Server-Serflix

It's focused on user and profile management and response movies or series of TDMB, also it has Tensorflow as search through universal-sentence-encoder's model, moreover it has nodemailer to send messages, such as welcome, forgot-password, authenticate a user or reset a password.

Diagram database

Security levels in tokens

Most request need a token's role to access, moreover it's useful when a profile can't do something for user, on other hand each endpoint could have a token's role depend on the context and porpuse, like that:

Number	Role
0	Public access
1	Account access
2	Profile access
3	Authenticate user
4	Forgot password

API

All request have prefix /api/ and its parameters has format camelCase, the next table you can see the options about each Api:

Name	HTTP	Role	Option	Payload	Response
/user	POST	0	/register	username, email, password	msg, token
	POST	0	/login	username, password	msg, token
	PUT	0	/request/forgot-password	email	msg
	PUT	4	/forgot-password	newPassword	msg
	PUT	3	/auth	none	msg
	PUT	1,2	/password	oldPassword, newPassword	msg
	PUT	1	/username	newUsername	msg
	PUT	1	/premium	none	msg
	PUT	1	/free	none	msg
/profile	POST	1	/	idUser, name, img	msg, token
	GET	2	/	none	results[...]
	PUT	2	/	idProfile	msg, token
	POST	1	/login	none	name, img, results[...], goals[...]
	POST	2	/logout	name, img	msg
	POST	2	/password	name, img	msg
	PUT	2	/password	name, img	msg
	DELETE	2	/password	name, img	msg
	GET	1	/all	name, img	msg
/recommendation	GET	2	/	none	results
	GET	2	/all	none	last_viewed, recommendations
/movie	POST	2	/	idMovie	msg, goal
	GET	2	/	none	results[...]
	GET	2	/:id	none	result, is_favorite
	DELETE	2	/:id	none	msg
/serie	POST	2	/	idSerie	msg, goal
	GET	2	/	none	results[...]
	GET	2	/:id	none	result, is_favorite
	DELETE	2	/:id	none	msg
/score	POST	2	/movie/	idMovie, score, review	msg
	GET	2	/movie/:id	none	its_score, avg_score, scores[...]
	POST	2	/serie/	idSerie, score, review	msg
	GET	2	/serie/:id	none	its_score, avg_score, scores[...]
/search	GET	2	?query&times&itemsRelation	none	total_pages, total_results, results[]

When you make a request of search you must add queries for IA (tensorflow/tfjs-node and universal-sentence-encoder) can work to research on the data of database:

Where:

• query: only can string, this query reference about movie's description or title
• times: it's a number of data's area to research, if you want big area to get more presition, you could put high number but your request could be slow
• itemsRelation:: number of how to many item returned after research but there movies will have similar movies added each of them

Run

When you get up this server in development mode you just introduced in your terminal npm run dev but if you want to get up in production you just intruduced npm start where just uses node to get up it without any script by ts.

Now, this project is able to up on docker with the command docker compose up but if you have problem and change other line of code and you update change, you only put docker compose up --build for rebuild and compile changes

Email system

Now, it's able to send emails through nodemailer.

When you add a new user, automatic it sends an email to give welcome for user but if you want other templates for each method that required, you add a new folder on email's templates and add a index.html into it, also you can add a style.css.

Handling error

To avoid repeat many code's line and get different results, there's a middleware to centralizes the errors, moreover most errors has a kind of error such as PermissionDeniedError.

Variable's environment

For this project have to have many variable's environment to functionaly, to add those variables you must add new file with extension .env on root project and has the following variables:

If you don't work on docker, discard the docker's variables

NODE_ENV

This is the most important because asign role for the server, the only values are 'development', 'testing' and 'production'.

PORT

This variables is you can communicate to the backend, you could put 3000 or any value.

DOCKER_PORT

If you work on docker in this project you must put one port where other client can communicate to node.

TOKEN_TMDB

It must need to put a TMDB's token because some methods need to get some informaton by that service.

API_TMDB

This is just for organizing some line of code, for example https://api.themoviedb.org/3 that url is base of TMDB's API.

DB_HOST

This part is essential to connect your database, if you work on docker you don't put this variable.

DB_USER

It's name of user who connected for database.

USE_SSL

It's a boolean's state to use ssl or not, because on production it's better use ssl for database.

DB_NAME_FILE_SSL_CA

File's name of ssl to open it on connection to database.

DB_PASSWORD

It's essential put your password of your database, if you work on docker this password is now your password of your new database.

DB_NAME

You could put any name like serflix_db or my_new_database, but its name will where store all information.

DB_LOCAL_PORT

It's just database's port and it's commonly use 3306, but if you work on docker you could any number.

DB_DOCKER_PORT

When you work on docker is must be neccesary put one number because is where you can communicate with a client like DBeaver or workbench.

CLAVE_ENCRYPTATION

It's for encrypte the passwords with system AES.

SECRET_KEY_TOKEN

It's for JWS that is token's role.

EMAIL

it's to authenticate of service to send emails through nodemailer.

PASSWORD_EMAIL

it's to password of service that use nodemailer to authenticate.

API_SERFLIX

it's just reference in certain code's line, also it's important for email's controllers.

Common problems

If you have problem like Not found backend (don't care), To resolve this problem you navigate to node_modules where it's @tensorflow\tfjs-node\lib\napi-v9 and move tensorflow.dll from this directory to ../napi-v8, finally reset your backend.