arete-interpreter 0.2.0

AST transformation runtime and VM for Arete streaming pipelines
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
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
use std::collections::HashMap;
use std::future::Future;
use std::pin::Pin;
use std::sync::OnceLock;
use std::time::Instant;

use serde::{Deserialize, Serialize};
use serde_json::Value;

use crate::ast::{HttpMethod, ResolverType};
use crate::compiler::MultiEntityBytecode;
use crate::resolvers::{TokenMetadataResolverClient, UrlResolverClient};
use crate::vm::{
    CachedResolverValue, ResolverRequest, TokenResolverMissAction, UpdateContext, VmContext,
};
use crate::Mutation;

pub type ResolverBatchResult =
    Result<HashMap<String, Value>, Box<dyn std::error::Error + Send + Sync>>;
pub type ResolverBatchFuture<'a> = Pin<Box<dyn Future<Output = ResolverBatchResult> + Send + 'a>>;
pub type ResolverApplyFuture<'a> = Pin<Box<dyn Future<Output = Vec<Mutation>> + Send + 'a>>;
pub type SharedRuntimeResolver = std::sync::Arc<dyn RuntimeResolver>;

#[derive(Clone, Copy)]
struct RuntimeResolverProfileConfig {
    enabled: bool,
    slow_ms: f64,
}

fn runtime_resolver_profile_config() -> RuntimeResolverProfileConfig {
    static CONFIG: OnceLock<RuntimeResolverProfileConfig> = OnceLock::new();
    *CONFIG.get_or_init(|| RuntimeResolverProfileConfig {
        enabled: std::env::var("ARETE_PROFILE_CODEPATHS")
            .ok()
            .or_else(|| std::env::var("ARETE_PROFILE_RUNTIME_RESOLVER").ok())
            .map(|value| {
                matches!(
                    value.to_ascii_lowercase().as_str(),
                    "1" | "true" | "yes" | "on"
                )
            })
            .unwrap_or(false),
        slow_ms: std::env::var("ARETE_PROFILE_CODEPATHS_SLOW_MS")
            .ok()
            .or_else(|| std::env::var("ARETE_PROFILE_RUNTIME_RESOLVER_SLOW_MS").ok())
            .and_then(|value| value.parse::<f64>().ok())
            .unwrap_or(5.0),
    })
}

fn elapsed_ms(start: Instant) -> f64 {
    start.elapsed().as_secs_f64() * 1000.0
}

fn should_log_runtime_resolver_profile(total_ms: f64) -> bool {
    let config = runtime_resolver_profile_config();
    config.enabled && total_ms >= config.slow_ms
}

#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq, Hash)]
#[serde(tag = "type", rename_all = "snake_case")]
pub enum RuntimeResolverRequest {
    TokenMetadata {
        key: String,
        mint: String,
    },
    UrlJson {
        key: String,
        url: String,
        method: HttpMethod,
    },
}

impl RuntimeResolverRequest {
    pub fn key(&self) -> &str {
        match self {
            Self::TokenMetadata { key, .. } | Self::UrlJson { key, .. } => key,
        }
    }
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct RuntimeResolverBatchRequest {
    pub requests: Vec<RuntimeResolverRequest>,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct RuntimeResolverResponse {
    pub key: String,
    pub value: Value,
}

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct RuntimeResolverBatchResponse {
    #[serde(default)]
    pub resolved: Vec<RuntimeResolverResponse>,
}

#[derive(Debug)]
struct PendingRuntimeResolverRequest {
    request: ResolverRequest,
    backend_request: RuntimeResolverRequest,
}

pub trait RuntimeResolver: Send + Sync {
    fn resolve_batch<'a>(
        &'a self,
        requests: &'a [RuntimeResolverRequest],
    ) -> ResolverBatchFuture<'a>;

    fn resolve_and_apply<'a>(
        &'a self,
        vm: &'a std::sync::Mutex<VmContext>,
        bytecode: &'a MultiEntityBytecode,
        requests: Vec<ResolverRequest>,
        apply_context: Option<UpdateContext>,
    ) -> ResolverApplyFuture<'a> {
        Box::pin(async move {
            let total_started_at = Instant::now();
            if requests.is_empty() {
                return Vec::new();
            }

            let request_count = requests.len();
            let mut cached = Vec::new();
            let mut pending = Vec::new();
            let mut invalid = Vec::new();
            let mut negative_cached = Vec::new();

            let classify_started_at = Instant::now();
            let (
                classify_ms,
                cached_count,
                pending_count,
                invalid_count,
                negative_cached_count,
                unique_request_count,
            ) = {
                let mut vm_guard = vm.lock().unwrap_or_else(|e| e.into_inner());

                for request in requests {
                    let canonical_key =
                        runtime_resolver_cache_key(&request.resolver, &request.input);

                    if let Some(cached_value) = vm_guard.get_cached_resolver_value(&canonical_key) {
                        match cached_value {
                            CachedResolverValue::Resolved(resolved_value) => {
                                cached.push((request, resolved_value));
                            }
                            CachedResolverValue::Negative => {
                                negative_cached.push(request);
                            }
                        }
                        continue;
                    }

                    match runtime_request_from_vm_request(&request) {
                        Some(backend_request) => pending.push(PendingRuntimeResolverRequest {
                            request,
                            backend_request,
                        }),
                        None => invalid.push(request),
                    }
                }

                let invalid_count = invalid.len();
                if !invalid.is_empty() {
                    vm_guard.restore_resolver_requests(invalid);
                }

                let classify_ms = elapsed_ms(classify_started_at);
                let cached_count = cached.len();
                let pending_count = pending.len();
                let negative_cached_count = negative_cached.len();
                let unique_request_count = if pending.is_empty() {
                    0
                } else {
                    let mut unique = HashMap::new();
                    for entry in &pending {
                        unique
                            .entry(entry.backend_request.key().to_string())
                            .or_insert_with(|| entry.backend_request.clone());
                    }
                    unique.len()
                };

                (
                    classify_ms,
                    cached_count,
                    pending_count,
                    invalid_count,
                    negative_cached_count,
                    unique_request_count,
                )
            };

            let resolve_batch_started_at = Instant::now();
            let resolved_map = if pending.is_empty() {
                Ok(HashMap::new())
            } else {
                let mut unique = HashMap::new();
                for entry in &pending {
                    unique
                        .entry(entry.backend_request.key().to_string())
                        .or_insert_with(|| entry.backend_request.clone());
                }

                let unique_requests: Vec<RuntimeResolverRequest> = unique.into_values().collect();
                self.resolve_batch(&unique_requests).await
            };
            let resolve_batch_ms = elapsed_ms(resolve_batch_started_at);

            let mut mutations = Vec::new();
            let mut failed = Vec::new();
            let mut retry_scheduled_count = 0usize;
            let mut negative_cached_miss_count = 0usize;

            let apply_cached_started_at = Instant::now();
            let mut vm_guard = vm.lock().unwrap_or_else(|e| e.into_inner());
            let previous_context = if apply_context.is_some() {
                let previous = vm_guard.current_context().cloned();
                vm_guard.set_current_context(apply_context.clone());
                previous
            } else {
                None
            };

            for (request, resolved_value) in cached {
                match vm_guard.apply_resolver_result(bytecode, &request.cache_key, resolved_value) {
                    Ok(mut new_mutations) => mutations.append(&mut new_mutations),
                    Err(err) => {
                        tracing::warn!(
                            cache_key = %request.cache_key,
                            error = %err,
                            "Failed to apply cached resolver result"
                        );
                        failed.push(request);
                    }
                }
            }
            let apply_cached_ms = elapsed_ms(apply_cached_started_at);

            let apply_pending_started_at = Instant::now();
            match resolved_map {
                Ok(resolved_map) => {
                    for entry in pending {
                        match resolved_map.get(entry.backend_request.key()) {
                            Some(resolved_value) => match vm_guard.apply_resolver_result(
                                bytecode,
                                &entry.request.cache_key,
                                resolved_value.clone(),
                            ) {
                                Ok(mut new_mutations) => mutations.append(&mut new_mutations),
                                Err(err) => {
                                    tracing::warn!(
                                        cache_key = %entry.request.cache_key,
                                        error = %err,
                                        "Failed to apply resolver result"
                                    );
                                    failed.push(entry.request);
                                }
                            },
                            None => match entry.backend_request {
                                RuntimeResolverRequest::TokenMetadata { .. } => {
                                    match vm_guard
                                        .handle_token_resolver_miss(&entry.request.cache_key)
                                    {
                                        TokenResolverMissAction::RetryScheduled => {
                                            retry_scheduled_count += 1;
                                        }
                                        TokenResolverMissAction::NegativeCached => {
                                            negative_cached_miss_count += 1;
                                        }
                                        TokenResolverMissAction::Dropped => {}
                                    }
                                }
                                RuntimeResolverRequest::UrlJson { .. } => {
                                    failed.push(entry.request)
                                }
                            },
                        }
                    }
                }
                Err(err) => {
                    tracing::warn!(error = %err, "Runtime resolver backend request failed");
                    for entry in pending {
                        match entry.backend_request {
                            RuntimeResolverRequest::TokenMetadata { .. } => {
                                match vm_guard.handle_token_resolver_miss(&entry.request.cache_key)
                                {
                                    TokenResolverMissAction::RetryScheduled => {
                                        retry_scheduled_count += 1;
                                    }
                                    TokenResolverMissAction::NegativeCached => {
                                        negative_cached_miss_count += 1;
                                    }
                                    TokenResolverMissAction::Dropped => {}
                                }
                            }
                            RuntimeResolverRequest::UrlJson { .. } => failed.push(entry.request),
                        }
                    }
                }
            }
            let apply_pending_ms = elapsed_ms(apply_pending_started_at);

            let failed_count = failed.len();
            let restore_failed_started_at = Instant::now();
            if !failed.is_empty() {
                vm_guard.restore_resolver_requests(failed);
            }
            let restore_failed_ms = elapsed_ms(restore_failed_started_at);
            if apply_context.is_some() {
                vm_guard.set_current_context(previous_context);
            }

            let total_ms = elapsed_ms(total_started_at);
            if should_log_runtime_resolver_profile(total_ms) {
                tracing::info!(
                    stage = "runtime.resolver_apply",
                    request_count,
                    cached_count,
                    pending_count,
                    invalid_count,
                    negative_cached_count,
                    unique_request_count,
                    failed_count,
                    retry_scheduled_count,
                    negative_cached_miss_count,
                    mutation_count = mutations.len(),
                    total_ms,
                    classify_ms,
                    resolve_batch_ms,
                    apply_cached_ms,
                    apply_pending_ms,
                    restore_failed_ms,
                    "Runtime resolver profile"
                );
            }

            mutations
        })
    }
}

pub struct InProcessResolver {
    token_client: Option<TokenMetadataResolverClient>,
    url_client: UrlResolverClient,
}

impl InProcessResolver {
    pub fn from_env() -> Result<Self, Box<dyn std::error::Error + Send + Sync>> {
        Ok(Self {
            token_client: TokenMetadataResolverClient::from_env()?,
            url_client: UrlResolverClient::new(),
        })
    }

    pub fn new(
        token_client: Option<TokenMetadataResolverClient>,
        url_client: UrlResolverClient,
    ) -> Self {
        Self {
            token_client,
            url_client,
        }
    }

    pub async fn resolve_batch_internal(
        &self,
        requests: &[RuntimeResolverRequest],
    ) -> ResolverBatchResult {
        let total_started_at = Instant::now();
        let mut results = HashMap::new();
        let mut token_requests = Vec::new();
        let mut url_requests = Vec::new();

        let classify_started_at = Instant::now();
        for request in requests {
            match request {
                RuntimeResolverRequest::TokenMetadata { key, mint } => {
                    token_requests.push((key.clone(), mint.clone()));
                }
                RuntimeResolverRequest::UrlJson { key, url, method } => {
                    url_requests.push((key.clone(), url.clone(), method.clone()));
                }
            }
        }
        let classify_ms = elapsed_ms(classify_started_at);

        let token_request_count = token_requests.len();
        let url_request_count = url_requests.len();
        let mut token_batch_ms = 0.0_f64;
        let mut url_dedupe_ms = 0.0_f64;
        let mut url_batch_ms = 0.0_f64;
        let mut url_unique_request_count = 0usize;

        if !token_requests.is_empty() {
            let token_started_at = Instant::now();
            if let Some(token_client) = &self.token_client {
                let mints: Vec<String> = token_requests
                    .iter()
                    .map(|(_, mint)| mint.clone())
                    .collect();
                match token_client.resolve_token_metadata(&mints).await {
                    Ok(resolved) => {
                        for (key, mint) in token_requests {
                            if let Some(value) = resolved.get(&mint) {
                                results.insert(key, value.clone());
                            }
                        }
                    }
                    Err(err) => {
                        tracing::warn!(error = %err, "Failed to resolve token metadata batch");
                    }
                }
            } else {
                tracing::warn!(
                    count = token_requests.len(),
                    "DAS_API_ENDPOINT not set; token resolver requests will be re-queued"
                );
            }
            token_batch_ms = elapsed_ms(token_started_at);
        }

        if !url_requests.is_empty() {
            let url_dedupe_started_at = Instant::now();
            let mut unique = HashMap::new();
            for (key, url, method) in &url_requests {
                unique
                    .entry((url.clone(), method.clone()))
                    .or_insert_with(Vec::new)
                    .push(key.clone());
            }
            url_unique_request_count = unique.len();

            let batch_input: Vec<(String, HttpMethod)> = unique.keys().cloned().collect();
            url_dedupe_ms = elapsed_ms(url_dedupe_started_at);
            let url_batch_started_at = Instant::now();
            let resolved = self.url_client.resolve_batch(&batch_input).await;
            url_batch_ms = elapsed_ms(url_batch_started_at);

            for ((url, method), keys) in unique {
                if let Some(value) = resolved.get(&(url, method)) {
                    for key in keys {
                        results.insert(key, value.clone());
                    }
                }
            }
        }

        let total_ms = elapsed_ms(total_started_at);
        if should_log_runtime_resolver_profile(total_ms) {
            tracing::info!(
                stage = "runtime.resolver_batch",
                request_count = requests.len(),
                token_request_count,
                url_request_count,
                url_unique_request_count,
                resolved_count = results.len(),
                total_ms,
                classify_ms,
                token_batch_ms,
                url_dedupe_ms,
                url_batch_ms,
                "Runtime resolver batch profile"
            );
        }

        Ok(results)
    }
}

impl RuntimeResolver for InProcessResolver {
    fn resolve_batch<'a>(
        &'a self,
        requests: &'a [RuntimeResolverRequest],
    ) -> ResolverBatchFuture<'a> {
        Box::pin(async move { self.resolve_batch_internal(requests).await })
    }
}

pub fn runtime_resolver_cache_key(resolver: &ResolverType, input: &Value) -> String {
    crate::vm::resolver_cache_key(resolver, input)
}

fn runtime_request_from_vm_request(request: &ResolverRequest) -> Option<RuntimeResolverRequest> {
    match &request.resolver {
        ResolverType::Token => extract_mint_from_input(&request.input).map(|mint| {
            RuntimeResolverRequest::TokenMetadata {
                key: request.cache_key.clone(),
                mint,
            }
        }),
        ResolverType::Url(config) => match &request.input {
            Value::String(url) if !url.is_empty() => Some(RuntimeResolverRequest::UrlJson {
                key: request.cache_key.clone(),
                url: url.clone(),
                method: config.method.clone(),
            }),
            _ => None,
        },
    }
}

fn extract_mint_from_input(input: &Value) -> Option<String> {
    match input {
        Value::String(value) if !value.is_empty() => Some(value.clone()),
        Value::Object(map) => map
            .get("mint")
            .and_then(|value| value.as_str())
            .filter(|value| !value.is_empty())
            .map(str::to_string),
        _ => None,
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn token_request_extracts_mint_from_object_input() {
        let request = ResolverRequest {
            cache_key: "token:mint".to_string(),
            resolver: ResolverType::Token,
            input: serde_json::json!({ "mint": "abc" }),
        };

        let runtime_request = runtime_request_from_vm_request(&request).unwrap();
        assert_eq!(
            runtime_request,
            RuntimeResolverRequest::TokenMetadata {
                key: "token:mint".to_string(),
                mint: "abc".to_string(),
            }
        );
    }

    #[test]
    fn url_request_uses_existing_cache_key() {
        let request = ResolverRequest {
            cache_key: "url:get:https://example.com".to_string(),
            resolver: ResolverType::Url(crate::ast::UrlResolverConfig {
                url_source: crate::ast::UrlSource::FieldPath("metadata_url".to_string()),
                method: HttpMethod::Get,
                extract_path: None,
            }),
            input: serde_json::json!("https://example.com"),
        };

        let runtime_request = runtime_request_from_vm_request(&request).unwrap();
        assert_eq!(
            runtime_request,
            RuntimeResolverRequest::UrlJson {
                key: "url:get:https://example.com".to_string(),
                url: "https://example.com".to_string(),
                method: HttpMethod::Get,
            }
        );
    }
}