[Game Service] Question fetcher component

[carlosfernandezmartinez]

Question fetcher component

The question fetcher is in charge of querying WikiData for questions. Remember that the main function of this component is processing the data returned from Wikidata and storing it into the database. The format of storing the data into the MongoDB database with the following schema (have a look at the user-model):

{
subject: {
type: String,
required: true,
},
image_name: {
type: String,
required: true,
},
answer: {
type: String,
required: true,
}
}

GET /generate with parameter the number of questions we want to generate, as well as the theme
Something like: /generate?n=100&sub=capital

Note that, in this initial implementation, only questions about capitals should be correctly generated and stored.

The image should be downloaded into a folder images inside the game service.

[carlosfernandezmartinez]

Update: it seems that mongoose assigns by default a '_id' field to each object of type ObjectId. This is the test I've done:

We can take that hash to uniquely identify each Question object in the db, so that the image file linked to that question also has the same name. I suggest this schema:

const mongoose = require('mongoose');

const questionSchema = new mongoose.Schema({
    subject: String,
    answer: String,
});

const Question = mongoose.model('Question', questionSchema);

module.exports = Question

since both image names and question id can be extracted from _id @Arquisoft/course2425_en2b.

[carlosfernandezmartinez]

I've found that it is not so hard to request generation together with the subject and amount of generated questions. I think that the best approach would be to directly pass a WikiData code as a request parameter (ex. Q515, which stands for cities).

GET /generate/wikiDataCode/amount - Ex. /generate/Q515/10 would download 10 images and store in the 'game' DB the answer to them.

The implementation right now is far from being perfect, it has a single query, so the subjects should be very specific. These are some subjects that I've checked are good enough to generate a reasonably large (around 10) amount of images with their answer:

• Q31629 - sports type
• Q515 - cities
• Q4989906 - monuments
• Q11344 - chemical elements

Therefore, this is the way in which the data is stored in the DB, together with the storage of the image inside in the /public/images folder. Please note that the __v field is also something MongoDB creates automatically, so I've took advantage of that to keep track of the version of the generated questions:

Exposing an endpoint to retrieve those images shouldn't be that hard, and the good part is that we can use the question id to retrieve the right image.

[carlosfernandezmartinez]

Thanks to @adrianmfuentes and his previous experience in node, I've discovered that I had an OOOoooold version of Node installed in my Ubuntu. I have already updated it and I will fix the existing errors soon.

[carlosfernandezmartinez]

😡 🐰 🏄‍♀️ DISCALIMER 💃 🥇 🧑‍🦯

Images can be obtained from url/image/[question_id].jpg
Here you got an example: