zoey-core 0.1.1

ZoeyAI core runtime and types — privacy-first AI agent framework optimized for local models
Documentation
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
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
use crate::detectors::analyze_all;
use crate::nlp::{double_metaphone, normalized_similarity};
use crate::pipeline::mps::*;
use crate::planner::complexity::ComplexityAssessment;
use crate::types::*;
use crate::Result;
use std::sync::{Arc, RwLock};
use tokio::task::JoinSet;
use tracing::{debug, info};

#[derive(Clone)]
pub struct Phase0Output {
    pub language: Option<String>,
    pub intent: Option<String>,
    pub sentiment: Option<String>,
    pub tone: Option<String>,
    pub topics: Vec<String>,
    pub keywords: Vec<String>,
    pub entities: Vec<String>,
    pub complexity: Option<ComplexityAssessment>,
    pub agent_candidates: Vec<String>,
    pub available_actions: Vec<String>,
}

impl Phase0Output {
    pub fn new() -> Self {
        Self {
            language: None,
            intent: None,
            sentiment: None,
            tone: None,
            topics: Vec::new(),
            keywords: Vec::new(),
            entities: Vec::new(),
            complexity: None,
            agent_candidates: Vec::new(),
            available_actions: Vec::new(),
        }
    }
}

pub struct Phase0Preprocessor {
    runtime: Arc<RwLock<crate::AgentRuntime>>,
}

impl Phase0Preprocessor {
    pub fn new(runtime: Arc<RwLock<crate::AgentRuntime>>) -> Self {
        Self { runtime }
    }

    pub async fn execute(&self, message: &Memory) -> Result<Phase0Output> {
        let mut js: JoinSet<(String, serde_json::Value, i64)> = JoinSet::new();
        let text = message.content.text.clone();

        {
            let t = text.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let lang = mp1_language(&t);
                (
                    "phase0:language".to_string(),
                    serde_json::json!(lang),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let t = text.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let (intent, sentiment, tone) = mp2_classify(&t);
                (
                    "phase0:classification".to_string(),
                    serde_json::json!({
                        "intent": intent,
                        "sentiment": sentiment,
                        "tone": tone
                    }),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let t = text.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let (topics, keywords) = mp3_topics_keywords(&t);
                (
                    "phase0:topics_keywords".to_string(),
                    serde_json::json!({
                        "topics": topics,
                        "keywords": keywords
                    }),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let t = text.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let entities = mp4_entities_simple(&t);
                (
                    "phase0:entities".to_string(),
                    serde_json::json!(entities),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let t = text.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let det = analyze_all(&t, 0);
                (
                    "phase0:detectors".to_string(),
                    serde_json::to_value(&det).unwrap_or(serde_json::Value::Null),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let rt = self.runtime.clone();
            let msg = message.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let assessment = mp5_complexity(rt.clone(), &msg).await;
                (
                    "phase0:complexity".to_string(),
                    serde_json::to_value(&assessment).unwrap_or(serde_json::Value::Null),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let rt_arc = self.runtime.clone();
            let enabled = {
                let rt = rt_arc.read().unwrap();
                rt.get_setting("ui:phase0_embeddings")
                    .and_then(|v| v.as_bool())
                    .unwrap_or(false)
            };
            let adapter_opt = rt_arc.read().unwrap().adapter.read().unwrap().clone();
            let handler_opt = rt_arc
                .read()
                .unwrap()
                .models
                .read()
                .unwrap()
                .get("TEXT_EMBEDDING")
                .and_then(|v| v.first().map(|p| p.handler.clone()));
            let prompt_text = {
                let rt = rt_arc.read().unwrap();
                rt.get_setting("ui:lastPrompt")
                    .and_then(|v| v.as_str().map(|s| s.to_string()))
                    .unwrap_or_else(|| text.clone())
            };
            let msg_owned = message.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let done = mp6_embedding_queue(rt_arc.clone(), &msg_owned, prompt_text).await;
                (
                    "phase0:embedding_queue".to_string(),
                    serde_json::json!(done),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let rt = self.runtime.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let actions = mp7_available_actions(rt.clone());
                (
                    "phase0:available_actions".to_string(),
                    serde_json::json!(actions),
                    st.elapsed().as_millis() as i64,
                )
            });
        }
        {
            let rt = self.runtime.clone();
            js.spawn(async move {
                let st = std::time::Instant::now();
                let candidates: Vec<String> = mp8_agent_candidates(rt.clone());
                (
                    "phase0:agent_candidates".to_string(),
                    serde_json::json!(candidates),
                    st.elapsed().as_millis() as i64,
                )
            });
        }

        let mut out = Phase0Output::new();
        let start = std::time::Instant::now();
        let mut profiles: Vec<(String, i64)> = Vec::new();
        while let Some(res) = js.join_next().await {
            if let Ok((key, val, dur)) = res {
                let mut rt = self.runtime.write().unwrap();
                rt.set_setting(&key, val.clone(), false);
                profiles.push((key.clone(), dur));
                match key.as_str() {
                    "phase0:language" => {
                        out.language = val.as_str().map(|s| s.to_string());
                    }
                    "phase0:classification" => {
                        if let Some(obj) = val.as_object() {
                            out.intent = obj
                                .get("intent")
                                .and_then(|v| v.as_str().map(|s| s.to_string()));
                            out.sentiment = obj
                                .get("sentiment")
                                .and_then(|v| v.as_str().map(|s| s.to_string()));
                            out.tone = obj
                                .get("tone")
                                .and_then(|v| v.as_str().map(|s| s.to_string()));
                        }
                    }
                    "phase0:topics_keywords" => {
                        if let Some(obj) = val.as_object() {
                            if let Some(arr) = obj.get("topics").and_then(|v| v.as_array()) {
                                out.topics = arr
                                    .iter()
                                    .filter_map(|v| v.as_str().map(|s| s.to_string()))
                                    .collect();
                            }
                            if let Some(arr) = obj.get("keywords").and_then(|v| v.as_array()) {
                                out.keywords = arr
                                    .iter()
                                    .filter_map(|v| v.as_str().map(|s| s.to_string()))
                                    .collect();
                                let (metaphones, sim_hint) =
                                    mp9_keywords_phonetic_similarity(&out.keywords);
                                rt.set_setting(
                                    "phase0:keywords:phonetic",
                                    serde_json::json!(metaphones),
                                    false,
                                );
                                rt.set_setting(
                                    "phase0:keywords:similarity_hint",
                                    serde_json::json!(sim_hint),
                                    false,
                                );
                                // If embeddings are queued, schedule background task
                                let queued = rt
                                    .get_setting("phase0:embedding:queued")
                                    .and_then(|v| v.as_bool())
                                    .unwrap_or(false);
                                // Background embedding will be processed by runtime services
                            }
                        }
                    }
                    "phase0:entities" => {
                        if let Some(arr) = val.as_array() {
                            out.entities = arr
                                .iter()
                                .filter_map(|v| v.as_str().map(|s| s.to_string()))
                                .collect();
                        }
                    }
                    "phase0:detectors" => {
                        if let Some(obj) = val.as_object() {
                            if let Some(amb) = obj.get("ambiguity_score").and_then(|v| v.as_f64()) {
                                let mut rt = self.runtime.write().unwrap();
                                rt.set_setting("ui:ambiguity", serde_json::json!(amb), false);
                            }
                            if let Some(hint) =
                                obj.get("urgency_markers").and_then(|v| v.as_array())
                            {
                                let mut rt = self.runtime.write().unwrap();
                                rt.set_setting(
                                    "ui:urgency_markers",
                                    serde_json::json!(hint.len()),
                                    false,
                                );
                            }
                            if let Some(incomplete) =
                                obj.get("incomplete").and_then(|v| v.as_bool())
                            {
                                let mut rt = self.runtime.write().unwrap();
                                rt.set_setting(
                                    "ui:incomplete",
                                    serde_json::json!(incomplete),
                                    false,
                                );
                            }
                        }
                    }
                    "phase0:complexity" => {
                        if !val.is_null() {
                            if let Ok(assess) =
                                serde_json::from_value::<ComplexityAssessment>(val.clone())
                            {
                                out.complexity = Some(assess);
                            }
                        }
                    }
                    "phase0:available_actions" => {
                        if let Some(arr) = val.as_array() {
                            out.available_actions = arr
                                .iter()
                                .filter_map(|v| v.as_str().map(|s| s.to_string()))
                                .collect();
                        }
                    }
                    "phase0:embedding_queue" => {}
                    "phase0:agent_candidates" => {
                        if let Some(arr) = val.as_array() {
                            out.agent_candidates = arr
                                .iter()
                                .filter_map(|v| v.as_str().map(|s| s.to_string()))
                                .collect();
                        }
                    }
                    _ => {}
                }
            }
        }
        let elapsed = start.elapsed().as_millis() as i64;
        profiles.sort_by(|a, b| b.1.cmp(&a.1));
        let top: Vec<String> = profiles
            .iter()
            .take(3)
            .map(|(k, d)| format!("{}:{}ms", k, d))
            .collect();
        debug!(total_ms = elapsed, top = ?top, "phase0_profiles");
        {
            let mut rt = self.runtime.write().unwrap();
            rt.set_setting(
                "phase0:profile",
                serde_json::json!({"total_ms": elapsed, "top": top}),
                false,
            );
        }
        info!(duration_ms = elapsed, "phase0_complete");
        Ok(out)
    }
}

fn detect_language(text: &str) -> String {
    let s = text.to_lowercase();
    if s.contains("¿") || s.contains("¡") || s.contains(" hola ") || s.contains(" gracias ") {
        return "es".to_string();
    }
    if s.contains(" bonjour ") || s.contains(" merci ") || s.contains(" français ") {
        return "fr".to_string();
    }
    if s.contains(" hallo ") || s.contains(" danke ") {
        return "de".to_string();
    }
    "en".to_string()
}

fn classify_text(text: &str) -> (String, String, String) {
    let t = text.to_lowercase();
    let intent =
        if t.contains('?') || t.starts_with("what") || t.starts_with("how") || t.starts_with("why")
        {
            "Question"
        } else if t.contains("please") || t.starts_with("can you") {
            "Request"
        } else if t.contains("hello") || t.contains("hi") {
            "Greeting"
        } else {
            "Statement"
        };
    let positive = ["great", "thanks", "awesome", "good", "love"];
    let negative = ["bad", "hate", "terrible", "awful", "worse"];
    let mut pos = 0;
    let mut neg = 0;
    for w in positive.iter() {
        if t.contains(w) {
            pos += 1;
        }
    }
    for w in negative.iter() {
        if t.contains(w) {
            neg += 1;
        }
    }
    let sentiment = if pos > neg {
        "Positive"
    } else if neg > pos {
        "Negative"
    } else {
        "Neutral"
    };
    let tone = if t.contains("please") || t.contains("thank") {
        "Formal"
    } else if t.contains("lol") || t.contains(":)") {
        "Casual"
    } else {
        "Professional"
    };
    (intent.to_string(), sentiment.to_string(), tone.to_string())
}

fn extract_topics_keywords(text: &str) -> (Vec<String>, Vec<String>) {
    let stop = [
        "the", "a", "an", "and", "or", "to", "of", "in", "on", "for", "with", "is", "it", "this",
        "that", "i", "you", "we", "they", "be", "are", "was", "were", "as", "at",
    ];
    let mut freq: std::collections::HashMap<String, usize> = std::collections::HashMap::new();
    for token in text
        .split(|c: char| !c.is_alphanumeric())
        .filter(|s| !s.is_empty())
    {
        let lc = token.to_lowercase();
        if stop.contains(&lc.as_str()) {
            continue;
        }
        *freq.entry(lc).or_insert(0) += 1;
    }
    let mut items: Vec<(String, usize)> = freq.into_iter().collect();
    items.sort_by(|a, b| b.1.cmp(&a.1));
    let keywords: Vec<String> = items.iter().take(8).map(|(k, _)| k.clone()).collect();
    let topics = keywords.iter().take(4).cloned().collect();
    (topics, keywords)
}

fn extract_entities_simple(text: &str) -> Vec<String> {
    let mut entities = Vec::new();
    let mut current = Vec::new();
    for w in text.split_whitespace() {
        let first = w.chars().next();
        if let Some(c) = first {
            if c.is_uppercase() {
                current.push(w.trim_matches(|ch: char| !ch.is_alphanumeric()).to_string());
            } else {
                if current.len() >= 1 {
                    entities.push(current.join(" "));
                    current.clear();
                }
            }
        } else {
            if current.len() >= 1 {
                entities.push(current.join(" "));
                current.clear();
            }
        }
    }
    if !current.is_empty() {
        entities.push(current.join(" "));
    }
    let mut uniq = std::collections::HashSet::new();
    let mut out = Vec::new();
    for e in entities {
        if !e.is_empty() && uniq.insert(e.to_lowercase()) {
            out.push(e);
        }
    }
    out
}