vectoramp 0.1.0

Official Rust client for the VectorAmp public API.
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
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
//! Core data types returned by and sent to the VectorAmp API.

use std::collections::HashMap;

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

/// Free-form key/value metadata attached to resources and results.
pub type Metadata = HashMap<String, Value>;

/// Default embedding provider used when a dataset is created without one.
pub const DEFAULT_EMBEDDING_PROVIDER: &str = "vectoramp";
/// Default embedding model used when a dataset is created without one.
pub const DEFAULT_EMBEDDING_MODEL: &str = "VectorAmp-Embedding-4B";
/// Inferred dimensionality of [`DEFAULT_EMBEDDING_MODEL`].
pub const DEFAULT_EMBEDDING_DIM: u32 = 2560;

/// Embedding provider/model selector attached to a dataset.
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct EmbeddingConfig {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub provider: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub model: Option<String>,
}

impl EmbeddingConfig {
    /// VectorAmp default embedding (`VectorAmp-Embedding-4B`, dim `2560`).
    pub fn vectoramp() -> Self {
        Self {
            provider: Some(DEFAULT_EMBEDDING_PROVIDER.into()),
            model: Some(DEFAULT_EMBEDDING_MODEL.into()),
        }
    }

    /// OpenAI embedding helper. Accepts the shorthands `"small"` /`"large"` or
    /// a full model id such as `text-embedding-3-small`.
    pub fn openai<S: Into<String>>(model: S) -> Self {
        let model = match model.into().as_str() {
            "small" => "text-embedding-3-small".to_string(),
            "large" => "text-embedding-3-large".to_string(),
            other => other.to_string(),
        };
        Self {
            provider: Some("openai".into()),
            model: Some(model),
        }
    }
}

/// Infer the embedding dimensionality from a `provider`/`model` pair using the
/// built-in table. Returns `None` for custom/unknown models, in which case the
/// caller must supply `dim` explicitly.
///
/// Known: `vectoramp/VectorAmp-Embedding-4B` → 2560,
/// `openai/text-embedding-3-small` → 1536, `openai/text-embedding-3-large` →
/// 3072.
pub fn infer_embedding_dim(provider: Option<&str>, model: Option<&str>) -> Option<u32> {
    match (provider, model) {
        (_, Some(DEFAULT_EMBEDDING_MODEL)) => Some(DEFAULT_EMBEDDING_DIM),
        (Some("openai"), Some("text-embedding-3-small")) => Some(1536),
        (Some("openai"), Some("text-embedding-3-large")) => Some(3072),
        // Allow inference even when the provider is omitted but the model is a
        // well-known OpenAI model.
        (_, Some("text-embedding-3-small")) => Some(1536),
        (_, Some("text-embedding-3-large")) => Some(3072),
        _ => None,
    }
}

/// VectorAmp dataset returned by the API.
///
/// Public dataset creation is SABLE-only, so newly created datasets always
/// come back with `index_type: "sable"`.
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct DatasetInfo {
    pub id: String,
    pub name: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub org_id: Option<String>,
    pub dim: u32,
    pub metric: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub tuning: Option<HashMap<String, Value>>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub embedding: Option<EmbeddingConfig>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub index_type: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub created_at: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub updated_at: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

/// Paginated list of datasets.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct DatasetList {
    #[serde(default)]
    pub datasets: Vec<DatasetInfo>,
    #[serde(default)]
    pub total: u32,
    #[serde(default)]
    pub limit: u32,
    #[serde(default)]
    pub offset: u32,
}

/// Request body for [`DatasetService::create`](crate::DatasetService::create).
///
/// Only `name` is required. When `dim` is omitted the SDK infers it from the
/// embedding model (defaulting to `VectorAmp-Embedding-4B` → `2560`). The SDK
/// always sends `index_type: "sable"` since public dataset creation is
/// SABLE-only and never exposes the index type.
///
/// Prefer the [`DatasetService::create`](crate::DatasetService::create) helper
/// (`create("name")`) or [`CreateDatasetRequest::builder`] over filling this
/// struct manually.
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct CreateDatasetRequest {
    pub name: String,
    /// Embedding dimensionality. Optional: inferred from the embedding model
    /// when omitted. Serialized as `dim` (never `dimension`).
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub dim: Option<u32>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metric: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub tuning: Option<HashMap<String, Value>>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub embedding: Option<EmbeddingConfig>,
    /// Enable hybrid (dense + sparse) indexing. Maps to `hybrid: true`.
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub hybrid: Option<bool>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

impl CreateDatasetRequest {
    /// Start a builder for a dataset with the given name. Everything else is
    /// optional and defaulted server-side or by the SDK.
    pub fn builder<S: Into<String>>(name: S) -> CreateDatasetBuilder {
        CreateDatasetBuilder::new(name)
    }
}

/// Fluent builder for [`CreateDatasetRequest`].
///
/// Only the dataset name is required. The default embedding is
/// `VectorAmp-Embedding-4B` (provider `vectoramp`), dim is inferred (`2560`),
/// and metric defaults to `cosine`.
#[derive(Debug, Clone)]
pub struct CreateDatasetBuilder {
    request: CreateDatasetRequest,
}

impl CreateDatasetBuilder {
    /// Begin building a create request for `name`.
    pub fn new<S: Into<String>>(name: S) -> Self {
        Self {
            request: CreateDatasetRequest {
                name: name.into(),
                ..Default::default()
            },
        }
    }

    /// Set the embedding dimensionality explicitly. Required only for
    /// custom/unknown embedding models the SDK cannot infer.
    pub fn dim(mut self, dim: u32) -> Self {
        self.request.dim = Some(dim);
        self
    }

    /// Set the distance metric (defaults to `cosine`).
    pub fn metric<S: Into<String>>(mut self, metric: S) -> Self {
        self.request.metric = Some(metric.into());
        self
    }

    /// Set the embedding provider/model.
    pub fn embedding(mut self, embedding: EmbeddingConfig) -> Self {
        self.request.embedding = Some(embedding);
        self
    }

    /// Use an OpenAI embedding model. Accepts `"small"`/`"large"` (or the full
    /// model id). Dimensionality is inferred (`1536`/`3072`).
    pub fn openai<S: Into<String>>(mut self, model: S) -> Self {
        self.request.embedding = Some(EmbeddingConfig::openai(model));
        self
    }

    /// Enable hybrid (dense + sparse) indexing.
    pub fn hybrid(mut self, hybrid: bool) -> Self {
        self.request.hybrid = Some(hybrid);
        self
    }

    /// Attach arbitrary metadata to the dataset.
    pub fn metadata(mut self, metadata: Metadata) -> Self {
        self.request.metadata = Some(metadata);
        self
    }

    /// Finish building. The SDK fills in defaults and dim inference at create
    /// time.
    pub fn build(self) -> CreateDatasetRequest {
        self.request
    }
}

impl From<CreateDatasetBuilder> for CreateDatasetRequest {
    fn from(b: CreateDatasetBuilder) -> Self {
        b.build()
    }
}

impl From<&str> for CreateDatasetRequest {
    fn from(name: &str) -> Self {
        CreateDatasetRequest {
            name: name.to_owned(),
            ..Default::default()
        }
    }
}

impl From<String> for CreateDatasetRequest {
    fn from(name: String) -> Self {
        CreateDatasetRequest {
            name,
            ..Default::default()
        }
    }
}

/// Identifier for a vector record.
///
/// A vector id may be a string **or** an integer. Integer ids are serialized as
/// JSON numbers (not strings) so the API preserves their numeric type and does
/// not rewrite them.
#[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
#[serde(untagged)]
pub enum VectorId {
    /// Integer id, serialized as a JSON number.
    Int(i64),
    /// String id, serialized as a JSON string.
    Str(String),
}

impl Default for VectorId {
    fn default() -> Self {
        VectorId::Str(String::new())
    }
}

impl std::fmt::Display for VectorId {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            VectorId::Int(n) => write!(f, "{n}"),
            VectorId::Str(s) => write!(f, "{s}"),
        }
    }
}

impl From<String> for VectorId {
    fn from(s: String) -> Self {
        VectorId::Str(s)
    }
}

impl From<&str> for VectorId {
    fn from(s: &str) -> Self {
        VectorId::Str(s.to_owned())
    }
}

macro_rules! vector_id_from_int {
    ($($t:ty),*) => {$(
        impl From<$t> for VectorId {
            fn from(n: $t) -> Self {
                VectorId::Int(n as i64)
            }
        }
    )*};
}
vector_id_from_int!(i8, i16, i32, i64, u8, u16, u32, u64, usize, isize);

/// One vector record to insert into a dataset.
///
/// The `id` accepts a string or integer via [`VectorId`]; integer ids are
/// preserved as JSON numbers on the wire.
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct Vector {
    pub id: VectorId,
    pub values: Vec<f64>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

impl Vector {
    /// Construct a vector from an id (string or integer) and values.
    pub fn new<I: Into<VectorId>>(id: I, values: Vec<f64>) -> Self {
        Self {
            id: id.into(),
            values,
            metadata: None,
        }
    }

    /// Attach metadata to the vector.
    pub fn with_metadata(mut self, metadata: Metadata) -> Self {
        self.metadata = Some(metadata);
        self
    }
}

#[derive(Debug, Default, Clone, Serialize)]
pub(crate) struct InsertVectorsRequest {
    pub vectors: Vec<Vector>,
}

/// Number of vectors written by [`DatasetService::insert`](crate::DatasetService::insert).
#[derive(Debug, Default, Clone, Deserialize)]
pub struct InsertVectorsResponse {
    #[serde(default)]
    pub inserted: u32,
}

/// One text document to be embedded and inserted by `add_texts`.
///
/// `id` accepts a string or integer via [`VectorId`]. When omitted in the
/// convenience helpers the SDK generates `text-1`, `text-2`, … ids.
#[derive(Debug, Default, Clone)]
pub struct TextDocument {
    pub id: VectorId,
    pub text: String,
    pub metadata: Option<Metadata>,
}

/// Counts returned by `add_texts`.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct AddTextsResponse {
    #[serde(default)]
    pub inserted: u32,
    #[serde(default)]
    pub embeddings: u32,
}

/// Embedding endpoint request body.
#[derive(Debug, Default, Clone, Serialize)]
pub struct EmbedRequest {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub text: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub texts: Option<Vec<String>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub embedding_provider: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub embedding_model: Option<String>,
}

/// Embedding endpoint response.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct EmbedResponse {
    #[serde(default)]
    pub embeddings: Vec<Vec<f64>>,
    #[serde(default)]
    pub embedding: Option<Vec<f64>>,
}

/// Dataset search request body.
///
/// Provide either [`SearchRequest::query`] for vector search or
/// [`SearchRequest::query_text`] for text search. `top_k` defaults to 10 when
/// the SDK's `search` convenience helpers are used.
/// VectorAmp rerank settings. Only `enabled` is required; provider defaults to
/// `vectoramp` and model defaults to `VectorAmp-Rerank-v1`.
#[derive(Debug, Clone, Serialize)]
#[serde(untagged)]
pub enum Rerank {
    Enabled(bool),
    Config(RerankConfig),
}

#[derive(Debug, Default, Clone, Serialize)]
pub struct RerankConfig {
    pub enabled: bool,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub provider: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub model: Option<String>,
}

#[derive(Debug, Default, Clone, Serialize)]
pub struct SearchRequest {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub query: Option<Vec<f64>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub query_text: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub embedding_model: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub embedding_provider: Option<String>,
    pub top_k: u32,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub filters: Option<HashMap<String, String>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub advanced_filters: Option<Vec<AdvancedFilter>>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub nprobe_override: Option<u32>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub rerank_depth_override: Option<u32>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub hybrid: Option<bool>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub sparse_query: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub alpha: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub include_embeddings: Option<bool>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub include_documents: Option<bool>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub include_metadata: Option<bool>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub rerank: Option<Rerank>,
}

/// Structured metadata filter expression.
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct AdvancedFilter {
    pub field: String,
    pub op: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub value: Option<Value>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub values: Option<Vec<Value>>,
}

/// One ranked search result.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SearchResult {
    pub id: Value,
    #[serde(default)]
    pub score: f64,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub embedding: Option<Vec<f64>>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub doc_kind: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub doc_value: Option<String>,
}

/// Dataset search response.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SearchResponse {
    #[serde(default)]
    pub results: Vec<SearchResult>,
    #[serde(default)]
    pub dataset_id: Option<String>,
    #[serde(default)]
    pub query_time_ms: Option<f64>,
}

/// Ingestion source returned by the API.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct Source {
    #[serde(default)]
    pub id: Option<String>,
    #[serde(default)]
    pub source_id: Option<String>,
    #[serde(default)]
    pub uuid: Option<String>,
    #[serde(default)]
    pub name: String,
    #[serde(default, rename = "type")]
    pub kind: Option<String>,
    #[serde(default)]
    pub source_type: Option<String>,
    #[serde(default)]
    pub description: Option<String>,
    #[serde(default)]
    pub config: Option<HashMap<String, Value>>,
    #[serde(default)]
    pub metadata: Option<Metadata>,
    #[serde(default)]
    pub created_at: Option<String>,
    #[serde(default)]
    pub updated_at: Option<String>,
}

impl Source {
    /// Return the most authoritative identifier set on the source.
    pub fn identifier(&self) -> Option<&str> {
        self.id
            .as_deref()
            .or(self.source_id.as_deref())
            .or(self.uuid.as_deref())
    }
}

/// Paginated list of ingestion sources.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SourceList {
    #[serde(default)]
    pub sources: Vec<Source>,
    #[serde(default)]
    pub total: u32,
    #[serde(default)]
    pub limit: u32,
    #[serde(default)]
    pub offset: u32,
}

/// Request body for creating an ingestion source. Prefer the typed builders in
/// [`crate::sources`] over filling this struct manually.
#[derive(Debug, Default, Clone, Serialize)]
pub struct CreateSourceRequest {
    pub source_type: String,
    pub name: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub description: Option<String>,
    pub config: HashMap<String, Value>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

/// Ingestion job returned by the API.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct Job {
    #[serde(default)]
    pub job_id: String,
    #[serde(default)]
    pub status: String,
    #[serde(default)]
    pub message: Option<String>,
    #[serde(default)]
    pub documents_processed: Option<u64>,
    #[serde(default)]
    pub vectors_inserted: Option<u64>,
    #[serde(default)]
    pub processing_time_seconds: Option<f64>,
    #[serde(default)]
    pub pipeline_result: Option<Value>,
    #[serde(default)]
    pub error_details: Option<Value>,
    #[serde(default)]
    pub started_at: Option<String>,
    #[serde(default)]
    pub completed_at: Option<String>,
    #[serde(default)]
    pub progress_percentage: Option<f64>,
    #[serde(default)]
    pub current_step: Option<String>,
}

/// Paginated list of ingestion jobs.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct JobList {
    #[serde(default)]
    pub jobs: Vec<Job>,
    #[serde(default)]
    pub total: u32,
    #[serde(default)]
    pub limit: u32,
    #[serde(default)]
    pub offset: u32,
}

/// Request body for [`IngestionService::start_job`](crate::IngestionService::start_job).
#[derive(Debug, Default, Clone, Serialize)]
pub struct StartIngestionRequest {
    pub source_id: String,
    pub dataset_id: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub pipeline_id: Option<String>,
}

/// Local file descriptor used when initializing a presigned upload.
#[derive(Debug, Default, Clone, Serialize)]
pub struct UploadFile {
    pub name: String,
    pub size_bytes: u64,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub content_type: Option<String>,
}

#[derive(Debug, Default, Clone, Serialize)]
pub(crate) struct InitUploadRequest {
    pub files: Vec<UploadFile>,
}

/// One presigned upload destination returned by the upload init endpoint.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct UploadTarget {
    pub file_id: String,
    #[serde(default)]
    pub file_name: Option<String>,
    pub upload_url: String,
}

/// Response from the upload init endpoint.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct InitUploadResponse {
    pub job_id: String,
    #[serde(default)]
    pub uploads: Vec<UploadTarget>,
}

#[derive(Debug, Default, Clone, Serialize)]
pub(crate) struct CompleteUploadRequest {
    pub job_id: String,
    pub file_ids: Vec<String>,
}

/// Conversation turn for [`AskRequest::conversation_history`].
#[derive(Debug, Default, Clone, Serialize, Deserialize)]
pub struct ConversationMessage {
    pub role: String,
    pub content: String,
}

/// Intelligence query request body.
///
/// `dataset_id` is optional. Pass a dataset id, `"all"`, or leave it unset for
/// the API default. Use the `Client::ask` and `Dataset::ask` helpers when you
/// want a single-call API.
#[derive(Debug, Default, Clone, Serialize)]
pub struct AskRequest {
    pub query: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub dataset_id: Option<Value>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub top_k: Option<u32>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub conversation_history: Option<Vec<ConversationMessage>>,
    pub stream: bool,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub include_sources: Option<bool>,
}

/// Non-streaming intelligence response.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct AskResponse {
    #[serde(default)]
    pub answer: String,
    #[serde(default)]
    pub sources: Vec<SourceCitation>,
    #[serde(default)]
    pub chunks: Vec<RagChunk>,
    #[serde(default)]
    pub message: Option<String>,
    #[serde(default)]
    pub metadata: Option<Metadata>,
}

/// Source cited by an intelligence answer.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SourceCitation {
    #[serde(default)]
    pub name: Option<String>,
    #[serde(default)]
    pub path: Option<String>,
    #[serde(default)]
    pub url: Option<String>,
    #[serde(default)]
    pub source_type: Option<String>,
    #[serde(default)]
    pub content_type: Option<String>,
    #[serde(default)]
    pub relevance: Option<f64>,
    #[serde(default)]
    pub pages: Vec<u32>,
    #[serde(default)]
    pub chunk_count: Option<u32>,
    #[serde(default)]
    pub preview: Option<String>,
    #[serde(default)]
    pub chunks: Vec<RagChunk>,
    #[serde(default)]
    pub file_id: Option<String>,
    #[serde(default)]
    pub thumbnail_url: Option<String>,
}

/// Retrieved chunk used by an intelligence answer.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct RagChunk {
    #[serde(default)]
    pub id: Option<String>,
    #[serde(default)]
    pub chunk_id: Option<String>,
    #[serde(default)]
    pub text: Option<String>,
    #[serde(default)]
    pub score: Option<f64>,
    #[serde(default)]
    pub source: Option<String>,
    #[serde(default)]
    pub source_url: Option<String>,
    #[serde(default)]
    pub page: Option<Value>,
    #[serde(default)]
    pub metadata: Option<Metadata>,
}

/// Body for creating an intelligence session.
#[derive(Debug, Default, Clone, Serialize)]
pub struct CreateSessionRequest {
    #[serde(skip_serializing_if = "Option::is_none")]
    pub title: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub workspace_id: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub dataset_id: Option<String>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

impl From<()> for CreateSessionRequest {
    fn from(_: ()) -> Self {
        CreateSessionRequest::default()
    }
}

impl From<&str> for CreateSessionRequest {
    fn from(title: &str) -> Self {
        CreateSessionRequest {
            title: Some(title.to_owned()),
            ..Default::default()
        }
    }
}

impl From<String> for CreateSessionRequest {
    fn from(title: String) -> Self {
        CreateSessionRequest {
            title: Some(title),
            ..Default::default()
        }
    }
}

/// An intelligence session: a durable container for a RAG conversation.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct IntelligenceSession {
    #[serde(default)]
    pub id: String,
    #[serde(default)]
    pub organization_id: Option<String>,
    #[serde(default)]
    pub user_id: Option<String>,
    #[serde(default)]
    pub workspace_id: Option<String>,
    #[serde(default)]
    pub dataset_id: Option<String>,
    #[serde(default)]
    pub title: Option<String>,
    #[serde(default)]
    pub status: Option<String>,
    #[serde(default)]
    pub metadata: Option<Metadata>,
    #[serde(default)]
    pub created_at: Option<String>,
    #[serde(default)]
    pub updated_at: Option<String>,
}

/// Paginated list of intelligence sessions.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SessionList {
    #[serde(default)]
    pub sessions: Vec<IntelligenceSession>,
}

/// Body for appending a message to an intelligence session.
#[derive(Debug, Default, Clone, Serialize)]
pub struct AppendMessageRequest {
    pub role: String,
    pub content: String,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

/// One message stored in an intelligence session.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct SessionMessage {
    #[serde(default)]
    pub id: String,
    #[serde(default)]
    pub session_id: Option<String>,
    #[serde(default)]
    pub role: String,
    #[serde(default)]
    pub content: String,
    #[serde(default)]
    pub metadata: Option<Metadata>,
    #[serde(default)]
    pub created_at: Option<String>,
}

/// Paginated list of session messages.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct MessageList {
    #[serde(default)]
    pub messages: Vec<SessionMessage>,
}

/// Cursor pagination options for dataset document listing.
#[derive(Debug, Default, Clone)]
pub struct DocumentListOpts {
    pub limit: Option<u32>,
    pub cursor: Option<String>,
    pub status: Option<String>,
}

/// Source/original document metadata returned by the dataset document catalog.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct DatasetDocument {
    pub id: String,
    #[serde(default)]
    pub dataset_id: Option<String>,
    #[serde(default)]
    pub source_id: Option<String>,
    #[serde(default)]
    pub source_type: Option<String>,
    #[serde(default)]
    pub external_id: Option<String>,
    #[serde(default)]
    pub file_name: Option<String>,
    #[serde(default)]
    pub mime_type: Option<String>,
    #[serde(default)]
    pub size_bytes: Option<i64>,
    #[serde(default)]
    pub content_hash: Option<String>,
    #[serde(default)]
    pub status: Option<String>,
    #[serde(default)]
    pub version: Option<i32>,
    #[serde(default)]
    pub chunk_count: Option<i32>,
    #[serde(default)]
    pub embeddings_count: Option<i32>,
    #[serde(default)]
    pub download_available: bool,
    #[serde(default)]
    pub created_at: Option<String>,
    #[serde(default)]
    pub updated_at: Option<String>,
}

/// Recurring ingestion schedule returned by the API.
///
/// A schedule pairs a source with a target dataset and a cron expression. The
/// server's ingestion scheduler daemon polls for due schedules and creates jobs
/// as they fire.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct Schedule {
    #[serde(default)]
    pub id: String,
    #[serde(default)]
    pub organization_id: Option<String>,
    #[serde(default)]
    pub source_id: Option<String>,
    #[serde(default)]
    pub dataset_id: Option<String>,
    #[serde(default)]
    pub pipeline_id: Option<String>,
    #[serde(default)]
    pub cron: Option<String>,
    #[serde(default)]
    pub timezone: Option<String>,
    #[serde(default)]
    pub enabled: bool,
    #[serde(default)]
    pub name: Option<String>,
    #[serde(default)]
    pub next_run_at: Option<String>,
    #[serde(default)]
    pub last_run_at: Option<String>,
    #[serde(default)]
    pub metadata: Option<Metadata>,
    #[serde(default)]
    pub created_at: Option<String>,
    #[serde(default)]
    pub updated_at: Option<String>,
}

/// Paginated list of schedules.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct ScheduleList {
    #[serde(default)]
    pub schedules: Vec<Schedule>,
    #[serde(default)]
    pub total: u32,
    #[serde(default)]
    pub limit: u32,
    #[serde(default)]
    pub offset: u32,
}

/// Body for creating an ingestion schedule.
#[derive(Debug, Default, Clone, Serialize)]
pub struct CreateScheduleRequest {
    pub source_id: String,
    pub dataset_id: String,
    pub cron: String,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub timezone: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub pipeline_id: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub enabled: Option<bool>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub name: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

/// Body for updating an ingestion schedule. Only `Some` fields are sent.
#[derive(Debug, Default, Clone, Serialize)]
pub struct UpdateScheduleRequest {
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub cron: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub timezone: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub pipeline_id: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub enabled: Option<bool>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub name: Option<String>,
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub metadata: Option<Metadata>,
}

/// Response from an immediate schedule trigger.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct TriggerScheduleResponse {
    #[serde(default)]
    pub job_id: Option<String>,
}

/// Cursor-paginated dataset document list.
#[derive(Debug, Default, Clone, Deserialize)]
pub struct DatasetDocumentList {
    #[serde(default)]
    pub documents: Vec<DatasetDocument>,
    #[serde(default)]
    pub next_cursor: Option<String>,
    #[serde(default)]
    pub limit: Option<u32>,
}