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
//! Repositories are abstraction over a specific mongo collection for a given `Model`

use crate::Model;
use mongodb::bson::doc;
use mongodb::bson::from_bson;
use mongodb::bson::to_bson;
use mongodb::bson::Bson;
use mongodb::bson::Document;
use mongodb::options::*;
use mongodb::results::*;
use serde::Deserialize;
use std::collections::HashMap;

/// Associate a mongo client to a `Database` and a `Model`.
///
/// This type can safely be copied and passed around because `std::sync::Arc` is used internally.
#[derive(Debug, Clone)]
pub struct Repository<M: Model> {
    db: mongodb::Database, // TODO: once indexes are officially supported in the driver, we should get all required opterations in `Collection` and remove this field
    coll: mongodb::Collection,
    _pd: std::marker::PhantomData<M>,
}

impl<M: Model> Repository<M> {
    /// Create a new repository from the given mongo client.
    pub fn new(db: mongodb::Database) -> Self {
        let coll = if let Some(options) = M::coll_options() {
            db.collection_with_options(M::coll_name(), options)
        } else {
            db.collection(M::coll_name())
        };

        Self {
            db,
            coll,
            _pd: std::marker::PhantomData,
        }
    }

    /// Create a new repository with associated collection options (override `Model::coll_options`).
    pub fn new_with_options(db: mongodb::Database, options: CollectionOptions) -> Self {
        Self {
            coll: db.collection_with_options(M::coll_name(), options),
            db,
            _pd: std::marker::PhantomData,
        }
    }

    /// Returns associated `M::COLL_NAME`.
    pub fn coll_name(&self) -> &'static str {
        M::coll_name()
    }

    /// Returns underlying `mongodb::Collection`.
    pub fn get_underlying(&self) -> mongodb::Collection {
        mongodb::Collection::clone(&self.coll)
    }

    /// Drops the underlying collection, deleting all data, users, and indexes stored inside.
    pub async fn drop(
        &self,
        options: impl Into<Option<DropCollectionOptions>>,
    ) -> mongodb::error::Result<()> {
        self.coll.drop(options).await
    }

    /// Runs an aggregation operation.
    ///
    /// [Mongo manual](https://docs.mongodb.com/manual/aggregation/)
    pub async fn aggregate(
        &self,
        pipeline: impl IntoIterator<Item = Document>,
        options: impl Into<Option<AggregateOptions>>,
    ) -> mongodb::error::Result<crate::cursor::ModelCursor<M>> {
        self.coll
            .aggregate(pipeline, options)
            .await
            .map(crate::cursor::ModelCursor::from)
    }

    /// Estimates the number of documents in the collection using collection metadata.
    pub async fn estimated_document_count(
        &self,
        options: impl Into<Option<EstimatedDocumentCountOptions>>,
    ) -> mongodb::error::Result<i64> {
        self.coll.estimated_document_count(options).await
    }

    /// Gets the number of documents matching `filter`.
    ///
    /// Note that using `Repository::estimated_document_count` is recommended instead of this method is most cases.
    pub async fn count_documents(
        &self,
        filter: impl Into<Option<Document>>,
        options: impl Into<Option<CountOptions>>,
    ) -> mongodb::error::Result<i64> {
        self.coll.count_documents(filter, options).await
    }

    /// Deletes all documents stored in the collection matching `query`.
    pub async fn delete_many(
        &self,
        query: Document,
        options: impl Into<Option<DeleteOptions>>,
    ) -> mongodb::error::Result<DeleteResult> {
        self.coll.delete_many(query, options).await
    }

    /// Deletes up to one document found matching `query`.
    pub async fn delete_one(
        &self,
        query: Document,
        options: impl Into<Option<DeleteOptions>>,
    ) -> mongodb::error::Result<DeleteResult> {
        self.coll.delete_one(query, options).await
    }

    /// Finds the distinct values of the field specified by `field_name` across the collection.
    pub async fn distinct(
        &self,
        field_name: &str,
        filter: impl Into<Option<Document>>,
        options: impl Into<Option<DistinctOptions>>,
    ) -> mongodb::error::Result<Vec<M>> {
        let bson_items = self.coll.distinct(field_name, filter, options).await?;
        let mut items = Vec::with_capacity(bson_items.len());
        for bson in bson_items {
            let item =
                from_bson(bson).map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
            items.push(item);
        }
        Ok(items)
    }

    /// Finds the documents in the collection matching `filter`.
    pub async fn find(
        &self,
        filter: impl Into<Option<Document>>,
        options: impl Into<Option<FindOptions>>,
    ) -> mongodb::error::Result<crate::cursor::ModelCursor<M>> {
        self.coll
            .find(filter, options)
            .await
            .map(crate::cursor::ModelCursor::from)
    }

    /// Finds a single document in the collection matching `filter`.
    pub async fn find_one(
        &self,
        filter: impl Into<Option<Document>>,
        options: impl Into<Option<FindOneOptions>>,
    ) -> mongodb::error::Result<Option<M>> {
        let doc_opt = self.coll.find_one(filter, options).await?;
        if let Some(doc) = doc_opt {
            let item = Self::h_doc_to_model(doc)?;
            Ok(Some(item))
        } else {
            Ok(None)
        }
    }

    /// Atomically finds up to one document in the collection matching `filter` and deletes it.
    pub async fn find_one_and_delete(
        &self,
        filter: Document,
        options: impl Into<Option<FindOneAndDeleteOptions>>,
    ) -> mongodb::error::Result<Option<M>> {
        let doc_opt = self.coll.find_one_and_delete(filter, options).await?;
        if let Some(doc) = doc_opt {
            let item = Self::h_doc_to_model(doc)?;
            Ok(Some(item))
        } else {
            Ok(None)
        }
    }

    /// Atomically finds up to one document in the collection matching `filter` and replaces it with
    /// `replacement`.
    pub async fn find_one_and_replace(
        &self,
        filter: Document,
        replacement: &M,
        options: impl Into<Option<FindOneAndReplaceOptions>>,
    ) -> mongodb::error::Result<Option<M>> {
        let replacement = Self::h_model_to_doc(replacement)?;
        let doc_opt = self
            .coll
            .find_one_and_replace(filter, replacement, options)
            .await?;
        if let Some(doc) = doc_opt {
            let item = Self::h_doc_to_model(doc)?;
            Ok(Some(item))
        } else {
            Ok(None)
        }
    }

    /// Atomically finds up to one model in the collection matching `filter` and updates it.
    ///
    /// Both `Document` and `Vec<Document>` implement `Into<UpdateModifications>`, so either can be
    /// passed in place of constructing the enum case. Note: pipeline updates are only supported
    /// in MongoDB 4.2+.
    pub async fn find_one_and_update(
        &self,
        filter: Document,
        update: impl Into<UpdateModifications>,
        options: impl Into<Option<FindOneAndUpdateOptions>>,
    ) -> mongodb::error::Result<Option<M>> {
        let doc_opt = self
            .coll
            .find_one_and_update(filter, update, options)
            .await?;
        if let Some(doc) = doc_opt {
            let item = Self::h_doc_to_model(doc)?;
            Ok(Some(item))
        } else {
            Ok(None)
        }
    }

    /// Inserts the models into the collection.
    pub async fn insert_many(
        &self,
        models: impl IntoIterator<Item = M>,
        options: impl Into<Option<InsertManyOptions>>,
    ) -> mongodb::error::Result<InsertManyResult> {
        let mut docs = Vec::new();
        for model in models.into_iter() {
            docs.push(Self::h_model_to_doc(&model)?)
        }

        self.coll.insert_many(docs, options).await
    }

    /// Inserts model `M` into the collection.
    pub async fn insert_one(
        &self,
        model: &M,
        options: impl Into<Option<InsertOneOptions>>,
    ) -> mongodb::error::Result<InsertOneResult> {
        let doc = Self::h_model_to_doc(model)?;
        self.coll.insert_one(doc, options).await
    }

    /// Replaces up to one document matching `query` in the collection with `replacement`.
    pub async fn replace_one(
        &self,
        query: Document,
        replacement: &M,
        options: impl Into<Option<ReplaceOptions>>,
    ) -> mongodb::error::Result<UpdateResult> {
        let replacement = Self::h_model_to_doc(replacement)?;
        self.coll.replace_one(query, replacement, options).await
    }

    /// Updates all documents matching `query` in the collection.
    ///
    /// Both `Document` and `Vec<Document>` implement `Into<UpdateModifications>`, so either can be
    /// passed in place of constructing the enum case. Note: pipeline updates are only supported
    /// in MongoDB 4.2+. See the official MongoDB
    /// [documentation](https://docs.mongodb.com/manual/reference/command/update/#behavior) for more information on specifying updates.
    pub async fn update_many(
        &self,
        query: Document,
        update: impl Into<UpdateModifications>,
        options: impl Into<Option<UpdateOptions>>,
    ) -> mongodb::error::Result<UpdateResult> {
        self.coll.update_many(query, update, options).await
    }

    /// Updates up to one document matching `query` in the collection.
    ///
    /// Both `Document` and `Vec<Document>` implement `Into<UpdateModifications>`, so either can be
    /// passed in place of constructing the enum case. Note: pipeline updates are only supported
    /// in MongoDB 4.2+. See the official MongoDB
    /// [documentation](https://docs.mongodb.com/manual/reference/command/update/#behavior) for more information on specifying updates.
    pub async fn update_one(
        &self,
        query: Document,
        update: impl Into<UpdateModifications>,
        options: impl Into<Option<UpdateOptions>>,
    ) -> mongodb::error::Result<UpdateResult> {
        self.coll.update_one(query, update, options).await
    }

    /// Synchronize model with underlying mongo collection.
    ///
    /// This should be called once per model on startup to synchronize indexes defined
    /// by the `Model`. Indexes found in the backend and not defined in the model are
    /// destroyed except for the special index "_id".
    pub async fn sync_indexes(&self) -> Result<(), mongodb::error::Error> {
        let mut indexes = M::indexes();

        match self
            .h_run_command(doc! { "listIndexes": M::coll_name() })
            .await
        {
            Ok(ret) => {
                let parsed_ret: ListIndexesRet = from_bson(Bson::Document(ret))
                    .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;

                if parsed_ret.cursor.id != 0 {
                    // batch isn't complete
                    Err(std::io::Error::new(
                        std::io::ErrorKind::Other,
                        format!("couldn't list all indexes from '{}'", M::coll_name()),
                    ))?;
                }

                let mut existing_indexes = HashMap::new();
                for index in parsed_ret.cursor.first_batch {
                    if let Some(key) = index.get("key") {
                        existing_indexes.insert(key.to_string(), index);
                    }
                }

                let mut already_sync = Vec::new();
                let mut to_drop = Vec::new();
                for (i, index) in indexes.0.clone().into_iter().enumerate() {
                    let index_doc = index.into_document();
                    let key = index_doc.get("key").ok_or_else(|| {
                        std::io::Error::new(std::io::ErrorKind::Other, "index doc is missing 'key'")
                    })?;
                    if let Some(mut existing_index) = existing_indexes.remove(&key.to_string()) {
                        // "ns" and "v" in the response should not be used for the comparison
                        existing_index.remove("ns");
                        existing_index.remove("v");

                        if doc_are_eq(&index_doc, &existing_index) {
                            already_sync.push(i);
                        } else {
                            // An index with the same specification already exists, we need to drop it.
                            to_drop.push(
                                index_doc
                                    .get_str("name")
                                    .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?
                                    .to_owned(),
                            );
                        }
                    }
                }

                // Drop all remaining existing index expect "_id_" (for the "_id" key)
                // "_id" is special and cannot be deleted.
                // https://api.mongodb.com/wiki/current/Indexes.html#Indexes-The%5CidIndex
                for existing_index in existing_indexes.values() {
                    let name = existing_index
                        .get_str("name")
                        .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?
                        .to_owned();
                    if name != "_id_" {
                        to_drop.push(name);
                    }
                }

                if !to_drop.is_empty() {
                    // Actually send the drop command
                    // Dropping multiple indexes is available only starting MongoDB 4.2
                    // If this fails, we fallback to a loop dropping all indexes individually
                    // TODO: it would be better to select the method by checking mongo version, but db.version()
                    // is not yet exposed by the driver.
                    if self
                        .h_run_command(doc! { "dropIndexes": M::coll_name(), "index": &to_drop })
                        .await
                        .is_err()
                    {
                        for index_name in to_drop {
                            self.h_run_command(
                                doc! { "dropIndexes": M::coll_name(), "index": index_name },
                            )
                            .await?;
                        }
                    }
                }

                // Ignore index already in sync
                for i in already_sync.into_iter().rev() {
                    indexes.0.remove(i);
                }
            }
            Err(e) => {
                match e.kind.as_ref() {
                    mongodb::error::ErrorKind::CommandError(err) if err.code == 26 => {
                        // Namespace doesn't exists yet as such no index is present either.
                    }
                    _ => return Err(e),
                }
            }
        }

        if !indexes.0.is_empty() {
            self.h_run_command(indexes.create_indexes_command(M::coll_name()))
                .await?;
        }

        Ok(())
    }

    async fn h_run_command(
        &self,
        command_doc: Document,
    ) -> Result<Document, mongodb::error::Error> {
        let ret = self
            .db
            .run_command(
                command_doc,
                Some(SelectionCriteria::ReadPreference(ReadPreference::Primary)),
            )
            .await?;
        if let Ok(err) = from_bson::<mongodb::error::CommandError>(Bson::Document(ret.clone())) {
            Err(mongodb::error::Error::from(
                mongodb::error::ErrorKind::CommandError(err),
            ))
        } else {
            Ok(ret)
        }
    }

    fn h_doc_to_model(doc: Document) -> mongodb::error::Result<M> {
        let item = from_bson(Bson::Document(doc))
            .map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
        Ok(item)
    }

    fn h_model_to_doc(model: &M) -> mongodb::error::Result<Document> {
        let bson =
            to_bson(&model).map_err(|e| std::io::Error::new(std::io::ErrorKind::Other, e))?;
        if let Bson::Document(doc) = bson {
            Ok(doc)
        } else {
            Err(mongodb::error::Error::from(std::io::Error::new(
                std::io::ErrorKind::Other,
                "model can't be serialized into a `Bson::Document`",
            )))
        }
    }
}

#[derive(Deserialize)]
struct ListIndexesRet {
    pub cursor: Cursor,
}

#[derive(Deserialize)]
struct Cursor {
    pub id: i64,
    #[serde(rename = "firstBatch", default)]
    pub first_batch: Vec<Document>,
}

fn doc_are_eq(a: &Document, b: &Document) -> bool {
    if a.len() != b.len() {
        return false;
    }

    for (key, a_val) in a {
        match b.get(key) {
            Some(b_val) if a_val != b_val => {
                return false;
            }
            Some(_) => {}
            None => {
                return false;
            }
        }
    }

    true
}