* Create routes /enrollment/gender and /enrollment/ethnic_group
* Add gender and ethnic_group to enrollment params in RQF

SQL now is using new schema

SQL is now compatible with DB schema in simcaqdb3

- Mount route under '/api/v1/enrollment/location'
- It should list the location names for enrollments and school classes
- The reponse format is a lista of objects with each object having a location_id (CEBES014T1) and a description

Signed-off-by: João Victor Risso <joaovictor.risso@gmail.com>

Signed-off-by: João Victor Tozatti Risso <jvtr12@inf.ufpr.br>

Change SQL queries in the following routes:

* city
* state
* region
* school
* enrollment

Also, change tests to reflect the new schema where appropriate. Specially in the case of cities, the primary key now becomes pk_cod_ibge (city IBGE code) and, therefore, any query that relies on the city id should use this value instead of the old primary key.

Aliases of fields that have different names were added to ensure compatibility with the UI of LDE.

Please refer to the database repository in order to understand how the current schema works. The corresponding ER diagram can be located there.

Sanitize the location in the result

Related: simcaq/SCRUM#18

Changing apps to follow suggested practices from documentation
and centralizing prefix /api/v1 in the app module.

All routes now use squel and the sql, query, response chain

Signed-off-by: João Victor Tozatti Risso <jvtr12@inf.ufpr.br>

Signed-off-by: João Victor Tozatti Risso <jvtr12@inf.ufpr.br>

Changes in this commit:
- Singularize all routes and components
- Separate different modules into their own source files
- Rename the API module to enrollment and add a more general API module
- Resources that are related to a route, are now under its scope. For
  example, the year range of the enrollment component is now under the
  route /v1/enrollment/year_range
- Implement a thenable function to execute the queries against the
  database to reduce the amount of duplicated code in the application. This
  approach also allows several concurrent queries to run concurrently in the
  same route.

Major modifications:
- Source code is now in the src directory
- All source files were adapted to ECMAScript6
- Add Gulpfile and Babel to transcompile the project from ES6 to ES5 (until Node.js fully supports ES6)
	- By running gulp one generates the build directory with the files transcompiled, from which the API can be run
- Add ESLint to check for syntax errors and enforce code standards
	- Javascript coding standard currently adopted is AirBnB with 4 spaces
- Replace monetdb package with monetdb-pool, which allows for concurrent queries over a pool of connections while maintaining the same interface

TODO (in order of priority, from high to low):
- Add Gulp tasks to handle automatic building, tests and running the server in order to deprecate using npm
- Implement decorator to execute the SQL queries and reduce code duplication.
- Implement SQL query builder (e.g. squel.js) to erradicate the need for embedding SQL directly into the code (which is error-prone).
- Change enrollments route not to use route chaining in order to decide which SQL query is appropriate to respond the user's request.
- Implement decorator for API responses and also reduce code duplication.
- Split up tests into and add more test cases.

Signed-off-by: João Victor Risso <jvtr12@c3sl.ufpr.br>

Changelog:
- Add education level to all queries

- Add route '/education_level' to query the available education levels.
  The response is a set of pairs (id, education_level)

- Change var declarations to their respective const and let counterparts

- Fix wrong foreign key name in the aggregation by states. When a
  adm_dependency_id parameter was supplied, the query would fail.

Changes:
- Add error handling in the database query function. When a database error
occurs, it is logged in the server and a plain text response is returned to the
user.
- Implement the building of queries inside the API, which removes the
  need to have stored functions/procedures on the database, which is not
  flexible or efficient to maintain/extend.
- Use route chaining to determine which function will build the query
  that needs to be executed. The last function is the one that actually
  sends the query to the database and send the response.

Future Work:
- Instead of returning a plain text response, return a HTTP 501 status,
  which is currently not possible. When one attempts to send such status
  the framework throws an error that the headers were already sent for
  the current request.
- Chaining based on the route might not be the best solution here.