crud-mongoose

1.0.46 • Public • Published

Crud-Mongoose

Crud mongoose is a library of middleware for connecting crud to mongoose. It make it pretty easy to create simple APIs, with plenty of default and configurable capabilities.

Simple Example:

var crud = require('crud'),
  cm = require('crud-mongoose'),
  mongoose = require('mongoose'),
  Model = mongoose.model(
    'users',
    new mongoose.Schema({
      firstName: { type: String, required: true },
      lastName: { type: String, required: true },
      gender: { type: String, required: true, enum: ['M', 'F'] },
      created: { type: Date, default: Date.now },
    })
  );

// All Users -------------------------------------------------------------------

crud
  .entity('/users')
  .Create()
  .pipe(cm.createNew(Model));

crud
  .entity('/users')
  .Read()
  .pipe(cm.findAll(Model));

crud
  .entity('/users')
  .Delete()
  .pipe(cm.removeAll(Model));

// One User --------------------------------------------------------------------

crud
  .entity('/users/:_id')
  .Read()
  .pipe(cm.findOne(Model));

crud
  .entity('/users/:_id')
  .Update()
  .pipe(cm.updateOne(Model));

crud
  .entity('/users/:_id')
  .Delete()
  .pipe(cm.removeOne(Model));

With this example, you can do simple GET, POST, PUT, and DELETE routes to:

  • Create a user - POST on /api/users
  • Read many users - GET on /api/users
  • Delete many users - DELETE on /api/users
  • Read one user - GET on /api/users/<id>
  • Update one user - PUT on /api/users/<id>
  • Delete one user - DELETE on /api/users/<id>

There are also a number of default query-able parameters. For example, the following GET requests could be used to get different subsets of data:

  • /api/users?gender=F - get all female users
  • /api/users?lastName=Thompson - get all users with lastName "Thompson"
  • /api/users?limit=10 - only get 10 users
  • /api/users?skip=50&limit=100&sortBy=created:asc - Sorts users by created date ascending, skips 50, and then returns max 100
  • /api/users?page=2&perPage=100&sortBy=created - Sorts users by created date ascending (note it's ascending by default), and then skips the first 200 because it's page 2 (starts at 0) and there are only 100 per page.
  • /api/users?fields=firstName,lastName - get all users but only return firstName and lastName fields.

There are other queries you can make, but this gives a pretty good example of why this library makes API creation so easy. BUT, it's not only easy, it's also really configurable so you are not boxed in. And, if you need to configure things this library doesn't allow, you can always drop in your own middleware function into Crud, which is pretty easy anyway.

API

The API has two types of middleware functions:

  • Middleware that interacts directly with the MongoDB through mongoose - here.
  • Middleware that modifies the crud data and query objects - here.

All middleware can be placed into a Crud Pipe, like this:

crud('/users')
  .Read()
  .pipe(cm.findAll());

Mongoose Middleware

The following middleware functions all interact with a Mongoose Model, using the crud data and query objects. (See here for info on the meaning of those objects).

# cm.createNew(Model)

This method creates a new record in the Model based on the crud data object. The middleware is basically this:

function(data, query, callback) {
  new Model(data).save(callback);
}

Example, allowing you to POST on /api/users to create a new user:

crud
  .entity('/users')
  .Create()
  .pipe(cm.createNew(Model));

# cm.findAll(Model, [fields])

This method does a find on the Model using the query object. The fields parameter is an optional array of fields you want to allow in the response. So, if you only want to show certain fields you can provide an array like ['firstName', 'lastName']. Or, if you wish to restrict a field ALWAYS, you can put a - before it (e.g. ['firstname', 'lastname', '-password'] will by default only show first and last names, but will not allow you to ever query for the password fields).

For starters, you can think of it as something like this:

function (data, query, callback) {
  Model.find(query).select(fields.join(' ')).lean().exec(callback);
  // see here: http://mongoosejs.com/docs/api.html#query_Query-select for mongoose select
}

That's where it started, but then we added a lot more functionality. There are special query parameters that are parsed here:

  • limit - limits the number of records responded with
  • skip - skips this many records on the response
  • page - used for pagination. Indexed starting at zero.
  • perPage - number of records per page. Pagination doesn't work well without a value here.
  • sortBy - sortable values. This is formatted like this: firstvalue:asc,secondvalue:desc, where the comma denotes different fields and the colon is used to show asc for ascending or desc for descending. Default is ascending.
  • fields - a comma separated listed of fields to select.

An example, allowing you to query all users with a GET on /api/users:

crud
  .entity('/users')
  .Read()
  .pipe(cm.findAll(Model, ['firstName', 'lastName']));

The following are the chainable properties:

  • findAll(Model).stream() - sets the mode to streaming. This will not pass things through the crud chain anymore and instead stream responses. This is useful for queries that expect huge responses.
  • findAll(Model).exports(object) - turns on exporting capabilities. This allows you to set export functionality so, for example, query ?export=csv will export the data as a csv file. The object parameter specifies keys as the export keys like 'csv' and the values are functions that handle the exporting. An example would be findAll(Model).exports({ csv: cm.exporters.csv() }). See exporter functions for information on how they are defined and the pre-packaged functions in crud-mongoose.
  • findAll(Model).metadata(allow) - allow is boolean deciding whether or not to send metadata with responses. The default is false. Metadata will only be sent when a query has set page and perPage, because it's not useful otherwise.

# cm.findAggregation(Model, [fields])

This method does an aggregate one the Model using the query object as a $match pipeline.

NOTE: This feature is in beta. We haven't had a ton of time to test, but needed it for some work.

Everything operates the same as findAll, with these changes:

Removed Functionality from findAll:

  • Export: no exporting functionality
  • Metadata: no metadata functionality

Adding Functionality from findAll:

  • findAggregation(Model).additionalStages(spot='end', stages) - adds additional pipeline stages to the aggregation. This runs after the projection and before the sort/limit/skip functionality. The argument spot indicates where to put these stages in the pipeline. The options are start, end, preSort, and afterLimit (good for joins that don't affect the $match but are expensive). The argument stages is an array, where each value is a MongoDB pipeline stage. If the value is an object, it will be put in the pipeline. If the value is a function it will be called with one argument (getDefaultProjection), which when called returns an object projecting all standard fields for the Model so you can add new ones. The context of the function call will be the same as the crud context.
  • findAggregation(Model).sortPresets(presets) - This allows you to add new fields for sorting. For example, you can add a new field that is the sum of two fields. presets is a key-value lookup that will be added to the $project stage. These values will not be added unless the sort key is used in the sortBy query parameter to prevent additional work that's unnecessary.
  • findAggregation(Model).cursor(fn) - This allows you to modify the cursor. This function is passed (cursor) and must respond with the cursor. An example: .cursor(cursor => { return cursor.allowDiskUse(true); }

Example:

```js
.pipe(cm.findAggregation(Model).additionalStages('start', [
  { $lookup: {
      from: 'users',
      localField: 'user',
      foreignField: '_id',
      as: 'user'
  } }
]).sortPresets({
  totalLeads: { $sum: [
    '$applicationStats.facebook.qualified',
    '$applicationStats.facebook.unqualified',
    '$applicationStats.jobBoard.qualified',
    '$applicationStats.jobBoard.unqualified',
    '$applicationStats.embedded.qualified',
    '$applicationStats.embedded.unqualified'
  ] }
}))
```

# cm.findOne(Model, [fields])

This method does a findOne on the Model using the query object. The fields parameter is an optional array of fields you want to allow in the response. So, if you only want to show certain fields you can provide an array like ['firstName', 'lastName']. Or, if you wish to restrict a field ALWAYS, you can put a - before it (e.g. ['firstname', 'lastname', '-password'] will by default only show first and last names, but will not allow you to ever query for the password fields).

For starters, you can think of it as something like this:

function (data, query, callback) {
  Model.findOne(query).select(fields.join(' '))
       .lean().exec(callback);
  // see here: http://mongoosejs.com/docs/api.html#query_Query-select for mongoose select
}

That's where it started, but then we added more functionality. There are special query parameters that are parsed here: (well, there is only one right now).

  • fields - a comma separated listed of fields to select.

An example, allowing you to query one user with GET on /api/users/<id>:

crud
  .entity('/users/:_id')
  .Read()
  .pipe(cm.findOne(Model, ['firstName', 'lastName']));

Note: It's important for the URL to be formatted like /users/:_id, because then the query will have { _id: <id_in_url> }, which is needed to find the individual user.

# cm.removeAll(Model)

This method does a remove on the Model using the query object. The middleware is basically this:

function(data, query, callback) {
  Model.remove(query).lean().callback();
}

An example so you can remove all users with a DELETE on /api/users:

crud
  .entity('/users')
  .Delete()
  .pipe(cm.removeAll(Model));

# cm.removeOne(Model)

This method does a findOneAndRemove on the Model using the query object. The middleware is basically this:

function(data, query, callback) {
  Model.findOneAndRemove(query).lean().callback();
}

An example so you can remove one user with a DELETE on /api/users/<id>:

crud
  .entity('/users/:_id')
  .Delete()
  .pipe(cm.removeOne(Model));

Note: It's important for the URL to be formatted like /users/:_id, because then the query will have { _id: <id_in_url> }, which is needed to find the individual user.

# cm.updateOne(Model, Options)

This method does a findOne and then and update on the Model using the query object for querying and the data object for the update. The middleware is basically this:

function(data, query, callback) {
  Model.findOne(query, function(e, d) {
    d.set(data);
    d.save(callback);
  });
}

If the data value has a value for **v, it is removed: if ('**v' in data) delete data.\_\_v. This is because document versioning was causing problems. If this removal causes problems, we should revisit this.

Options

  • findOneAndUpdate (Default=false) - We used to use findOneAndUpdate, but decided to do a find then update because updates do not use mongoose validators. Unfortunately, this does not allow you to apply MongoDB updates like $push, $pull, etc. So, if you need to use other update methods, you can set this to true. Just remember this will not obey the Mongoose validation. Additionally, you may need this if you're trying to update a Mixed mongoose object because Mixed objects cannot be updated with the object.set function.

Example - An example so you can update a user with a PUT on /api/users/<id>:

crud
  .entity('/users/:_id')
  .Update()
  .pipe(cm.updateOne(Model));

Note: It's important for the URL to be formatted like /users/:_id, because then the query will have { _id: <id_in_url> }, which is needed to find the individual user.

Modify Middleware

These middleware functions modify crud data and query objects. (See here for info on the meaning of those objects). The purpose of these middleware functions are for the instances where you want to set default parameters (e.g. default limit on a read of 100 records), prevent the use of certain parameters (e.g. don't allow querying of users by age), establishing max query limits (e.g. cannot query for more than 100 records), or overriding certain values (e.g. the data on an update always updating the "updated" date field).

# cm.parseQuery()

This method allows you to parse the query object with special options. Calling this function returns a middleware function that has chainable properties so you can configure the function. For example, you could chain configurations like this:

crud('/users')
  .Read()
  .pipe(cm.parseQuery().defaults({ limit: 10 }))
  .pipe(cm.findAll());

The following are the chainable properties:

  • parseQuery().defaults(defaults) - defaults is a key-value object that will set any key-value parameters on the query object if they are not set. For example, if you have { limit: 10 } as the defaults, then if you were to query /api/users it would be equivalent to /api/users?limit=10. Also, if any values in the key-values are functions, then they will be called when set on the query.

  • parseQuery().removes(key1, [key2, key3 ... ]) - the keys, passed as individual arguments, are keys that will be removed from the query so API users cannot query by this field. For example, if you have .removes('age'), then you cannot query like this: /api/users?age=23, because age will be removed from the query.

  • parseData().required(key1, [key2, key3 ... ]) - the keys, passed as individual arguments, are keys indicate fields that need to be present on the query object. An empty string does not count as "present".

  • parseQuery().overrides(overrides) - overrides is a key-value object that will force its key-value parameters on the query object. For example, if you have { active: true } as the overrides, then if you were to query /api/users?active=false it would be equivalent to /api/users?active=true. You could use this to prevent API users from querying inactive records. Also, if any values in the key-values are functions, then they will be called when set on the query.

  • parseQuery().maxes(maxes) - maxes is a key-value object that will force any key-value parameters on the query object that are in the maxes object to be no-more-than the max. For example, if you have { limit: 100 } as the maxes, then if you were to query /api/users?limit=1000, it would override the limit parameter to be 100, since that is the max.

# cm.parseData()

This method allows you to parse the data object with special options. Calling this function returns a middleware function that has chainable properties so you can configure the function. For example, you could chain configurations like this:

crud('/users/:_id')
  .Update()
  .pipe(cm.parseData().overrides({ updated: Date.now }))
  .pipe(cm.findAll());

The chainable properties are the SAME as those from parseQuery, except these important things:

  • They modify the data object (not the query object)
  • There is no maxes method, because that is made specifically to handle the query limits, pages, etc.
  • On required, anything that is not truthy AND is not the boolean false is not considered present. This allows you to pass false to a required field, but not null or ''.
  • You can request which fields should not be flattened by doing .dontFlatten('fieldNotToFlatten'), which is useful for Mixed types in Mongoose.

Note, if you remove the ability to send certain data, like parseData().removes('info'), this does not mean the user cannot updated pass something like { 'info.age' : 7 }, which does update mongoose documents because everything is treated as flat in mongoose. Likewise, if you have an array such as { favcolors: ['red', 'blue', 'green'] }, you could pass { 'favcolors.0' : 'black' }. Because of this, if key in the data object has a period (meaning it's already been flattened somewhat), parseData will remove this key-value from the resultant data object. This is solely for security reasons.

Exporter Functions

Exporter functions are used to export the api data into something other than JSON, say CSV for example.

The are passed three arguments:

  • response - express response object.
  • cursor - The mongoose cursor with the rest of the query parameters set
  • callback - callback to be called when done. The first argument should be the error message if there is one or null.

Prepackaged in crud-mongoose are the folloowing exporters, which can be accessed by cm.exporters. If you're using the csv exporter you could do the following:

crud.entity('/users').Read()
  .pipe(cm.findAll(Model)
          .exports({ csv: cm.exporters.csv() })

# cm.exporters.csv([headers])

Exports the resulting query as a csv file. The optional headers argument allows you to specify an array of csv headers, which will be the first line and will be used as the accessors for the data values for each row. If not provided, the key's on the first row will be used.

# cm.exporters.xlsx([headers])

Exports the resulting query as a xlsx file. The optional headers argument allows you to specify an array of xlsx headers, which will be the first line and will be used as the accessors for the data values for each row. If not provided, the key's on the first row will be used.

Debug

The Crud module has sprinkled some debug messages throughout the module. If you wish to turn these on, run your sever with the environment variable DEBUG=crud-mongoose* set. Or, you can turn them on for only one method, like DEBUG=crud-mongoose:findOne.

Readme

Keywords

none

Package Sidebar

Install

npm i crud-mongoose

Weekly Downloads

117

Version

1.0.46

License

MIT

Unpacked Size

49 kB

Total Files

17

Last publish

Collaborators

  • rthomps7
  • uhray