anno 0.10.0

NER, coreference resolution, relation extraction, PII detection, and zero-shot entity types
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

use super::*;

// -------------------------------------------------------------------------
// Config / struct smoke tests
// -------------------------------------------------------------------------

#[test]
fn test_coref_config_default() {
    let config = T5CorefConfig::default();
    assert_eq!(config.max_input_length, 512);
    assert_eq!(config.num_beams, 1);
}

#[test]
fn test_cluster_struct() {
    let cluster = CorefCluster {
        id: 0,
        mentions: vec!["Marie Curie".to_string(), "She".to_string()],
        spans: vec![(0, 11), (50, 53)],
        canonical: "Marie Curie".to_string(),
    };
    assert_eq!(cluster.mentions.len(), 2);
    assert_eq!(cluster.canonical, "Marie Curie");
}

// -------------------------------------------------------------------------
// mark_mentions_for_t5
// -------------------------------------------------------------------------

#[test]
fn mark_mentions_wraps_pronouns() {
    let out = mark_mentions_for_t5("she said he agreed");
    assert!(
        out.contains("<m> she </m>"),
        "pronoun 'she' should be marked"
    );
    assert!(out.contains("<m> he </m>"), "pronoun 'he' should be marked");
    assert!(out.contains("said"), "'said' should pass through unmarked");
    assert!(
        out.contains("agreed"),
        "'agreed' should pass through unmarked"
    );
}

#[test]
fn mark_mentions_wraps_capitalized() {
    let out = mark_mentions_for_t5("Sophie Wilson designed ARM.");
    assert!(out.contains("<m> Sophie </m>"));
    assert!(out.contains("<m> Wilson </m>"));
    assert!(out.contains("<m> ARM. </m>"));
    assert!(out.contains("designed"));
}

#[test]
fn mark_mentions_empty_string() {
    assert_eq!(mark_mentions_for_t5(""), "");
}

#[test]
fn mark_mentions_no_pronouns_no_caps() {
    let text = "the quick brown fox";
    assert_eq!(mark_mentions_for_t5(text), text);
}

// -------------------------------------------------------------------------
// extract_t5_mentions
// -------------------------------------------------------------------------

#[test]
fn extract_mentions_basic() {
    let (plain, spans) = extract_t5_mentions("<m> Sophie Wilson </m> designed ARM.").unwrap();
    assert_eq!(plain, "Sophie Wilson designed ARM.");
    assert_eq!(spans.len(), 1);
    assert_eq!(spans[0].0, "Sophie Wilson");
}

#[test]
fn extract_mentions_two_spans() {
    let marked = "<m> Ada </m> founded <m> Lovelace Labs </m>.";
    let (plain, spans) = extract_t5_mentions(marked).unwrap();
    assert_eq!(spans.len(), 2);
    assert_eq!(spans[0].0, "Ada");
    assert_eq!(spans[1].0, "Lovelace Labs");
    assert!(plain.contains("Ada"));
    assert!(plain.contains("Lovelace Labs"));
    assert!(plain.contains("founded"));
}

#[test]
fn extract_mentions_empty_input() {
    let (plain, spans) = extract_t5_mentions("").unwrap();
    assert_eq!(plain, "");
    assert!(spans.is_empty());
}

#[test]
fn extract_mentions_no_markers() {
    let (plain, spans) = extract_t5_mentions("no markers here").unwrap();
    assert_eq!(plain, "no markers here");
    assert!(spans.is_empty());
}

#[test]
fn extract_mentions_span_offsets_are_consistent() {
    let marked = "<m> Marie </m> discovered polonium. <m> She </m> won.";
    let (plain, spans) = extract_t5_mentions(marked).unwrap();
    for (text, start, end) in &spans {
        let extracted: String = plain.chars().skip(*start).take(end - start).collect();
        assert_eq!(
            &extracted, text,
            "span offsets must index back to mention text"
        );
    }
}

// -------------------------------------------------------------------------
// parse_t5_coref_output
// -------------------------------------------------------------------------

#[test]
fn parse_output_basic_two_mentions() {
    // "Ada | 1 founded Lovelace Labs | 2. She | 1 led ENIAC."
    let decoded = "Ada | 1 founded Lovelace Labs | 2. She | 1 led ENIAC.";
    let clusters = parse_t5_coref_output(decoded);
    let cluster1 = clusters.iter().find(|c| c.id == 1);
    assert!(cluster1.is_some(), "cluster 1 (Ada/She) should be parsed");
    let c = cluster1.unwrap();
    assert!(c.mentions.len() >= 2, "cluster 1 should have ≥2 members");
    assert!(c.mentions.contains(&"Ada".to_string()));
    assert!(c.mentions.contains(&"She".to_string()));
}

#[test]
fn parse_output_singletons_filtered() {
    // Word with marker but only appearing once → should be filtered (singleton)
    let decoded = "Marie | 1 discovered polonium.";
    let clusters = parse_t5_coref_output(decoded);
    assert!(
        clusters.is_empty(),
        "singleton cluster should be filtered out"
    );
}

#[test]
fn parse_output_empty_string() {
    assert!(parse_t5_coref_output("").is_empty());
}

#[test]
fn parse_output_no_markers() {
    let clusters = parse_t5_coref_output("This text has no cluster markers at all.");
    assert!(clusters.is_empty());
}

#[test]
fn parse_output_canonical_is_longest_mention() {
    // Each token gets a marker individually.
    // "Marie | 1 Curie | 1" produces mentions ["Marie", "Curie"] for cluster 1.
    // With "She" also in cluster 1, the longest single-token mention is "Marie" (5 chars)
    // vs "Curie" (5) vs "She" (3) — tie between Marie/Curie, resolved by first-seen.
    // Use a clearer example: "Alexandra | 1 won . She | 1 left."
    let decoded = "Alexandra | 1 won . She | 1 left .";
    let clusters = parse_t5_coref_output(decoded);
    if let Some(c) = clusters.iter().find(|c| c.id == 1) {
        assert_eq!(
            c.canonical, "Alexandra",
            "canonical should be the longest mention"
        );
    }
}

#[test]
fn parse_output_sorted_by_cluster_id() {
    let decoded = "B | 2 and A | 1 were colleagues . them | 2 worked together . B | 1 also .";
    let clusters = parse_t5_coref_output(decoded);
    if clusters.len() >= 2 {
        let ids: Vec<u32> = clusters.iter().map(|c| c.id).collect();
        let mut sorted = ids.clone();
        sorted.sort();
        assert_eq!(ids, sorted, "clusters should be sorted by ID");
    }
}

// -------------------------------------------------------------------------
// Additional pure-function tests
// -------------------------------------------------------------------------

#[test]
fn config_custom_values() {
    let config = T5CorefConfig {
        max_input_length: 256,
        max_output_length: 128,
        num_beams: 4,
        optimization_level: 1,
        num_threads: 2,
    };
    assert_eq!(config.max_input_length, 256);
    assert_eq!(config.max_output_length, 128);
    assert_eq!(config.num_beams, 4);
    assert_eq!(config.optimization_level, 1);
    assert_eq!(config.num_threads, 2);
}

#[test]
fn cluster_clone_is_independent() {
    let original = CorefCluster {
        id: 1,
        mentions: vec!["Ada Lovelace".into(), "She".into()],
        spans: vec![(0, 12), (30, 33)],
        canonical: "Ada Lovelace".into(),
    };
    let mut cloned = original.clone();
    cloned.mentions.push("her".into());
    assert_eq!(original.mentions.len(), 2, "original must be unaffected");
    assert_eq!(cloned.mentions.len(), 3);
}

#[test]
fn mark_mentions_mixed_pronouns_and_caps() {
    let out = mark_mentions_for_t5("Alice told him about Bob");
    assert!(out.contains("<m> Alice </m>"), "capitalized 'Alice' marked");
    assert!(out.contains("<m> him </m>"), "pronoun 'him' marked");
    assert!(out.contains("<m> Bob </m>"), "capitalized 'Bob' marked");
    assert!(out.contains("told"), "'told' unmarked");
    assert!(out.contains("about"), "'about' unmarked");
}

#[test]
fn mark_mentions_possessive_pronouns() {
    let out = mark_mentions_for_t5("his car and their house");
    assert!(out.contains("<m> his </m>"));
    assert!(out.contains("<m> their </m>"));
    // "car", "and", "house" should not be marked
    assert!(!out.contains("<m> car </m>"));
    assert!(!out.contains("<m> and </m>"));
    assert!(!out.contains("<m> house </m>"));
}

#[test]
fn extract_mentions_unclosed_tag_keeps_remaining_text() {
    // An unclosed <m> tag should not panic; the remainder is appended as plain text.
    let (plain, spans) = extract_t5_mentions("before <m> orphan text after").unwrap();
    assert!(
        plain.contains("before"),
        "text before unclosed tag preserved"
    );
    assert!(
        plain.contains("orphan text after"),
        "text inside unclosed tag appended as plain"
    );
    assert!(spans.is_empty(), "no complete span from unclosed tag");
}

#[test]
fn extract_mentions_adjacent_spans() {
    let marked = "<m> A </m><m> B </m>";
    let (plain, spans) = extract_t5_mentions(marked).unwrap();
    assert_eq!(spans.len(), 2);
    assert_eq!(spans[0].0, "A");
    assert_eq!(spans[1].0, "B");
    // Offsets must not overlap
    assert!(
        spans[0].2 <= spans[1].1,
        "first span end ({}) must be <= second span start ({})",
        spans[0].2,
        spans[1].1
    );
    // Round-trip: extracting from plain text at the stored offsets must match
    for (text, start, end) in &spans {
        let extracted: String = plain.chars().skip(*start).take(end - start).collect();
        assert_eq!(&extracted, text);
    }
}

#[test]
fn parse_output_multiple_distinct_clusters() {
    // Two separate clusters: 1 (Alice/She) and 2 (Bob/He)
    let decoded = "Alice | 1 met Bob | 2 yesterday . She | 1 greeted He | 2 warmly .";
    let clusters = parse_t5_coref_output(decoded);
    assert_eq!(clusters.len(), 2, "two multi-mention clusters expected");
    let c1 = clusters.iter().find(|c| c.id == 1).unwrap();
    let c2 = clusters.iter().find(|c| c.id == 2).unwrap();
    assert!(c1.mentions.contains(&"Alice".to_string()));
    assert!(c1.mentions.contains(&"She".to_string()));
    assert!(c2.mentions.contains(&"Bob".to_string()));
    assert!(c2.mentions.contains(&"He".to_string()));
}

#[test]
fn parse_output_cluster_id_with_trailing_punctuation() {
    // Cluster ID followed by punctuation (e.g. "1." or "1,") should still parse
    let decoded = "Marie | 1 discovered radium . She | 1 won .";
    let clusters = parse_t5_coref_output(decoded);
    assert_eq!(clusters.len(), 1);
    let c = &clusters[0];
    assert_eq!(c.id, 1);
    assert!(c.mentions.contains(&"Marie".to_string()));
    assert!(c.mentions.contains(&"She".to_string()));
}