scrape-le 0.1.4

Check whether a page is scrapeable before the scraper is written, and say when it cannot tell
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427

//! The exit codes are the API — a caller programs against "0 means
//! yes, 1 means a real no, 2 means the question was malformed" — so
//! they are pinned by tests that drive the built binary against a
//! local fixture server. No display, no browser, no internet, so they
//! run anywhere on every push.

use std::fmt::Write as _;
use std::io::{BufRead, BufReader, Read, Write};
use std::net::{TcpListener, TcpStream};
use std::process::{Command, Stdio};
use std::sync::Arc;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::thread;

/// A response the fixture server should give for a path.
struct Route {
    path: &'static str,
    status: &'static str,
    headers: &'static [(&'static str, &'static str)],
    body: &'static str,
}

const ROUTES: &[Route] = &[
    Route {
        path: "/robots.txt",
        status: "200 OK",
        headers: &[("content-type", "text/plain")],
        body: "User-agent: *\nDisallow: /forbidden\nCrawl-delay: 0\n\nUser-agent: PickyBot\nDisallow: /\n",
    },
    Route {
        path: "/open",
        status: "200 OK",
        headers: &[("content-type", "text/html")],
        body: "<html><head><title>Open</title></head><body>hello</body></html>",
    },
    Route {
        path: "/forbidden",
        status: "200 OK",
        headers: &[("content-type", "text/html")],
        body: "<html><head><title>Forbidden by robots</title></head><body>hi</body></html>",
    },
    Route {
        path: "/limited",
        status: "429 Too Many Requests",
        headers: &[("retry-after", "60"), ("x-ratelimit-limit", "100")],
        body: "slow down",
    },
    Route {
        path: "/secret",
        status: "401 Unauthorized",
        headers: &[("content-type", "text/html")],
        body: "<html><head><title>Members</title></head><body>no</body></html>",
    },
    Route {
        path: "/cloudflare",
        status: "200 OK",
        headers: &[("cf-ray", "8abc123-EWR"), ("content-type", "text/html")],
        body: "<html><head><title>Fronted</title></head><body>hi</body></html>",
    },
];

/// Serves the routes above until the process ends. Bound to port 0 so
/// concurrent test binaries never collide.
fn start_server() -> u16 {
    start_counting_server(0).0
}

/// The same server, plus the number of `/robots.txt` requests it has
/// been sent and an optional run of leading failures on that one route.
///
/// Counting is how "one robots.txt per origin" is asserted: the win is a
/// request that no longer happens, and a count says that exactly, where
/// a duration would mostly measure whichever machine is running the
/// suite. `fail_first` closes the connection without writing a response,
/// which is what a dropped or refused one looks like from the client.
fn start_counting_server(fail_first: usize) -> (u16, Arc<AtomicUsize>) {
    let listener = TcpListener::bind("127.0.0.1:0").expect("bind");
    let port = listener.local_addr().expect("addr").port();
    let robots_requests = Arc::new(AtomicUsize::new(0));
    let counter = Arc::clone(&robots_requests);
    thread::spawn(move || {
        for stream in listener.incoming() {
            let Ok(stream) = stream else { continue };
            let counter = Arc::clone(&counter);
            thread::spawn(move || serve_one(stream, &counter, fail_first));
        }
    });
    (port, robots_requests)
}

fn serve_one(mut stream: TcpStream, robots_requests: &AtomicUsize, fail_first: usize) {
    let mut reader = BufReader::new(stream.try_clone().expect("clone"));
    let mut request_line = String::new();
    if reader.read_line(&mut request_line).is_err() {
        return;
    }
    let path = request_line.split_whitespace().nth(1).unwrap_or("/");

    if path == "/robots.txt" && robots_requests.fetch_add(1, Ordering::SeqCst) < fail_first {
        return;
    }

    // `/page/<n>` is any number of distinct, unremarkable pages on one
    // host — what a batch of fifty URLs on one site actually looks like,
    // which no fixed route table can supply. They answer as `/open`.
    let wanted = if path.starts_with("/page/") {
        "/open"
    } else {
        path
    };
    let route = ROUTES.iter().find(|route| route.path == wanted);
    let response = match route {
        Some(route) => {
            let mut head = format!("HTTP/1.1 {}\r\n", route.status);
            for (name, value) in route.headers {
                let _ = write!(head, "{name}: {value}\r\n");
            }
            let _ = write!(head, "content-length: {}\r\n", route.body.len());
            head.push_str("connection: close\r\n\r\n");
            format!("{head}{}", route.body)
        }
        None => {
            "HTTP/1.1 404 Not Found\r\ncontent-length: 0\r\nconnection: close\r\n\r\n".to_string()
        }
    };
    let _ = stream.write_all(response.as_bytes());
    let _ = stream.flush();
}

struct Run {
    code: i32,
    stdout: String,
}

/// Runs the built binary. `--no-render` throughout: these tests pin the
/// exit-code contract, which must hold on a machine with no browser.
fn run(args: &[&str]) -> Run {
    let binary = env!("CARGO_BIN_EXE_scrape-le");
    let mut command = Command::new(binary);
    command.arg("--no-render");
    command.args(args);
    let output = command.output().expect("binary runs");
    Run {
        code: output.status.code().expect("exit code"),
        stdout: String::from_utf8_lossy(&output.stdout).to_string(),
    }
}

fn verdict_of(stdout: &str) -> String {
    let line = stdout.lines().next().unwrap_or_default();
    let report: serde_json::Value = serde_json::from_str(line).expect("a JSON report");
    report["verdict"].as_str().unwrap_or_default().to_string()
}

#[test]
fn a_clean_page_without_render_is_inconclusive_and_exits_1() {
    let port = start_server();
    let run = run(&[&format!("http://127.0.0.1:{port}/open")]);
    // Never `clear`: the anti-bot and auth checks cannot complete
    // without a rendered page, and `clear` is a claim about absence.
    assert_eq!(verdict_of(&run.stdout), "inconclusive");
    assert_eq!(run.code, 1);
}

#[test]
fn a_robots_disallow_is_restricted_and_exits_1() {
    let port = start_server();
    let run = run(&[&format!("http://127.0.0.1:{port}/forbidden")]);
    assert_eq!(verdict_of(&run.stdout), "restricted");
    assert_eq!(run.code, 1);
    assert!(run.stdout.contains("Disallow: /forbidden"));
}

#[test]
fn an_agent_specific_group_only_applies_when_asked_for() {
    let port = start_server();
    let url = format!("http://127.0.0.1:{port}/open");

    let generic = run(&[&url]);
    assert_eq!(verdict_of(&generic.stdout), "inconclusive");

    let picky = run(&["--agent", "PickyBot/1.0", &url]);
    assert_eq!(verdict_of(&picky.stdout), "restricted");
    assert!(picky.stdout.contains("User-agent: pickybot"));
    assert_eq!(picky.code, 1);
}

#[test]
fn http_429_is_restricted_and_exits_1() {
    let port = start_server();
    let run = run(&[&format!("http://127.0.0.1:{port}/limited")]);
    assert_eq!(verdict_of(&run.stdout), "restricted");
    assert_eq!(run.code, 1);
    assert!(run.stdout.contains("rate limited right now"));
}

#[test]
fn http_401_is_restricted_and_exits_1() {
    let port = start_server();
    let run = run(&[&format!("http://127.0.0.1:{port}/secret")]);
    assert_eq!(verdict_of(&run.stdout), "restricted");
    assert_eq!(run.code, 1);
    assert!(run.stdout.contains("HTTP 401 Unauthorized"));
}

#[test]
fn a_vendor_header_warns_and_names_its_evidence() {
    let port = start_server();
    let run = run(&[&format!("http://127.0.0.1:{port}/cloudflare")]);
    assert_eq!(verdict_of(&run.stdout), "restricted");
    assert!(run.stdout.contains("Cloudflare (cf-ray header)"));
    assert!(run.stdout.contains("response-header"));
}

#[test]
fn an_unparseable_url_exits_2() {
    let run = run(&["not a url at all"]);
    assert_eq!(run.code, 2);
}

#[test]
fn a_dns_failure_exits_2() {
    let run = run(&["https://this-host-does-not-exist-9f2c1a.invalid/"]);
    assert_eq!(run.code, 2);
}

#[test]
fn an_unknown_flag_exits_2() {
    let run = run(&["--not-a-flag", "https://example.com"]);
    assert_eq!(run.code, 2);
}

#[test]
fn a_flag_missing_its_value_exits_2() {
    let run = run(&["--agent"]);
    assert_eq!(run.code, 2);
}

#[test]
fn help_and_version_exit_0() {
    for flag in ["--help", "--version"] {
        let run = run(&[flag]);
        assert_eq!(run.code, 0, "{flag}");
        assert!(!run.stdout.is_empty(), "{flag}");
    }
}

#[test]
fn a_batch_exits_with_its_worst_verdict() {
    let port = start_server();
    let directory = std::env::temp_dir().join(format!("scrape-le-contract-{port}"));
    std::fs::create_dir_all(&directory).expect("temp dir");
    let input = directory.join("urls.txt");
    std::fs::write(
        &input,
        format!("http://127.0.0.1:{port}/open\nhttp://127.0.0.1:{port}/forbidden\n"),
    )
    .expect("write input");

    let run = run(&["--input", input.to_str().expect("path")]);
    assert_eq!(run.code, 1);
    let verdicts: Vec<String> = run.stdout.lines().map(verdict_of).collect();
    assert_eq!(verdicts.len(), 2);
    assert!(verdicts.contains(&"restricted".to_string()));

    std::fs::remove_dir_all(&directory).ok();
}

#[test]
fn a_batch_reports_every_input_index() {
    let port = start_server();
    let directory = std::env::temp_dir().join(format!("scrape-le-index-{port}"));
    std::fs::create_dir_all(&directory).expect("temp dir");
    let input = directory.join("urls.json");
    std::fs::write(
        &input,
        format!(
            r#"["http://127.0.0.1:{port}/open","http://127.0.0.1:{port}/limited","http://127.0.0.1:{port}/open"]"#
        ),
    )
    .expect("write input");

    let run = run(&["--input", input.to_str().expect("path")]);
    let mut indices: Vec<u64> = run
        .stdout
        .lines()
        .map(|line| {
            let report: serde_json::Value = serde_json::from_str(line).expect("JSON");
            report["index"].as_u64().expect("index")
        })
        .collect();
    indices.sort_unstable();
    assert_eq!(indices, [0, 1, 2]);

    std::fs::remove_dir_all(&directory).ok();
}

/// Writes a batch input file and returns its path plus the directory to
/// clean up.
fn batch_input(port: u16, name: &str, urls: &[String]) -> (std::path::PathBuf, std::path::PathBuf) {
    let directory = std::env::temp_dir().join(format!("scrape-le-{name}-{port}"));
    std::fs::create_dir_all(&directory).expect("temp dir");
    let input = directory.join("urls.txt");
    std::fs::write(&input, urls.join("\n")).expect("write input");
    (input, directory)
}

/// **What grouping by host was always for.** Twenty paths on one origin
/// is one robots.txt request, not twenty — and the twenty answers are
/// the ones each path had when each fetched and parsed its own copy.
#[test]
fn a_batch_asks_one_origin_for_robots_txt_once() {
    let (port, robots_requests) = start_counting_server(0);
    let mut urls: Vec<String> = (0..19)
        .map(|i| format!("http://127.0.0.1:{port}/page/{i}"))
        .collect();
    urls.push(format!("http://127.0.0.1:{port}/forbidden"));
    let (input, directory) = batch_input(port, "cache", &urls);

    let run = run(&["--input", input.to_str().expect("path")]);

    assert_eq!(
        robots_requests.load(Ordering::SeqCst),
        1,
        "20 URLs on one origin asked for robots.txt more than once"
    );
    let verdicts: Vec<String> = run.stdout.lines().map(verdict_of).collect();
    assert_eq!(verdicts.len(), 20);
    assert_eq!(
        verdicts.iter().filter(|v| *v == "inconclusive").count(),
        19,
        "every unremarkable page answers as it does alone: {verdicts:?}"
    );
    assert_eq!(
        verdicts.iter().filter(|v| *v == "restricted").count(),
        1,
        "the disallowed path is still disallowed: {verdicts:?}"
    );
    assert!(run.stdout.contains("Disallow: /forbidden"));
    assert_eq!(run.code, 1);

    std::fs::remove_dir_all(&directory).ok();
}

/// **Regression.** A robots.txt that fails to arrive must not answer for
/// the rest of the run. The cache holds what an origin served, never a
/// failure to serve it, so the next URL on that host asks again — and
/// the rules that come back are the ones that decide. Caching the miss
/// would have had one dropped connection read as "nothing forbids you"
/// for every remaining URL on the host, which is the single direction
/// this tool may not be wrong in.
#[test]
fn a_failed_robots_txt_does_not_answer_for_the_rest_of_the_run() {
    let (port, robots_requests) = start_counting_server(1);
    let urls = [
        format!("http://127.0.0.1:{port}/page/0"),
        format!("http://127.0.0.1:{port}/forbidden"),
    ];
    let (input, directory) = batch_input(port, "retry", &urls);

    let run = run(&["--input", input.to_str().expect("path")]);

    // URLs on one host run in input order, so the first request is the
    // one that fails and the second is the retry.
    assert_eq!(robots_requests.load(Ordering::SeqCst), 2);
    let verdicts: Vec<String> = run.stdout.lines().map(verdict_of).collect();
    assert!(
        verdicts.contains(&"restricted".to_string()),
        "the retry's rules never reached the second URL: {verdicts:?}"
    );
    assert!(run.stdout.contains("Disallow: /forbidden"));

    std::fs::remove_dir_all(&directory).ok();
}

/// The CLI and the MCP server must answer the same URL identically —
/// asserted directly, so neither surface can drift from the other.
#[test]
fn the_two_surfaces_return_identical_findings() {
    let port = start_server();
    let url = format!("http://127.0.0.1:{port}/forbidden");

    let cli = run(&[&url]);
    let cli_report: serde_json::Value =
        serde_json::from_str(cli.stdout.lines().next().expect("a report")).expect("JSON");

    let request = format!(
        r#"{{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{{"name":"scrape_le_check","arguments":{{"url":"{url}","render":false}}}}}}"#
    );
    let binary = env!("CARGO_BIN_EXE_scrape-le");
    let mut child = Command::new(binary)
        .arg("mcp")
        .stdin(Stdio::piped())
        .stdout(Stdio::piped())
        .stderr(Stdio::null())
        .spawn()
        .expect("mcp starts");
    child
        .stdin
        .as_mut()
        .expect("stdin")
        .write_all(format!("{request}\n").as_bytes())
        .expect("write request");
    drop(child.stdin.take());
    let mut raw = String::new();
    child
        .stdout
        .as_mut()
        .expect("stdout")
        .read_to_string(&mut raw)
        .expect("read response");
    child.wait().expect("mcp exits");

    let response: serde_json::Value =
        serde_json::from_str(raw.lines().next().expect("a response")).expect("JSON");
    let envelope = &response["result"]["structuredContent"];
    let mcp_report = &envelope["data"];

    assert_eq!(cli_report["verdict"], mcp_report["verdict"]);
    assert_eq!(cli_report["findings"], mcp_report["findings"]);
    assert_eq!(cli_report["checks"], mcp_report["checks"]);
    // `ok` says the check ran; the answer is the verdict. A restricted
    // page must not read as a broken tool.
    assert_eq!(envelope["ok"], true);
    assert_eq!(mcp_report["verdict"], "restricted");
    assert_eq!(response["result"]["isError"], false);
}