^\]]+)\]')
 74
 75        links = []
 76
 77        # Collect reference definitions
 78        reference_defs = {}
 79        for match in REFERENCE_DEF.finditer(content):
 80            ref_name = match.group(1).lower()
 81            ref_target = match.group(2).strip()
 82            reference_defs[ref_name] = ref_target
 83
 84        # Parse line by line
 85        in_code_block = False
 86        for line_num, line in enumerate(content.split('\n'), start=1):
 87            if line.strip().startswith("```"):
 88                in_code_block = not in_code_block
 89                continue
 90
 91            if in_code_block:
 92                continue
 93
 94            for match in INLINE_LINK.finditer(line):
 95                links.append({
 96                    "text": match.group(1),
 97                    "target": match.group(2),
 98                    "line": line_num
 99                })
100
101            for match in REFERENCE_USE.finditer(line):
102                ref_name = match.group(2).lower()
103                if ref_name in reference_defs:
104                    links.append({
105                        "text": match.group(1),
106                        "target": reference_defs[ref_name],
107                        "line": line_num
108                    })
109
110        return links
111
112    def _filter_internal(self, links: List[Dict]) -> List[Dict]:
113        """Filter to internal links only"""
114        external_prefixes = (
115            'http://', 'https://', 'mailto:', 'tel:', 'ftp://'
116        )
117        return [
118            link for link in links
119            if not link['target'].startswith('#')
120            and not link['target'].startswith(external_prefixes)
121        ]
122
123# Convenience functions
124def check_file(file_path: str, use_rust: bool = True) -> Dict:
125    """
126    Check a single markdown file for broken links
127
128    Args:
129        file_path: Path to the markdown file
130        use_rust: Whether to use Rust acceleration
131
132    Returns:
133        Dict with file_path, total_links, and internal_links count
134    """
135    checker = MarkdownLinkChecker(use_rust=use_rust)
136    path = Path(file_path)
137    content = path.read_text(encoding='utf-8')
138
139    all_links = checker.parse_markdown_links(content)
140    internal_links = checker.parse_internal_links(content)
141
142    return {
143        "file_path": str(path),
144        "total_links": len(all_links),
145        "internal_links": len(internal_links),
146        "links": internal_links
147    }
148
149if __name__ == "__main__":
150    # Example usage
151    sample = (
152        "# Sample Document\n\n"
153        "Check the [documentation](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/docs/README.md) for more info.\n\n"
154        "External link: [Google](https://google.com)\n\n"
155        "Reference style: [API docs][api]\n\n"
156        "[api]: ./api/reference.md\n\n"
157        "~~~python\n"
158        "# This [link](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/should_be_ignored.md) is in a code block\n"
159        "~~~\n"
160    )
161
162    checker = MarkdownLinkChecker()
163    links = checker.parse_markdown_links(sample)
164
165    print("All links found:")
166    for link in links:
167        print(f"  Line {link['line']}: [{link['text']}](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/{link['target']})")
168
169    print("\nInternal links only:")
170    internal = checker.parse_internal_links(sample)
171    for link in internal:
172        print(f"  Line {link['line']}: [{link['text']}](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/{link['target']})")

  1"""
  2Performance comparison: Python vs Cython vs Rust
  3
  4This script benchmarks the three implementations on
  5various markdown file sizes.
  6"""
  7
  8import time
  9import statistics
 10from pathlib import Path
 11from typing import Callable, List, Tuple
 12
 13# Generate test data
 14def generate_markdown(num_links: int) -> str:
 15    """Generate markdown content with specified number of links"""
 16    lines = ["# Test Document\n"]
 17
 18    for i in range(num_links):
 19        if i % 5 == 0:
 20            # Inline link
 21            lines.append(f"Check [link{i}](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/path/to/file{i}.md) for info.\n")
 22        elif i % 5 == 1:
 23            # External link (should be filtered)
 24            lines.append(f"Visit [site{i}](https://example{i}.com)\n")
 25        elif i % 5 == 2:
 26            # Reference style link
 27            lines.append(f"See [doc{i}][ref{i}]\n")
 28            lines.append(f"[ref{i}]: ./docs/page{i}.md\n")
 29        elif i % 5 == 3:
 30            # Anchor link (should be filtered)
 31            lines.append(f"Jump to [section{i}](#section-{i})\n")
 32        else:
 33            # Regular text
 34            lines.append(f"This is paragraph {i} with some text.\n")
 35
 36        # Add occasional code blocks (using ~~~ to avoid markdown parsing issues)
 37        if i % 20 == 0:
 38            lines.append("~~~python\n")
 39            lines.append(f"# [fake link](/python-advanced/06-rust-extensions/case-studies/pyo3-parser/should_ignore_{i}.md)\n")
 40            lines.append("print('hello')\n")
 41            lines.append("~~~\n")
 42
 43    return "".join(lines)
 44
 45def benchmark(
 46    func: Callable[[str], List],
 47    content: str,
 48    iterations: int = 100
 49) -> Tuple[float, float, float]:
 50    """
 51    Benchmark a function
 52
 53    Returns:
 54        Tuple of (mean_time_ms, min_time_ms, max_time_ms)
 55    """
 56    times = []
 57
 58    # Warmup
 59    for _ in range(5):
 60        func(content)
 61
 62    # Actual benchmark
 63    for _ in range(iterations):
 64        start = time.perf_counter()
 65        func(content)
 66        end = time.perf_counter()
 67        times.append((end - start) * 1000)  # Convert to ms
 68
 69    return (
 70        statistics.mean(times),
 71        min(times),
 72        max(times)
 73    )
 74
 75def run_benchmarks():
 76    """Run benchmarks comparing all implementations"""
 77
 78    # Import implementations
 79    import re
 80
 81    # Pure Python implementation
 82    INLINE_LINK = re.compile(r'(?<!!)\[([^\]]+)\]\(([^)]+)\)')
 83
 84    def parse_python(content: str) -> List:
 85        links = []
 86        in_code_block = False
 87        for line_num, line in enumerate(content.split('\n'), 1):
 88            if line.strip().startswith("```"):
 89                in_code_block = not in_code_block
 90                continue
 91            if in_code_block:
 92                continue
 93            for m in INLINE_LINK.finditer(line):
 94                links.append({"text": m.group(1), "target": m.group(2), "line": line_num})
 95        return links
 96
 97    # Try to import Rust implementation
 98    try:
 99        import markdown_parser_rs as rust_parser
100        has_rust = True
101    except ImportError:
102        has_rust = False
103        print("Rust module not available")
104
105    # Test sizes
106    sizes = [100, 500, 1000, 5000, 10000]
107
108    print("=" * 70)
109    print("Markdown Link Parser Benchmark")
110    print("=" * 70)
111    print()
112
113    results = []
114
115    for size in sizes:
116        content = generate_markdown(size)
117        content_kb = len(content.encode('utf-8')) / 1024
118
119        print(f"Test: {size} links (~{content_kb:.1f} KB)")
120        print("-" * 50)
121
122        # Python benchmark
123        py_mean, py_min, py_max = benchmark(parse_python, content)
124        print(f"  Python:  {py_mean:8.3f} ms (min: {py_min:.3f}, max: {py_max:.3f})")
125
126        # Rust benchmark
127        if has_rust:
128            rs_mean, rs_min, rs_max = benchmark(
129                rust_parser.parse_markdown_links,
130                content
131            )
132            speedup = py_mean / rs_mean
133            print(f"  Rust:    {rs_mean:8.3f} ms (min: {rs_min:.3f}, max: {rs_max:.3f})")
134            print(f"  Speedup: {speedup:.1f}x faster")
135
136        print()
137        results.append({
138            "size": size,
139            "python_ms": py_mean,
140            "rust_ms": rs_mean if has_rust else None,
141            "speedup": speedup if has_rust else None
142        })
143
144    # Summary table
145    print("=" * 70)
146    print("Summary")
147    print("=" * 70)
148    print(f"{'Links':<10} {'Python (ms)':<15} {'Rust (ms)':<15} {'Speedup':<10}")
149    print("-" * 50)
150    for r in results:
151        rust_str = f"{r['rust_ms']:.3f}" if r['rust_ms'] else "N/A"
152        speedup_str = f"{r['speedup']:.1f}x" if r['speedup'] else "N/A"
153        print(f"{r['size']:<10} {r['python_ms']:<15.3f} {rust_str:<15} {speedup_str:<10}")
154
155if __name__ == "__main__":
156    run_benchmarks()

1需要加速 Python 程式碼？
2├── 否 → 保持純 Python
3└── 是 → 效能需求多高？
4    ├── 2-5x 足夠 → 考慮 Cython
5    └── 需要 10x+ → 團隊有 Rust 經驗？
6        ├── 是 → 使用 PyO3
7        └── 否 → 效能瓶頸明確嗎？
8            ├── 是 → 值得學習 Rust
9            └── 否 → 先用 Cython，後續再評估

1// 目標：將 "Hello World! 你好" 轉換為 "hello-world-你好"
2#[pyfunction]
3fn slugify(title: &str) -> String {
4    // 你的實作
5    todo!()
6}

 1#[pyfunction]
 2fn slugify(title: &str) -> String {
 3    title
 4        .chars()
 5        .map(|c| {
 6            if c.is_alphanumeric() {
 7                c.to_lowercase().next().unwrap_or(c)
 8            } else if c.is_whitespace() {
 9                '-'
10            } else {
11                // Keep non-ASCII chars (like CJK)
12                if c.is_ascii() { '\0' } else { c }
13            }
14        })
15        .filter(|&c| c != '\0')
16        .collect::<String>()
17        // Clean up multiple consecutive dashes
18        .split('-')
19        .filter(|s| !s.is_empty())
20        .collect::<Vec<_>>()
21        .join("-")
22}

 1// 目標：提取 # 標題，## 標題，### 標題 等
 2#[pyclass]
 3struct Heading {
 4    #[pyo3(get)]
 5    level: usize,
 6    #[pyo3(get)]
 7    text: String,
 8    #[pyo3(get)]
 9    line: usize,
10}
11
12#[pyfunction]
13fn extract_headings(content: &str) -> Vec<Heading> {
14    // 你的實作
15    todo!()
16}