Grcov report - lowlevel.rs

1

use std::borrow::Borrow;

2

use std::collections::BTreeMap;

3

4

use arc_bytes::serde::Bytes;

5

use async_trait::async_trait;

6

7

use super::GroupedReductions;

8

use crate::connection::{

9

    AccessPolicy, HasSession, QueryKey, Range, RangeRef, SerializedQueryKey, Sort,

10

};

11

use crate::document::{

12

    CollectionDocument, CollectionHeader, Document, DocumentId, HasHeader, Header, OwnedDocument,

13

};

14

use crate::key::{self, ByteSource, Key, KeyEncoding};

15

use crate::schema::view::map::{

16

    CollectionMap, MappedDocuments, MappedSerializedValue, ViewMappings,

17

};

18

use crate::schema::view::{self};

19

use crate::schema::{self, CollectionName, MappedValue, Schematic, SerializedCollection, ViewName};

20

use crate::transaction::{OperationResult, Transaction};

21

use crate::Error;

22

23

/// The low-level interface to a database's [`schema::Schema`], giving access to

24

/// [`Collection`s](crate::schema::Collection) and

25

/// [`Views`s](crate::schema::View). This trait is not safe to use within async

26

/// contexts and will block the current thread. For async access, use

27

/// [`AsyncLowLevelConnection`].

28

///

29

/// This trait's methods are not designed for ergonomics. See

30

/// [`Connection`](super::Connection) for a higher-level interface.

31

pub trait LowLevelConnection: HasSchema + HasSession {

32

    /// Inserts a newly created document into the connected [`schema::Schema`]

33

    /// for the [`Collection`](schema::Collection) `C`. If `id` is `None` a unique id will be

34

    /// generated. If an id is provided and a document already exists with that

35

    /// id, a conflict error will be returned.

36

///

37

    /// This is the lower-level API. For better ergonomics, consider using

38

    /// one of:

39

///

40

    /// - [`SerializedCollection::push_into()`]

41

    /// - [`SerializedCollection::insert_into()`]

42

    /// - [`self.collection::<Collection>().insert()`](super::Collection::insert)

43

    /// - [`self.collection::<Collection>().push()`](super::Collection::push)

44

38772

    fn insert<C, PrimaryKey, B>(

45

38772

        &self,

46

38772

        id: Option<&PrimaryKey>,

47

38772

        contents: B,

48

38772

    ) -> Result<CollectionHeader<C::PrimaryKey>, Error>

49

38772

    where

50

38772

        C: schema::Collection,

51

38772

        B: Into<Bytes> + Send,

52

38772

        PrimaryKey: KeyEncoding<C::PrimaryKey> + Send + ?Sized,

53

38772

54

38772

        let contents = contents.into();

55

38772

        let results = self.apply_transaction(Transaction::insert(

56

38772

            C::collection_name(),

57

38772

            id.map(|id| DocumentId::new(id)).transpose()?,

58

38772

            contents,

59

54

        ))?;

60

38718

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

61

38718

            CollectionHeader::try_from(header)

62

        } else {

63

            unreachable!(

64

                "apply_transaction on a single insert should yield a single DocumentUpdated entry"

65

66

67

38772

68

69

    /// Updates an existing document in the connected [`schema::Schema`] for the

70

    /// [`Collection`](schema::Collection) `C`. Upon success, `doc.revision` will be updated with

71

    /// the new revision.

72

///

73

    /// This is the lower-level API. For better ergonomics, consider using

74

    /// one of:

75

///

76

    /// - [`CollectionDocument::update()`]

77

    /// - [`self.collection::<Collection>().update()`](super::Collection::update)

78

916

    fn update<C: schema::Collection, D: Document<C> + Send + Sync>(

79

916

        &self,

80

916

        doc: &mut D,

81

916

    ) -> Result<(), Error> {

82

916

        let results = self.apply_transaction(Transaction::update(

83

916

            C::collection_name(),

84

916

            doc.header().into_header()?,

85

916

            doc.bytes()?,

86

12

        ))?;

87

904

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

88

904

            doc.set_header(header)?;

89

904

            Ok(())

90

        } else {

91

            unreachable!(

92

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

93

94

95

916

96

97

    /// Overwrites an existing document, or inserts a new document. Upon success,

98

    /// `doc.revision` will be updated with the new revision information.

99

///

100

    /// This is the lower-level API. For better ergonomics, consider using

101

    /// one of:

102

///

103

    /// - [`SerializedCollection::overwrite()`]

104

    /// - [`SerializedCollection::overwrite_into()`]

105

    /// - [`self.collection::<Collection>().overwrite()`](super::Collection::overwrite)

106

9

    fn overwrite<C, PrimaryKey>(

107

9

        &self,

108

9

        id: &PrimaryKey,

109

9

        contents: Vec<u8>,

110

9

    ) -> Result<CollectionHeader<C::PrimaryKey>, Error>

111

9

    where

112

9

        C: schema::Collection,

113

9

        PrimaryKey: KeyEncoding<C::PrimaryKey>,

114

9

115

9

        let results = self.apply_transaction(Transaction::overwrite(

116

9

            C::collection_name(),

117

9

            DocumentId::new(id)?,

118

9

            contents,

119

        ))?;

120

9

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

121

9

            CollectionHeader::try_from(header)

122

        } else {

123

            unreachable!(

124

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

125

126

127

9

128

129

    /// Retrieves a stored document from [`Collection`](schema::Collection) `C` identified by `id`.

130

///

131

    /// This is a lower-level API. For better ergonomics, consider using one of:

132

///

133

    /// - [`SerializedCollection::get()`]

134

    /// - [`self.collection::<Collection>().get()`](super::Collection::get)

135

2335

    fn get<C, PrimaryKey>(&self, id: &PrimaryKey) -> Result<Option<OwnedDocument>, Error>

136

2335

    where

137

2335

        C: schema::Collection,

138

2335

        PrimaryKey: KeyEncoding<C::PrimaryKey> + ?Sized,

139

2335

140

2335

        self.get_from_collection(DocumentId::new(id)?, &C::collection_name())

141

2335

142

143

    /// Retrieves all documents matching `ids`. Documents that are not found are

144

    /// not returned, but no error will be generated.

145

///

146

    /// This is a lower-level API. For better ergonomics, consider using one of:

147

///

148

    /// - [`SerializedCollection::get_multiple()`]

149

    /// - [`self.collection::<Collection>().get_multiple()`](super::Collection::get_multiple)

150

1378

    fn get_multiple<'id, C, PrimaryKey, DocumentIds, I>(

151

1378

        &self,

152

1378

        ids: DocumentIds,

153

1378

    ) -> Result<Vec<OwnedDocument>, Error>

154

1378

    where

155

1378

        C: schema::Collection,

156

1378

        DocumentIds: IntoIterator<Item = &'id PrimaryKey, IntoIter = I> + Send + Sync,

157

1378

        I: Iterator<Item = &'id PrimaryKey> + Send + Sync,

158

1378

        PrimaryKey: KeyEncoding<C::PrimaryKey> + 'id + ?Sized,

159

1378

160

1378

        let ids = ids

161

1378

            .into_iter()

162

1390

            .map(|id| DocumentId::new(id))

163

1378

            .collect::<Result<Vec<_>, _>>()?;

164

1378

        self.get_multiple_from_collection(&ids, &C::collection_name())

165

1378

166

167

    /// Retrieves all documents within the range of `ids`. To retrieve all

168

    /// documents, pass in `..` for `ids`.

169

///

170

    /// This is a lower-level API. For better ergonomics, consider using one of:

171

///

172

    /// - [`SerializedCollection::all()`]

173

    /// - [`self.collection::<Collection>().all()`](super::Collection::all)

174

    /// - [`SerializedCollection::list()`]

175

    /// - [`self.collection::<Collection>().list()`](super::Collection::list)

176

24

    fn list<'id, C, R, PrimaryKey>(

177

24

        &self,

178

24

        ids: R,

179

24

        order: Sort,

180

24

        limit: Option<u32>,

181

24

    ) -> Result<Vec<OwnedDocument>, Error>

182

24

    where

183

24

        C: schema::Collection,

184

24

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

185

24

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

186

24

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

187

24

188

26

        let ids = ids.into().map_result(|id| DocumentId::new(id))?;

189

24

        self.list_from_collection(ids, order, limit, &C::collection_name())

190

24

191

192

    /// Retrieves all documents within the range of `ids`. To retrieve all

193

    /// documents, pass in `..` for `ids`.

194

///

195

    /// This is the lower-level API. For better ergonomics, consider using one

196

    /// of:

197

///

198

    /// - [`SerializedCollection::all_async().headers()`](schema::List::headers)

199

    /// - [`self.collection::<Collection>().all().headers()`](super::List::headers)

200

    /// - [`SerializedCollection::list_async().headers()`](schema::List::headers)

201

    /// - [`self.collection::<Collection>().list().headers()`](super::List::headers)

202

3

    fn list_headers<'id, C, R, PrimaryKey>(

203

3

        &self,

204

3

        ids: R,

205

3

        order: Sort,

206

3

        limit: Option<u32>,

207

3

    ) -> Result<Vec<Header>, Error>

208

3

    where

209

3

        C: schema::Collection,

210

3

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

211

3

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

212

3

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

213

3

214

6

        let ids = ids.into().map_result(|id| DocumentId::new(id))?;

215

3

        self.list_headers_from_collection(ids, order, limit, &C::collection_name())

216

3

217

218

    /// Counts the number of documents within the range of `ids`.

219

///

220

    /// This is a lower-level API. For better ergonomics, consider using one of:

221

///

222

    /// - [`SerializedCollection::all().count()`](schema::List::count)

223

    /// - [`self.collection::<Collection>().all().count()`](super::List::count)

224

    /// - [`SerializedCollection::list().count()`](schema::List::count)

225

    /// - [`self.collection::<Collection>().list().count()`](super::List::count)

226

6

    fn count<'id, C, R, PrimaryKey>(&self, ids: R) -> Result<u64, Error>

227

6

    where

228

6

        C: schema::Collection,

229

6

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

230

6

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

231

6

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

232

6

233

6

        self.count_from_collection(

234

6

            ids.into().map_result(|id| DocumentId::new(id))?,

235

6

            &C::collection_name(),

236

237

6

238

239

    /// Removes a `Document` from the database.

240

///

241

    /// This is a lower-level API. For better ergonomics, consider using

242

    /// one of:

243

///

244

    /// - [`CollectionDocument::delete()`]

245

    /// - [`self.collection::<Collection>().delete()`](super::Collection::delete)

246

18669

    fn delete<C: schema::Collection, H: HasHeader + Send + Sync>(

247

18669

        &self,

248

18669

        doc: &H,

249

18669

    ) -> Result<(), Error> {

250

18669

        let results =

251

18669

            self.apply_transaction(Transaction::delete(C::collection_name(), doc.header()?))?;

252

18669

        if let OperationResult::DocumentDeleted { .. } = &results[0] {

253

18669

            Ok(())

254

        } else {

255

            unreachable!(

256

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

257

258

259

18669

260

261

    /// Queries for view entries matching [`View`](schema::View).

262

///

263

    /// This is a lower-level API. For better ergonomics, consider querying the

264

    /// view using [`View::entries(self).query()`](super::View::query) instead. The

265

    /// parameters for the query can be customized on the builder returned from

266

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

267

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

268

    /// or [`Connection::view()`](super::Connection::view).

269

25039

    fn query<V: schema::SerializedView, Key>(

270

25039

        &self,

271

25039

        key: Option<QueryKey<'_, V::Key, Key>>,

272

25039

        order: Sort,

273

25039

        limit: Option<u32>,

274

25039

        access_policy: AccessPolicy,

275

25039

    ) -> Result<ViewMappings<V>, Error>

276

25039

    where

277

25039

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

278

25039

        V::Key: Borrow<Key> + PartialEq<Key>,

279

25039

280

25039

        let view = self.schematic().view::<V>()?;

281

25039

        let mappings = self.query_by_name(

282

25039

            &view.view_name(),

283

25039

            key.map(|key| key.serialized()).transpose()?,

284

25039

            order,

285

25039

            limit,

286

25039

            access_policy,

287

)?;

288

25039

        mappings

289

25039

            .into_iter()

290

25039

            .map(|mapping| {

291

21286

                Ok(CollectionMap {

292

21286

                    key: <V::Key as key::Key>::from_ord_bytes(ByteSource::Borrowed(&mapping.key))

293

21286

                        .map_err(view::Error::key_serialization)

294

21286

                        .map_err(Error::from)?,

295

21286

                    value: V::deserialize(&mapping.value)?,

296

21286

                    source: mapping.source.try_into()?,

297

})

298

25039

})

299

25039

            .collect::<Result<Vec<_>, Error>>()

300

25039

301

302

    /// Queries for view entries matching [`View`](schema::View) with their

303

    /// source documents.

304

///

305

    /// This is a lower-level API. For better ergonomics, consider querying the

306

    /// view using

307

    /// [`View::entries(self).query_with_docs()`](super::View::query_with_docs)

308

    /// instead. The parameters for the query can be customized on the builder

309

    /// returned from

310

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

311

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

312

    /// or [`Connection::view()`](super::Connection::view).

313

716

    fn query_with_docs<V: schema::SerializedView, Key>(

314

716

        &self,

315

716

        key: Option<QueryKey<'_, V::Key, Key>>,

316

716

        order: Sort,

317

716

        limit: Option<u32>,

318

716

        access_policy: AccessPolicy,

319

716

    ) -> Result<MappedDocuments<OwnedDocument, V>, Error>

320

716

    where

321

716

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

322

716

        V::Key: Borrow<Key> + PartialEq<Key>,

323

716

324

        // Query permission is checked by the query call

325

716

        let results = self.query::<V, Key>(key, order, limit, access_policy)?;

326

327

        // Verify that there is permission to fetch each document

328

716

        let documents = self

329

731

            .get_multiple::<V::Collection, _, _, _>(results.iter().map(|m| &m.source.id))?

330

716

            .into_iter()

331

731

            .map(|doc| {

332

729

                let id = doc.header.id.deserialize()?;

333

729

                Ok((id, doc))

334

731

})

335

716

            .collect::<Result<BTreeMap<_, _>, Error>>()?;

336

337

716

        Ok(MappedDocuments {

338

716

            mappings: results,

339

716

            documents,

340

716

})

341

716

342

343

    /// Queries for view entries matching [`View`](schema::View) with their

344

    /// source documents, deserialized.

345

///

346

    /// This is a lower-level API. For better ergonomics, consider querying the

347

    /// view using

348

    /// [`View::entries(self).query_with_collection_docs()`](super::View::query_with_collection_docs)

349

    /// instead. The parameters for the query can be customized on the builder

350

    /// returned from

351

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

352

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

353

    /// or [`Connection::view()`](super::Connection::view).

354

25

    fn query_with_collection_docs<V, Key>(

355

25

        &self,

356

25

        key: Option<QueryKey<'_, V::Key, Key>>,

357

25

        order: Sort,

358

25

        limit: Option<u32>,

359

25

        access_policy: AccessPolicy,

360

25

    ) -> Result<MappedDocuments<CollectionDocument<V::Collection>, V>, Error>

361

25

    where

362

25

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

363

25

        V::Key: Borrow<Key> + PartialEq<Key>,

364

25

        V: schema::SerializedView,

365

25

        V::Collection: SerializedCollection,

366

25

        <V::Collection as SerializedCollection>::Contents: std::fmt::Debug,

367

25

368

25

        let mapped_docs = self.query_with_docs::<V, Key>(key, order, limit, access_policy)?;

369

25

        let mut collection_docs = BTreeMap::new();

370

66

        for (id, doc) in mapped_docs.documents {

371

41

            collection_docs.insert(id, CollectionDocument::<V::Collection>::try_from(&doc)?);

372

373

25

        Ok(MappedDocuments {

374

25

            mappings: mapped_docs.mappings,

375

25

            documents: collection_docs,

376

25

})

377

25

378

379

    /// Reduces the view entries matching [`View`](schema::View).

380

///

381

    /// This is a lower-level API. For better ergonomics, consider reducing the

382

    /// view using [`View::entries(self).reduce()`](super::View::reduce)

383

    /// instead. The parameters for the query can be customized on the builder

384

    /// returned from

385

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

386

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

387

    /// or [`Connection::view()`](super::Connection::view).

388

42

    fn reduce<V: schema::SerializedView, Key>(

389

42

        &self,

390

42

        key: Option<QueryKey<'_, V::Key, Key>>,

391

42

        access_policy: AccessPolicy,

392

42

    ) -> Result<V::Value, Error>

393

42

    where

394

42

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

395

42

        V::Key: Borrow<Key> + PartialEq<Key>,

396

42

397

42

        let view = self.schematic().view::<V>()?;

398

42

        self.reduce_by_name(

399

42

            &view.view_name(),

400

42

            key.map(|key| key.serialized()).transpose()?,

401

42

            access_policy,

402

42

403

42

        .and_then(|value| V::deserialize(&value))

404

42

405

406

    /// Reduces the view entries matching [`View`](schema::View), reducing the

407

    /// values by each unique key.

408

///

409

    /// This is a lower-level API. For better ergonomics, consider reducing the

410

    /// view using

411

    /// [`View::entries(self).reduce_grouped()`](super::View::reduce_grouped)

412

    /// instead. The parameters for the query can be customized on the builder

413

    /// returned from

414

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

415

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

416

    /// or [`Connection::view()`](super::Connection::view).

417

18

    fn reduce_grouped<V: schema::SerializedView, Key>(

418

18

        &self,

419

18

        key: Option<QueryKey<'_, V::Key, Key>>,

420

18

        access_policy: AccessPolicy,

421

18

    ) -> Result<GroupedReductions<V>, Error>

422

18

    where

423

18

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

424

18

        V::Key: Borrow<Key> + PartialEq<Key>,

425

18

426

18

        let view = self.schematic().view::<V>()?;

427

18

        self.reduce_grouped_by_name(

428

18

            &view.view_name(),

429

18

            key.map(|key| key.serialized()).transpose()?,

430

18

            access_policy,

431

)?

432

18

        .into_iter()

433

45

        .map(|map| {

434

45

            Ok(MappedValue::new(

435

45

                V::Key::from_ord_bytes(ByteSource::Borrowed(&map.key))

436

45

                    .map_err(view::Error::key_serialization)?,

437

45

                V::deserialize(&map.value)?,

438

))

439

45

})

440

18

        .collect::<Result<Vec<_>, Error>>()

441

18

442

443

    /// Deletes all of the documents associated with this view.

444

///

445

    /// This is a lower-level API. For better ergonomics, consider querying the

446

    /// view using

447

    /// [`View::entries(self).delete_docs()`](super::View::delete_docs())

448

    /// instead. The parameters for the query can be customized on the builder

449

    /// returned from

450

    /// [`SerializedView::entries()`](schema::SerializedView::entries),

451

    /// [`SerializedView::entries_async()`](schema::SerializedView::entries_async),

452

    /// or [`Connection::view()`](super::Connection::view).

453

9

    fn delete_docs<V: schema::SerializedView, Key>(

454

9

        &self,

455

9

        key: Option<QueryKey<'_, V::Key, Key>>,

456

9

        access_policy: AccessPolicy,

457

9

    ) -> Result<u64, Error>

458

9

    where

459

9

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

460

9

        V::Key: Borrow<Key> + PartialEq<Key>,

461

9

462

9

        let view = self.schematic().view::<V>()?;

463

9

        self.delete_docs_by_name(

464

9

            &view.view_name(),

465

9

            key.map(|key| key.serialized()).transpose()?,

466

9

            access_policy,

467

468

9

469

470

    /// Applies a [`Transaction`] to the [`schema::Schema`]. If any operation in the

471

    /// [`Transaction`] fails, none of the operations will be applied to the

472

    /// [`schema::Schema`].

473

    fn apply_transaction(&self, transaction: Transaction) -> Result<Vec<OperationResult>, Error>;

474

475

    /// Retrieves the document with `id` stored within the named `collection`.

476

///

477

    /// This is a lower-level API. For better ergonomics, consider using

478

    /// one of:

479

///

480

    /// - [`SerializedCollection::get()`]

481

    /// - [`self.collection::<Collection>().get()`](super::Collection::get)

482

    fn get_from_collection(

483

        &self,

484

        id: DocumentId,

485

        collection: &CollectionName,

486

    ) -> Result<Option<OwnedDocument>, Error>;

487

488

    /// Retrieves all documents matching `ids` from the named `collection`.

489

    /// Documents that are not found are not returned, but no error will be

490

    /// generated.

491

///

492

    /// This is a lower-level API. For better ergonomics, consider using one of:

493

///

494

    /// - [`SerializedCollection::get_multiple()`]

495

    /// - [`self.collection::<Collection>().get_multiple()`](super::Collection::get_multiple)

496

    fn get_multiple_from_collection(

497

        &self,

498

        ids: &[DocumentId],

499

        collection: &CollectionName,

500

    ) -> Result<Vec<OwnedDocument>, Error>;

501

502

    /// Retrieves all documents within the range of `ids` from the named

503

    /// `collection`. To retrieve all documents, pass in `..` for `ids`.

504

///

505

    /// This is a lower-level API. For better ergonomics, consider using one of:

506

///

507

    /// - [`SerializedCollection::all()`]

508

    /// - [`self.collection::<Collection>().all()`](super::Collection::all)

509

    /// - [`SerializedCollection::list()`]

510

    /// - [`self.collection::<Collection>().list()`](super::Collection::list)

511

    fn list_from_collection(

512

        &self,

513

        ids: Range<DocumentId>,

514

        order: Sort,

515

        limit: Option<u32>,

516

        collection: &CollectionName,

517

    ) -> Result<Vec<OwnedDocument>, Error>;

518

519

    /// Retrieves all headers within the range of `ids` from the named

520

    /// `collection`. To retrieve all documents, pass in `..` for `ids`.

521

///

522

    /// This is a lower-level API. For better ergonomics, consider using one of:

523

///

524

    /// - [`SerializedCollection::all().headers()`](schema::List::headers)

525

    /// - [`self.collection::<Collection>().all().headers()`](super::AsyncCollection::all)

526

    /// - [`SerializedCollection::list().headers()`](schema::List::headers)

527

    /// - [`self.collection::<Collection>().list()`](super::AsyncCollection::list)

528

    fn list_headers_from_collection(

529

        &self,

530

        ids: Range<DocumentId>,

531

        order: Sort,

532

        limit: Option<u32>,

533

        collection: &CollectionName,

534

    ) -> Result<Vec<Header>, Error>;

535

536

    /// Counts the number of documents within the range of `ids` from the named

537

    /// `collection`.

538

///

539

    /// This is a lower-level API. For better ergonomics, consider using one of:

540

///

541

    /// - [`SerializedCollection::all().count()`](schema::List::count)

542

    /// - [`self.collection::<Collection>().all().count()`](super::List::count)

543

    /// - [`SerializedCollection::list().count()`](schema::List::count)

544

    /// - [`self.collection::<Collection>().list().count()`](super::List::count)

545

    fn count_from_collection(

546

        &self,

547

        ids: Range<DocumentId>,

548

        collection: &CollectionName,

549

    ) -> Result<u64, Error>;

550

551

    /// Compacts the collection to reclaim unused disk space.

552

///

553

    /// This process is done by writing data to a new file and swapping the file

554

    /// once the process completes. This ensures that if a hardware failure,

555

    /// power outage, or crash occurs that the original collection data is left

556

    /// untouched.

557

///

558

    /// ## Errors

559

///

560

    /// * [`Error::CollectionNotFound`]: database `name` does not exist.

561

    /// * [`Error::Other`]: an error occurred while compacting the database.

562

    fn compact_collection_by_name(&self, collection: CollectionName) -> Result<(), Error>;

563

564

    /// Queries for view entries from the named `view`.

565

///

566

    /// This is a lower-level API. For better ergonomics, consider querying the

567

    /// view using [`View::entries(self).query()`](super::View::query) instead. The

568

    /// parameters for the query can be customized on the builder returned from

569

    /// [`Connection::view()`](super::Connection::view).

570

    fn query_by_name(

571

        &self,

572

        view: &ViewName,

573

        key: Option<SerializedQueryKey>,

574

        order: Sort,

575

        limit: Option<u32>,

576

        access_policy: AccessPolicy,

577

    ) -> Result<Vec<schema::view::map::Serialized>, Error>;

578

579

    /// Queries for view entries from the named `view` with their source

580

    /// documents.

581

///

582

    /// This is a lower-level API. For better ergonomics, consider querying the

583

    /// view using

584

    /// [`View::entries(self).query_with_docs()`](super::View::query_with_docs)

585

    /// instead. The parameters for the query can be customized on the builder

586

    /// returned from [`Connection::view()`](super::Connection::view).

587

    fn query_by_name_with_docs(

588

        &self,

589

        view: &ViewName,

590

        key: Option<SerializedQueryKey>,

591

        order: Sort,

592

        limit: Option<u32>,

593

        access_policy: AccessPolicy,

594

    ) -> Result<schema::view::map::MappedSerializedDocuments, Error>;

595

596

    /// Reduces the view entries from the named `view`.

597

///

598

    /// This is a lower-level API. For better ergonomics, consider reducing the

599

    /// view using [`View::entries(self).reduce()`](super::View::reduce)

600

    /// instead. The parameters for the query can be customized on the builder

601

    /// returned from [`Connection::view()`](super::Connection::view).

602

    fn reduce_by_name(

603

        &self,

604

        view: &ViewName,

605

        key: Option<SerializedQueryKey>,

606

        access_policy: AccessPolicy,

607

    ) -> Result<Vec<u8>, Error>;

608

609

    /// Reduces the view entries from the named `view`, reducing the values by each

610

    /// unique key.

611

///

612

    /// This is a lower-level API. For better ergonomics, consider reducing

613

    /// the view using

614

    /// [`View::entries(self).reduce_grouped()`](super::View::reduce_grouped) instead.

615

    /// The parameters for the query can be customized on the builder returned

616

    /// from [`Connection::view()`](super::Connection::view).

617

    fn reduce_grouped_by_name(

618

        &self,

619

        view: &ViewName,

620

        key: Option<SerializedQueryKey>,

621

        access_policy: AccessPolicy,

622

    ) -> Result<Vec<MappedSerializedValue>, Error>;

623

624

    /// Deletes all source documents for entries that match within the named

625

    /// `view`.

626

///

627

    /// This is a lower-level API. For better ergonomics, consider querying the

628

    /// view using

629

    /// [`View::entries(self).delete_docs()`](super::View::delete_docs())

630

    /// instead. The parameters for the query can be customized on the builder

631

    /// returned from [`Connection::view()`](super::Connection::view).

632

    fn delete_docs_by_name(

633

        &self,

634

        view: &ViewName,

635

        key: Option<SerializedQueryKey>,

636

        access_policy: AccessPolicy,

637

    ) -> Result<u64, Error>;

638

639

640

/// The low-level interface to a database's [`schema::Schema`], giving access to

641

/// [`Collection`s](crate::schema::Collection) and

642

/// [`Views`s](crate::schema::View). This trait is for use within async

643

/// contexts. For access outside of async contexts, use [`LowLevelConnection`].

644

///

645

/// This trait's methods are not designed for ergonomics. See

646

/// [`AsyncConnection`](super::AsyncConnection) for a higher-level interface.

647

#[async_trait]

648

pub trait AsyncLowLevelConnection: HasSchema + HasSession + Send + Sync {

649

    /// Inserts a newly created document into the connected [`schema::Schema`]

650

    /// for the [`Collection`](schema::Collection) `C`. If `id` is `None` a unique id will be

651

    /// generated. If an id is provided and a document already exists with that

652

    /// id, a conflict error will be returned.

653

///

654

    /// This is the lower-level API. For better ergonomics, consider using

655

    /// one of:

656

///

657

    /// - [`SerializedCollection::push_into_async()`]

658

    /// - [`SerializedCollection::insert_into_async()`]

659

    /// - [`self.collection::<Collection>().insert()`](super::AsyncCollection::insert)

660

    /// - [`self.collection::<Collection>().push()`](super::AsyncCollection::push)

661

10734

    async fn insert<C: schema::Collection, PrimaryKey: Send, B: Into<Bytes> + Send>(

662

10734

        &self,

663

10734

        id: Option<&PrimaryKey>,

664

10734

        contents: B,

665

10734

    ) -> Result<CollectionHeader<C::PrimaryKey>, Error>

666

10734

    where

667

10734

        PrimaryKey: KeyEncoding<C::PrimaryKey> + ?Sized,

668

10734

669

10734

        let contents = contents.into();

670

10734

        let results = self

671

10734

            .apply_transaction(Transaction::insert(

672

10734

                C::collection_name(),

673

10734

                id.map(|id| DocumentId::new(id)).transpose()?,

674

10734

                contents,

675

))

676

25978

            .await?;

677

10211

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

678

10211

            CollectionHeader::try_from(header)

679

        } else {

680

            unreachable!(

681

                "apply_transaction on a single insert should yield a single DocumentUpdated entry"

682

683

684

21468

685

686

    /// Updates an existing document in the connected [`schema::Schema`] for the

687

    /// [`Collection`](schema::Collection)(schema::Collection) `C`. Upon success, `doc.revision`

688

    /// will be updated with the new revision.

689

///

690

    /// This is the lower-level API. For better ergonomics, consider using one

691

    /// of:

692

///

693

    /// - [`CollectionDocument::update_async()`]

694

    /// - [`self.collection::<Collection>().update()`](super::AsyncCollection::update)

695

5818

    async fn update<C: schema::Collection, D: Document<C> + Send + Sync>(

696

5818

        &self,

697

5818

        doc: &mut D,

698

5818

    ) -> Result<(), Error> {

699

5818

        let results = self

700

5818

            .apply_transaction(Transaction::update(

701

5818

                C::collection_name(),

702

5818

                doc.header().into_header()?,

703

5818

                doc.bytes()?,

704

))

705

12313

            .await?;

706

5798

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

707

5798

            doc.set_header(header)?;

708

5798

            Ok(())

709

        } else {

710

            unreachable!(

711

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

712

713

714

11636

715

716

    /// Overwrites an existing document, or inserts a new document. Upon success,

717

    /// `doc.revision` will be updated with the new revision information.

718

///

719

    /// This is the lower-level API. For better ergonomics, consider using

720

    /// one of:

721

///

722

    /// - [`SerializedCollection::overwrite_async()`]

723

    /// - [`SerializedCollection::overwrite_into_async()`]

724

    /// - [`self.collection::<Collection>().overwrite()`](super::AsyncCollection::overwrite)

725

15

    async fn overwrite<'a, C, PrimaryKey>(

726

15

        &self,

727

15

        id: &PrimaryKey,

728

15

        contents: Vec<u8>,

729

15

    ) -> Result<CollectionHeader<C::PrimaryKey>, Error>

730

15

    where

731

15

        C: schema::Collection,

732

15

        PrimaryKey: KeyEncoding<C::PrimaryKey>,

733

15

734

15

        let results = self

735

15

            .apply_transaction(Transaction::overwrite(

736

15

                C::collection_name(),

737

15

                DocumentId::new(id)?,

738

15

                contents,

739

))

740

15

            .await?;

741

15

        if let Some(OperationResult::DocumentUpdated { header, .. }) = results.into_iter().next() {

742

15

            CollectionHeader::try_from(header)

743

        } else {

744

            unreachable!(

745

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

746

747

748

30

749

750

    /// Retrieves a stored document from [`Collection`](schema::Collection) `C` identified by `id`.

751

///

752

    /// This is the lower-level API. For better ergonomics, consider using

753

    /// one of:

754

///

755

    /// - [`SerializedCollection::get_async()`]

756

    /// - [`self.collection::<Collection>().get()`](super::AsyncCollection::get)

757

16049

    async fn get<C, PrimaryKey>(&self, id: &PrimaryKey) -> Result<Option<OwnedDocument>, Error>

758

16049

    where

759

16049

        C: schema::Collection,

760

16049

        PrimaryKey: KeyEncoding<C::PrimaryKey> + ?Sized,

761

16049

762

16049

        self.get_from_collection(DocumentId::new(id)?, &C::collection_name())

763

19777

            .await

764

32098

765

766

    /// Retrieves all documents matching `ids`. Documents that are not found

767

    /// are not returned, but no error will be generated.

768

///

769

    /// This is the lower-level API. For better ergonomics, consider using

770

    /// one of:

771

///

772

    /// - [`SerializedCollection::get_multiple_async()`]

773

    /// - [`self.collection::<Collection>().get_multiple()`](super::AsyncCollection::get_multiple)

774

9626

    async fn get_multiple<'id, C, PrimaryKey, DocumentIds, I>(

775

9626

        &self,

776

9626

        ids: DocumentIds,

777

9626

    ) -> Result<Vec<OwnedDocument>, Error>

778

9626

    where

779

9626

        C: schema::Collection,

780

9626

        DocumentIds: IntoIterator<Item = &'id PrimaryKey, IntoIter = I> + Send + Sync,

781

9626

        I: Iterator<Item = &'id PrimaryKey> + Send + Sync,

782

9626

        PrimaryKey: KeyEncoding<C::PrimaryKey> + 'id + ?Sized,

783

9626

784

9626

        let ids = ids

785

9626

            .into_iter()

786

9626

            .map(DocumentId::new)

787

9626

            .collect::<Result<Vec<_>, _>>()?;

788

9626

        self.get_multiple_from_collection(&ids, &C::collection_name())

789

8788

            .await

790

19252

791

792

    /// Retrieves all documents within the range of `ids`. To retrieve all

793

    /// documents, pass in `..` for `ids`.

794

///

795

    /// This is the lower-level API. For better ergonomics, consider using one

796

    /// of:

797

///

798

    /// - [`SerializedCollection::all_async()`]

799

    /// - [`self.collection::<Collection>().all()`](super::AsyncCollection::all)

800

    /// - [`SerializedCollection::list_async()`]

801

    /// - [`self.collection::<Collection>().list()`](super::AsyncCollection::list)

802

20

    async fn list<'id, C, R, PrimaryKey>(

803

20

        &self,

804

20

        ids: R,

805

20

        order: Sort,

806

20

        limit: Option<u32>,

807

20

    ) -> Result<Vec<OwnedDocument>, Error>

808

20

    where

809

20

        C: schema::Collection,

810

20

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

811

20

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

812

20

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

813

20

814

30

        let ids = ids.into().map_result(|id| DocumentId::new(id))?;

815

20

        self.list_from_collection(ids, order, limit, &C::collection_name())

816

20

            .await

817

40

818

819

    /// Retrieves all documents within the range of `ids`. To retrieve all

820

    /// documents, pass in `..` for `ids`.

821

///

822

    /// This is the lower-level API. For better ergonomics, consider using one

823

    /// of:

824

///

825

    /// - [`SerializedCollection::all_async().headers()`](schema::AsyncList::headers)

826

    /// - [`self.collection::<Collection>().all()`](super::AsyncList::headers)

827

    /// - [`SerializedCollection::list_async().headers()`](schema::AsyncList::headers)

828

    /// - [`self.collection::<Collection>().list().headers()`](super::AsyncList::headers)

829

5

    async fn list_headers<'id, C, R, PrimaryKey>(

830

5

        &self,

831

5

        ids: R,

832

5

        order: Sort,

833

5

        limit: Option<u32>,

834

5

    ) -> Result<Vec<Header>, Error>

835

5

    where

836

5

        C: schema::Collection,

837

5

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

838

5

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

839

5

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

840

5

841

10

        let ids = ids.into().map_result(|id| DocumentId::new(id))?;

842

5

        self.list_headers_from_collection(ids, order, limit, &C::collection_name())

843

5

            .await

844

10

845

846

    /// Counts the number of documents within the range of `ids`.

847

///

848

    /// This is the lower-level API. For better ergonomics, consider using

849

    /// one of:

850

///

851

    /// - [`SerializedCollection::all_async().count()`](schema::AsyncList::count)

852

    /// - [`self.collection::<Collection>().all().count()`](super::AsyncList::count)

853

    /// - [`SerializedCollection::list_async().count()`](schema::AsyncList::count)

854

    /// - [`self.collection::<Collection>().list().count()`](super::AsyncList::count)

855

10

    async fn count<'id, C, R, PrimaryKey>(&self, ids: R) -> Result<u64, Error>

856

10

    where

857

10

        C: schema::Collection,

858

10

        R: Into<RangeRef<'id, C::PrimaryKey, PrimaryKey>> + Send,

859

10

        PrimaryKey: KeyEncoding<C::PrimaryKey> + PartialEq + 'id + ?Sized,

860

10

        C::PrimaryKey: Borrow<PrimaryKey> + PartialEq<PrimaryKey>,

861

10

862

10

        self.count_from_collection(

863

10

            ids.into().map_result(|id| DocumentId::new(id))?,

864

10

            &C::collection_name(),

865

866

10

        .await

867

20

868

869

    /// Removes a `Document` from the database.

870

///

871

    /// This is the lower-level API. For better ergonomics, consider using

872

    /// one of:

873

///

874

    /// - [`CollectionDocument::delete_async()`]

875

    /// - [`self.collection::<Collection>().delete()`](super::AsyncCollection::delete)

876

1046

    async fn delete<C: schema::Collection, H: HasHeader + Send + Sync>(

877

1046

        &self,

878

1046

        doc: &H,

879

1046

    ) -> Result<(), Error> {

880

1046

        let results = self

881

1046

            .apply_transaction(Transaction::delete(C::collection_name(), doc.header()?))

882

4137

            .await?;

883

1044

        if let OperationResult::DocumentDeleted { .. } = &results[0] {

884

1044

            Ok(())

885

        } else {

886

            unreachable!(

887

                "apply_transaction on a single update should yield a single DocumentUpdated entry"

888

889

890

2092

891

    /// Queries for view entries matching [`View`](schema::View)(super::AsyncView).

892

///

893

    /// This is the lower-level API. For better ergonomics, consider querying

894

    /// the view using [`View::entries(self).query()`](super::AsyncView::query)

895

    /// instead. The parameters for the query can be customized on the builder

896

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

897

9924

    async fn query<V: schema::SerializedView, Key>(

898

9924

        &self,

899

9924

        key: Option<QueryKey<'_, V::Key, Key>>,

900

9924

        order: Sort,

901

9924

        limit: Option<u32>,

902

9924

        access_policy: AccessPolicy,

903

9924

    ) -> Result<ViewMappings<V>, Error>

904

9924

    where

905

9924

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

906

9924

        V::Key: Borrow<Key> + PartialEq<Key>,

907

9924

908

9924

        let view = self.schematic().view::<V>()?;

909

9924

        let mappings = self

910

9924

            .query_by_name(

911

9924

                &view.view_name(),

912

9924

                key.map(|key| key.serialized()).transpose()?,

913

9924

                order,

914

9924

                limit,

915

9924

                access_policy,

916

917

9143

            .await?;

918

9924

        mappings

919

9924

            .into_iter()

920

10035

            .map(|mapping| {

921

9880

                Ok(CollectionMap {

922

9880

                    key: <V::Key as key::Key>::from_ord_bytes(ByteSource::Borrowed(&mapping.key))

923

9880

                        .map_err(view::Error::key_serialization)

924

9880

                        .map_err(Error::from)?,

925

9880

                    value: V::deserialize(&mapping.value)?,

926

9880

                    source: mapping.source.try_into()?,

927

})

928

10035

})

929

9924

            .collect::<Result<Vec<_>, Error>>()

930

19848

931

932

    /// Queries for view entries matching [`View`](schema::View) with their source documents.

933

///

934

    /// This is the lower-level API. For better ergonomics, consider querying

935

    /// the view using [`View::entries(self).query_with_docs()`](super::AsyncView::query_with_docs) instead.

936

    /// The parameters for the query can be customized on the builder returned

937

    /// from [`AsyncConnection::view()`](super::AsyncConnection::view).

938

    #[must_use]

939

9598

    async fn query_with_docs<V: schema::SerializedView, Key>(

940

9598

        &self,

941

9598

        key: Option<QueryKey<'_, V::Key, Key>>,

942

9598

        order: Sort,

943

9598

        limit: Option<u32>,

944

9598

        access_policy: AccessPolicy,

945

9598

    ) -> Result<MappedDocuments<OwnedDocument, V>, Error>

946

9598

    where

947

9598

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

948

9598

        V::Key: Borrow<Key> + PartialEq<Key>,

949

9598

950

        // Query permission is checked by the query call

951

9598

        let results = self

952

9598

            .query::<V, Key>(key, order, limit, access_policy)

953

8820

            .await?;

954

955

        // Verify that there is permission to fetch each document

956

9598

        let documents = self

957

9600

            .get_multiple::<V::Collection, _, _, _>(results.iter().map(|m| &m.source.id))

958

8760

            .await?

959

9598

            .into_iter()

960

9600

            .map(|doc| {

961

9512

                let id = doc.header.id.deserialize()?;

962

9512

                Ok((id, doc))

963

9600

})

964

9598

            .collect::<Result<BTreeMap<_, _>, Error>>()?;

965

966

9598

        Ok(MappedDocuments {

967

9598

            mappings: results,

968

9598

            documents,

969

9598

})

970

19196

971

972

    /// Queries for view entries matching [`View`](schema::View) with their source documents,

973

    /// deserialized.

974

///

975

    /// This is the lower-level API. For better ergonomics, consider querying

976

    /// the view using

977

    /// [`View::entries(self).query_with_collection_docs()`](super::AsyncView::query_with_collection_docs)

978

    /// instead. The parameters for the query can be customized on the builder

979

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

980

    #[must_use]

981

117

    async fn query_with_collection_docs<V, Key>(

982

117

        &self,

983

117

        key: Option<QueryKey<'_, V::Key, Key>>,

984

117

        order: Sort,

985

117

        limit: Option<u32>,

986

117

        access_policy: AccessPolicy,

987

117

    ) -> Result<MappedDocuments<CollectionDocument<V::Collection>, V>, Error>

988

117

    where

989

117

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

990

117

        V::Key: Borrow<Key> + PartialEq<Key>,

991

117

        V: schema::SerializedView,

992

117

        V::Collection: SerializedCollection,

993

117

        <V::Collection as SerializedCollection>::Contents: std::fmt::Debug,

994

117

995

117

        let mapped_docs = self

996

117

            .query_with_docs::<V, Key>(key, order, limit, access_policy)

997

232

            .await?;

998

117

        let mut collection_docs = BTreeMap::new();

999

236

        for (id, doc) in mapped_docs.documents {

1000

119

            collection_docs.insert(id, CollectionDocument::<V::Collection>::try_from(&doc)?);

1001

1002

117

        Ok(MappedDocuments {

1003

117

            mappings: mapped_docs.mappings,

1004

117

            documents: collection_docs,

1005

117

})

1006

234

1007

1008

    /// Reduces the view entries matching [`View`](schema::View).

1009

///

1010

    /// This is the lower-level API. For better ergonomics, consider querying

1011

    /// the view using

1012

    /// [`View::entries(self).reduce()`](super::AsyncView::reduce)

1013

    /// instead. The parameters for the query can be customized on the builder

1014

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1015

    #[must_use]

1016

18526

    async fn reduce<V: schema::SerializedView, Key>(

1017

18526

        &self,

1018

18526

        key: Option<QueryKey<'_, V::Key, Key>>,

1019

18526

        access_policy: AccessPolicy,

1020

18526

    ) -> Result<V::Value, Error>

1021

18526

    where

1022

18526

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

1023

18526

        V::Key: Borrow<Key> + PartialEq<Key>,

1024

18526

1025

18526

        let view = self.schematic().view::<V>()?;

1026

18526

        self.reduce_by_name(

1027

18526

            &view.view_name(),

1028

18526

            key.map(|key| key.serialized()).transpose()?,

1029

18526

            access_policy,

1030

1031

16982

        .await

1032

18526

        .and_then(|value| V::deserialize(&value))

1033

37052

1034

1035

    /// Reduces the view entries matching [`View`](schema::View), reducing the values by each

1036

    /// unique key.

1037

///

1038

    /// This is the lower-level API. For better ergonomics, consider querying

1039

    /// the view using

1040

    /// [`View::entries(self).reduce_grouped()`](super::AsyncView::reduce_grouped)

1041

    /// instead. The parameters for the query can be customized on the builder

1042

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1043

    #[must_use]

1044

12

    async fn reduce_grouped<V: schema::SerializedView, Key>(

1045

12

        &self,

1046

12

        key: Option<QueryKey<'_, V::Key, Key>>,

1047

12

        access_policy: AccessPolicy,

1048

12

    ) -> Result<GroupedReductions<V>, Error>

1049

12

    where

1050

12

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

1051

12

        V::Key: Borrow<Key> + PartialEq<Key>,

1052

12

1053

12

        let view = self.schematic().view::<V>()?;

1054

12

        self.reduce_grouped_by_name(

1055

12

            &view.view_name(),

1056

12

            key.map(|key| key.serialized()).transpose()?,

1057

12

            access_policy,

1058

1059

14

        .await?

1060

12

        .into_iter()

1061

31

        .map(|map| {

1062

31

            Ok(MappedValue::new(

1063

31

                V::Key::from_ord_bytes(ByteSource::Borrowed(&map.key))

1064

31

                    .map_err(view::Error::key_serialization)?,

1065

31

                V::deserialize(&map.value)?,

1066

))

1067

31

})

1068

12

        .collect::<Result<Vec<_>, Error>>()

1069

24

1070

1071

    /// Deletes all of the documents associated with this view.

1072

///

1073

    /// This is the lower-level API. For better ergonomics, consider querying

1074

    /// the view using

1075

    /// [`View::entries(self).delete_docs()`](super::AsyncView::delete_docs)

1076

    /// instead. The parameters for the query can be customized on the builder

1077

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1078

    #[must_use]

1079

13

    async fn delete_docs<V: schema::SerializedView, Key>(

1080

13

        &self,

1081

13

        key: Option<QueryKey<'_, V::Key, Key>>,

1082

13

        access_policy: AccessPolicy,

1083

13

    ) -> Result<u64, Error>

1084

13

    where

1085

13

        Key: KeyEncoding<V::Key> + PartialEq + ?Sized,

1086

13

        V::Key: Borrow<Key> + PartialEq<Key>,

1087

13

1088

13

        let view = self.schematic().view::<V>()?;

1089

13

        self.delete_docs_by_name(

1090

13

            &view.view_name(),

1091

13

            key.map(|key| key.serialized()).transpose()?,

1092

13

            access_policy,

1093

1094

13

        .await

1095

26

1096

1097

    /// Applies a [`Transaction`] to the [`Schema`](schema::Schema). If any

1098

    /// operation in the [`Transaction`] fails, none of the operations will be

1099

    /// applied to the [`Schema`](schema::Schema).

1100

    async fn apply_transaction(

1101

        &self,

1102

        transaction: Transaction,

1103

    ) -> Result<Vec<OperationResult>, Error>;

1104

1105

    /// Retrieves the document with `id` stored within the named `collection`.

1106

///

1107

    /// This is a lower-level API. For better ergonomics, consider using one of:

1108

///

1109

    /// - [`SerializedCollection::get_async()`]

1110

    /// - [`self.collection::<Collection>().get()`](super::AsyncCollection::get)

1111

    async fn get_from_collection(

1112

        &self,

1113

        id: DocumentId,

1114

        collection: &CollectionName,

1115

    ) -> Result<Option<OwnedDocument>, Error>;

1116

1117

    /// Retrieves all documents matching `ids` from the named `collection`.

1118

    /// Documents that are not found are not returned, but no error will be

1119

    /// generated.

1120

///

1121

    /// This is a lower-level API. For better ergonomics, consider using one of:

1122

///

1123

    /// - [`SerializedCollection::get_multiple_async()`]

1124

    /// - [`self.collection::<Collection>().get_multiple()`](super::AsyncCollection::get_multiple)

1125

    async fn get_multiple_from_collection(

1126

        &self,

1127

        ids: &[DocumentId],

1128

        collection: &CollectionName,

1129

    ) -> Result<Vec<OwnedDocument>, Error>;

1130

1131

    /// Retrieves all documents within the range of `ids` from the named

1132

    /// `collection`. To retrieve all documents, pass in `..` for `ids`.

1133

///

1134

    /// This is a lower-level API. For better ergonomics, consider using one of:

1135

///

1136

    /// - [`SerializedCollection::all().headers()`](schema::List::headers)

1137

    /// - [`self.collection::<Collection>().all().headers()`](super::List::headers)

1138

    /// - [`SerializedCollection::list().headers()`](schema::List::headers)

1139

    /// - [`self.collection::<Collection>().list().headers()`](super::List::headers)

1140

    async fn list_from_collection(

1141

        &self,

1142

        ids: Range<DocumentId>,

1143

        order: Sort,

1144

        limit: Option<u32>,

1145

        collection: &CollectionName,

1146

    ) -> Result<Vec<OwnedDocument>, Error>;

1147

1148

    /// Retrieves all headers within the range of `ids` from the named

1149

    /// `collection`. To retrieve all documents, pass in `..` for `ids`.

1150

///

1151

    /// This is a lower-level API. For better ergonomics, consider using one of:

1152

///

1153

    /// - [`SerializedCollection::all().headers()`](schema::AsyncList::headers)

1154

    /// - [`self.collection::<Collection>().all().headers()`](super::AsyncList::headers)

1155

    /// - [`SerializedCollection::list().headers()`](schema::AsyncList::headers)

1156

    /// - [`self.collection::<Collection>().list().headers()`](super::AsyncList::headers)

1157

    async fn list_headers_from_collection(

1158

        &self,

1159

        ids: Range<DocumentId>,

1160

        order: Sort,

1161

        limit: Option<u32>,

1162

        collection: &CollectionName,

1163

    ) -> Result<Vec<Header>, Error>;

1164

1165

    /// Counts the number of documents within the range of `ids` from the named

1166

    /// `collection`.

1167

///

1168

    /// This is a lower-level API. For better ergonomics, consider using one of:

1169

///

1170

    /// - [`SerializedCollection::all_async().count()`](schema::AsyncList::count)

1171

    /// - [`self.collection::<Collection>().all().count()`](super::AsyncList::count)

1172

    /// - [`SerializedCollection::list_async().count()`](schema::AsyncList::count)

1173

    /// - [`self.collection::<Collection>().list().count()`](super::AsyncList::count)

1174

    async fn count_from_collection(

1175

        &self,

1176

        ids: Range<DocumentId>,

1177

        collection: &CollectionName,

1178

    ) -> Result<u64, Error>;

1179

1180

    /// Compacts the collection to reclaim unused disk space.

1181

///

1182

    /// This process is done by writing data to a new file and swapping the file

1183

    /// once the process completes. This ensures that if a hardware failure,

1184

    /// power outage, or crash occurs that the original collection data is left

1185

    /// untouched.

1186

///

1187

    /// ## Errors

1188

///

1189

    /// * [`Error::CollectionNotFound`]: database `name` does not exist.

1190

    /// * [`Error::Other`]: an error occurred while compacting the database.

1191

    async fn compact_collection_by_name(&self, collection: CollectionName) -> Result<(), Error>;

1192

1193

    /// Queries for view entries from the named `view`.

1194

///

1195

    /// This is the lower-level API. For better ergonomics, consider querying

1196

    /// the view using [`View::entries(self).query()`](super::AsyncView::query)

1197

    /// instead. The parameters for the query can be customized on the builder

1198

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1199

    async fn query_by_name(

1200

        &self,

1201

        view: &ViewName,

1202

        key: Option<SerializedQueryKey>,

1203

        order: Sort,

1204

        limit: Option<u32>,

1205

        access_policy: AccessPolicy,

1206

    ) -> Result<Vec<schema::view::map::Serialized>, Error>;

1207

1208

    /// Queries for view entries from the named `view` with their source

1209

    /// documents.

1210

///

1211

    /// This is the lower-level API. For better ergonomics, consider querying

1212

    /// the view using [`View::entries(self).query_with_docs()`](super::AsyncView::query_with_docs) instead.

1213

    /// The parameters for the query can be customized on the builder returned

1214

    /// from [`AsyncConnection::view()`](super::AsyncConnection::view).

1215

    async fn query_by_name_with_docs(

1216

        &self,

1217

        view: &ViewName,

1218

        key: Option<SerializedQueryKey>,

1219

        order: Sort,

1220

        limit: Option<u32>,

1221

        access_policy: AccessPolicy,

1222

    ) -> Result<schema::view::map::MappedSerializedDocuments, Error>;

1223

1224

    /// Reduces the view entries from the named `view`.

1225

///

1226

    /// This is the lower-level API. For better ergonomics, consider querying

1227

    /// the view using

1228

    /// [`View::entries(self).reduce()`](super::AsyncView::reduce)

1229

    /// instead. The parameters for the query can be customized on the builder

1230

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1231

    async fn reduce_by_name(

1232

        &self,

1233

        view: &ViewName,

1234

        key: Option<SerializedQueryKey>,

1235

        access_policy: AccessPolicy,

1236

    ) -> Result<Vec<u8>, Error>;

1237

1238

    /// Reduces the view entries from the named `view`, reducing the values by each

1239

    /// unique key.

1240

///

1241

    /// This is the lower-level API. For better ergonomics, consider querying

1242

    /// the view using

1243

    /// [`View::entries(self).reduce_grouped()`](super::AsyncView::reduce_grouped)

1244

    /// instead. The parameters for the query can be customized on the builder

1245

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1246

    async fn reduce_grouped_by_name(

1247

        &self,

1248

        view: &ViewName,

1249

        key: Option<SerializedQueryKey>,

1250

        access_policy: AccessPolicy,

1251

    ) -> Result<Vec<MappedSerializedValue>, Error>;

1252

1253

    /// Deletes all source documents for entries that match within the named

1254

    /// `view`.

1255

///

1256

    /// This is the lower-level API. For better ergonomics, consider querying

1257

    /// the view using

1258

    /// [`View::entries(self).delete_docs()`](super::AsyncView::delete_docs)

1259

    /// instead. The parameters for the query can be customized on the builder

1260

    /// returned from [`AsyncConnection::view()`](super::AsyncConnection::view).

1261

    async fn delete_docs_by_name(

1262

        &self,

1263

        view: &ViewName,

1264

        key: Option<SerializedQueryKey>,

1265

        access_policy: AccessPolicy,

1266

    ) -> Result<u64, Error>;

1267

1268

1269

/// Access to a connection's schema.

1270

pub trait HasSchema {

1271

    /// Returns the schema for the database.

1272

    fn schematic(&self) -> &Schematic;

1273