I wrote a service that analyses videos with Google Cloud Video Intelligence
And I save the analysis results to the MongoDB with mongoose
This is the model I use (I've simplified everything to avoid confusion):
// Video.js
const mongoose = require('mongoose');
const videoSchema = new mongoose.Schema({
analysis_progress: {
percent: { type: Number, required: true },
details: {}
},
status: {
type: String,
enum: ['idle', 'processing', 'done', 'failed'],
default: 'idle'
}
});
module.exports = mongoose.model('Video', videoSchema);
When analyse operation ends, I call the function below and run update like this:
function detectFaces(video, results) {
//Build query
let update = {
$set: {
'analysis_results.face_annotations': results.faceDetectionAnnotations // results is the the test result
}
};
Video.findOneAndUpdate({ _id: video._id }, update, { new: true }, (err, result) => {
if (!err)
return console.log("Succesfully saved faces annotiations:", video._id);
throw err // This is the line error thrown
});
}
And this is the error I get:
Error: cyclic dependency detected
at serializeObject (C:\Users\murat\OneDrive\Masaüstü\bycape\media-analysis-api\node_modules\bson\lib\bson\parser\serializer.js:333:34)
at serializeInto (C:\Users\murat\OneDrive\Masaüstü\bycape\media-analysis-api\node_modules\bson\lib\bson\parser\serializer.js:947:17)
...
Solutions I tried:
Added {autoIndex: false} inside db config.
mongoose.connect(process.env.DB_CONNECTION, {useNewUrlParser: true, useUnifiedTopology: true, useFindAndModify: false, autoIndex: false });
Removing retryWrites=true from Mongo URI structure. (I didn't have that parameter in my connection URI already)
So, I think the source of the problem is that I am saving the whole test result but I don't have any other option to do that. I need to save as it is.
I am open to all kinds of suggestions.
Just as I guessed, the problem was that there was a cyclic dependency in the object that came to me from google.
With help of my colleague:
Then since JSON.stringify() changes an object into simple types:
string, number, array, object, boolean it is not capable of storing
references to objects therefor by using stringify and then parse you
destroy the information that stringify cannot convert.
Another way would be knowing which field held the cyclic reference and
then unsetting, or deleting that field.
I couldn't find which field has cycylic dependency so I used I JSON.stringfy() and JSON.parse() to remove it.
let videoAnnotiations = JSON.stringify(operationResult.annotationResults[0]);
videoAnnotiations = JSON.parse(videoAnnotiations);
Related
I'm trying to use Mongoose (MongoDB JS library) to create a basic database, but I can't figure out how to delete the documents / items, I'm not sure what the technical term for them is.
Everything seems to work fine, when I use Item.findById(result[i].id), it returns a valid id of the item, but when I use Item.findByIdAndDelete(result[i].id), the function doesn't seem to start at all.
This is a snippet the code that I have: (Sorry in advance for bad indentation)
const testSchema = new schema({
item: {
type: String,
required: true
},
detail: {
type: String,
required: true
},
quantity: {
type: String,
required: true
}
})
const Item = mongoose.model("testitems", testSchema)
Item.find()
.then((result) => {
for (i in result) {
Item.findByIdAndDelete(result[i].id), function(err, result) {
if (err) {
console.log(err)
}
else {
console.log("Deleted " + result)
}
}
}
mongoose.connection.close()
})
.catch((err) => {
console.log(err)
})
I'm not sure what I'm doing wrong, and I haven't been able to find anything on the internet.
Any help is appreciated, thanks.
_id is a special field on MongoDB documents that by default is the type ObjectId. Mongoose creates this field for you automatically. So a sample document in your testitems collection might look like:
{
_id: ObjectId("..."),
item: "xxx",
detail: "yyy",
quantity: "zzz"
}
However, you retrieve this value with id. The reason you get a value back even though the field is called _id is because Mongoose creates a virtual getter for id:
Mongoose assigns each of your schemas an id virtual getter by default which returns the document's _id field cast to a string, or in the case of ObjectIds, its hexString. If you don't want an id getter added to your schema, you may disable it by passing this option at schema construction time.
The key takeaway is that when you get this value with id it is a string, not an ObjectId. Because the types don't match, MongoDB will not delete anything.
To make sure the values and types match, you should use result[i]._id.
I'm running a Node.js server, connecting to a MongoDB database with mongoose.
Inside my controller, I have several methods that make operations to the database. One of them is this one:
async findMultiple(req, res) {
const [baseSkillsArray] = Array(req.body);
try {
// if there is not baseSkillsArray, skip
if (!baseSkillsArray) {
return res.status(200).send([]);
}
// find all baseSkills using the ids in the baseSkillsArray
const allBaseSkills = await BaseSkill.find({
_id: { $in: [baseSkillsArray.baseSkillArray] } //
});
console.log('test ' + allBaseSkills);
res.status(200).send(allBaseSkills);
} catch (error) {
console.error(error.message);
res.status(500).send('Server error find BaseSkills');
}
}
However, this returns me nothing. I did some debugging and I found the reason is the find id $in the array. So I tried hard coding a value, like '2', for instance.
// find all baseSkills using the ids in the baseSkillsArray
const allBaseSkills = await BaseSkill.find({ _id: { $in: ['2'] } });
No success. So I went to MongoDB Atlas, where my DB is stored. I tried filtering using the same line of code in my collections.
{ _id: { $in: ['2'] } }
Surprisingly, it returns my document as I wanted!
The issue is that I need to make it work with mongoose. Any ideas? Is this a known bug?
There is nothing wrong with the query, nor a bug regarding $in.
In fact, what's wrong is the actual collection name. I manually created a collection in MongoDB Atlas, called "baseSkills". However, mongoose by default transforms your collection name into lowercase and adds an "s" if your collection's name is not in the plural.
So every time I started my server, I noticed that there was a new collection called "baseskills". I assumed it was a bug and deleted it. Only after making this post that I realized the collection was there again.
So I exported the documents to this collection and my query was working fine.
FYI, there is a way to enforce the collection's name in mongoose. When you declare you model, add a second parameter to the Schema function called "collection". Here is an example:
const BaseSkillSchema = new mongoose.Schema({
_id: {
type: String,
required: true
}, ...
}, { collection: 'baseSkills' })
That's it! Sorry for the mess and thank you for your help!
you want to query over mongo db object ids. So you should create a new ObjectId to do that.
import {Types} from 'mongoose';
{ _id: { $in: [new Types.Object("2")] } }
Or if you have 2 ids one generated and one custom created as id then you can query without creating a new object.
{ id: { $in: ['2'] } }
I would like to know how to keep track of the values of a document in MongoDB.
It's a MongoDB Database with a Node and Express backend.
Say I have a document, which is part of the Patients collection.
{
"_id": "4k2lK49938d82kL",
"firstName": "John",
"objective": "Burn fat"
}
Then I edit the "objective" property, so the document results like this:
{
"_id": "4k2lK49938d82kL",
"firstName": "John",
"objective": "Gain muscle"
}
What's the best/most efficient way to keep track of that change? In other words, I would like to know that the "objective" property had the value "Burn fat" in the past, and access it in the future.
Thanks a lot!
Maintaining/tracking history in the same document is not all recommended. As the document size will keep on increasing leading to
probably if there are too many updates, 16mb document size limit
Performance degrades
Instead, you should maintain a separate collection for history. You might have use hibernates' Javers or envers for auditing for your relational databases. if not you can check how they work. A separate table (xyz_AUD) is maintained for each table (xyz). For each row (with primary key abc) in xyz table, there exist multiple rows in xyz_AUD table, where each row is version of that row.
Moreover, Javers also support MongoDB auditing. If you are using java you can directly use it. No need to write your own logic.
Refer - https://nullbeans.com/auditing-using-spring-boot-mongodb-and-javers/
One more thing, Javers Envers Hibernate are java libraries. But I'm sure for other programming languages also, similar libraries will be present.
There is a mongoose plugin as well -
https://www.npmjs.com/package/mongoose-audit (quite oudated 4 years)
https://github.com/nassor/mongoose-history#readme (better)
Maybe you can change the type of "objective" to array and track the changes in it. the last one of the array is the latest value.
Maintain it as a sub-document like below
{
"_id": "4k2lK49938d82kL",
"firstName": "John",
"objective": {
obj1: "Gain muscle",
obj2: "Burn fat"
}
}
You can also maintain it as an array field but remember, mongodb doesn't allow you to maintain uniqueness in an array field and if you plan to index the "objective" field, you'll have to create a multi key index
I think the simplest solution would be to use and update an array:
const patientSchema = new Schema({
firstName: { type: String, required: true },
lastName: { type: String, required: true },
objective: { type: String, required: true }
notes: [{
date: { type: Date, default: Date.now() },
note: { type: String, required: true }
}],
});
Then when you want to update the objective...
const updatePatientObjective = async (req, res) => {
try {
// check if _id and new objective exist in req.body
const { _id, objective, date } = req.body;
if (!_id || !objective) throw "Unable to update patient's objective.";
// make sure provided _id is valid
const existingPatient = await Patient.findOne({ _id });
if (!existingPatient) throw "Unable to locate that patient.";
// pull out objective as previousObjective
const { objective: previousObjective } = existingPatient;
// update patient's objective while pushing
// the previous objective into the notes sub document
await existingPatient.updateOne({
// update current objective
$set { objective },
// push an object with a date and note (previouseObjective)
// into a notes array
$push: {
notes: {
date,
note: previousObjective
},
},
}),
);
// send back response
res
.status(201)
.json({ message: "Successfully updated your objective!" });
} catch (err) {
return res.status(400).json({ err: err.toString() });
}
};
Document will look like:
firstName: "John",
lastName: "Smith",
objective: "Lose body fat.",
notes: [
{
date: 2019-07-19T17:45:43-07:00,
note: "Gain muscle".
},
{
date: 2019-08-09T12:00:38-07:00,
note: "Work on cardio."
}
{
date: 2019-08-29T19:00:38-07:00,
note: "Become a fullstack web developer."
}
...etc
]
Alternatively, if you're worried about document size, then create a separate schema for patient history and reference the user's id (or just store the patient's _id as a string instead of referencing an ObjectId, whichever you prefer):
const patientHistorySchema = new Schema({
_id: { type: Schema.Types.ObjectId, ref: "Patient", required: true },
objective: { type: String, required: true }
});
Then create a new patient history document when the objective is updated...
PatientHistory.create({ _id, objective: previousObjective });
And if you need to access to the patient history documents...
PatientHistory.find({ _id });
I have this schema:
var UserSchema = mongoose.Schema({
analytic: {
type: Object,
default: {
today:[],
weekly:[],
monthly:[],
yearly:[],
allTime:[]
}
}
});
let User = mongoose.model("bloger", UserSchema);
module.exports = {User};
and I am trying to save some data into one of the arrays like so:
User.findOne({username:username}, (e, user) => {
if (e) {
res.send('error fetching post')
}
else if (!user) {
res.send('no user found')
}
else if (user) {
user.analytic.today.push(req.body.visitor) // push the data object to the array
user.save((e, doc) => {
if (e) {
res.send(e)
}
if (doc) {
console.log('user saved')
res.send(doc)
}
})
}
})
})
I am getting the doc object on save() and not the e so I though it should have save it but it wasn't.
I have had a similar issue before this is because I am not defining a new Model I am just passing a JSON object.
Instead of saving the object you need to create a new model and save that.
Try creating a new model passing the save into it like below;
var newUser = new User(user);
newUser.save((e, doc) {
if (e) {
res.send(e)
}
if (doc) {
console.log('user saved')
res.send(doc)
}
});
Making sure you require the User Model inside the script.
Performing deep modifications in objects not in your schema makes Mongoose oblivious to those changes, preventing it from knowing what to save (and from making efficient atomic updates). The end result is that, when calling .save, Mongoose thinks there's nothing modified and therefore does nothing.
In your particular scenario, you have two options:
1. Add your analytics sub-arrays to your schema
This is the best option and allows for finer control of everything:
const UserSchema mongoose.Schema({
analytic: {
today: [{}],
weekly: [{}],
monthly: [{}],
yearly: [{}],
allTime: [{}],
}
});
With this, those arrays are now known to Mongoose and everything should work correctly.
Note that you don't need defaults in this case, as Mongoose is smart enough to create the arrays as needed.
2. Manually mark modified object as modified
If for any reason you don't want or can't modify the schema, you can manually mark the analytic object as modifies so Mongoose knows of it:
user.analytic.today.push(req.body.visitor) // push the data object to the array
user.markModified('analytic'); // mark field as modified
user.save(...);
This signals Mongoose that analytic or any of its children have changed and triggers an update when calling .save. Note however that Mongoose views this as a full change in the object, while with option 1 it can use $push instead.
I am using featherjs v2.0.3 with sequelize v3.29.0 and I just created three models, the third having a relationship to the other two.
I used the feathers-cli to generate services for each and then edit the model file of each.
So far, so good, the tables are created (using PostgreSQL), indexes are created, feathersjs takes care of the CRUD nicely, but not foreign keys yet.
So, when I try to tell feathersjs the relationship between the models, I get in trouble.
When I add role_permission.belongsTo(permissions) to the role_permission model, I get this error:
ReferenceError: permissions is not defined
As I've seen in the Sequelize documentation, the models are defined in the same "document", hence I suspect the problem is somewhere there, but I don't understand what need to be done.
Finally, here's the relevant parts of the model definitions of permission and role_permission:
// permission-model.js - A sequelize model
module.exports = function(sequelize) {
const permission = sequelize.define('permissions', {
permission_id: {
type: Sequelize.UUID,
defaultValue: Sequelize.UUIDV1,
primaryKey: true,
allowNull: false
}...
}, ...);
permission.sync();
return permission;
};
// role_permission-model.js - A sequelize model
module.exports = function(sequelize) {
const role_permission = sequelize.define('role_permissions', {
permission_id: {
type: Sequelize.UUID,
allowNull: false
}...
}, ...);
role_permission.belongsTo(permissions) //<-- undefined?
role_permission.sync();
return role_permission;
};
Do you have any pointers to help me solve this?
Thanks!
You need to import the permissions model. But it may or may not be defined yet.
Here's a method I discovered from #mrpatiwi on github to ensure every model is loaded before the associations are set up.
First, when you need to define a relationship add a classMethod called associate that accepts all the models and sets up the relationships.
module.exports = function(sequelize) {
const role_permission = sequelize.define('role_permisson', {
...
}, {
classMethods: {
associate(models) {
role_permission.belongsTo(models.permission);
},
},
});
// Don't add role_premission.sync() here
return role_permission;
};
Then, in src/services/index.js at the end of the module.exports function, add:
// Setup relationships
const models = sequelize.models;
Object.keys(models)
.map(name => models[name])
.filter(model => model.associate)
.forEach(model => model.associate(models));
sequalize.sync();
I had the same issue and resolved it simply by changing this:
role_permission.belongsTo(permissions)
to this:
role_permission.belongsTo(models.permissions)
(I used the Feathers CLI to generate the models today, 9 Aug 2018)