DB and Collections

data follow the strict mode representation of BSON, e.g. {"_id": {"$oid":"568106e67d3126b74dc5f9af"}} for ObjectId
the root embeds collections
dbs have properties and embed collections and file buckets

use page and pagesize query parameters for pagination
specify np (no properties) query paramenter to exclude properties from the request, i.e only embedded data.

PATCH modifies the passed properties while PUT replaces the whole properties set
all write operation can use the dot notation and all mongodb update operators
server or client can optionally require to check the ETag to avoid ghost writes, but by default DELETE /db and DELETE /db/coll requires it

Request	Description
`GET /`	returns the paginated, existing dbs
`PUT /db {"descr": "my first db"}`	upserts the db `db`
`PATCH /db {"number": 1}`	modifies the db `db`
`DELETE /db If-Match:<Etag>`	deletes the db `db`
`GET /db`	returns the properties and the paginated collections and file buckets of `db`
`PUT /db/coll {"descr": "my first collection"}`	upserts the collection `db.coll`
`PATCH /db/coll {"number": 1}`	modifies the properties of the collection `db.coll`
`PUT /db/coll/_indexes/name_idx { "keys":{"name":1} }`	create the index `name_idx` in the collection `db.coll`
`DELETE /db/coll If-Match:<Etag>`	deletes the collection `db.coll`

Documents

use the strict mode representation of BSON, e.g. {"_id": {"$oid":"568106e67d3126b74dc5f9af"}} for ObjectId
use shardkey query parameter to eventually specify the shard keys, e.g. shardkey={"sk":"value"}

use page and pagesize query parameters for pagination
specify np (no properties) query paramenter to exclude the collection properties from GET /db/coll response.
keys query parameter defines the properties to return (projection), e.g. keys={"a":1, "obj.prop":1}
filter query parameter specifies a mongodb query, e.g. filter={"n":{"$gt":5}}
use id_type optional query parameter to properly identify docs /db/coll/1 (string) vs /db/coll/1?id_type=number

PATCH modifies the passed properties while PUT replaces the whole properties set
if the _id is not specified, it will be generated as a new ObjectId
all write operation can use the dot notation and all mongodb update operators
server or client can optionally requires to check the ETag to avoid ghost writes

Request	Description
`GET /db/coll`	returns the properties and the paginated documents of `db.coll`
`GET /db/coll/doc`	returns the document `doc`
`POST /db/coll {"_id": "doc", "n": 1}`	upsert a document in collection `db.coll`
`POST /db/coll [ {"n": 1}, ... , {"n": 10} ]`	bulk upsert documents in `db.coll`
`PUT /db/coll/doc {"$inc":{"obj.n":1}}`	upsert the document `doc`
`PATCH /db/coll/doc {"n": 1}`	modifies the document `doc`
`PATCH /db/coll/*?filter={"n":{"$lt":1}} {"n": 1}`	bulk modifies the documents matching the `filter` in `db.coll`
`DELETE /db/coll/doc`	deletes the document `doc`
`DELETE /db/coll/*?filter={"n":{"$gt":100}}`	bulk deletes the documents matching the `filter` in `db.coll`

Request	Description
`PUT /db/bucket.files {"descr": "my first file bucket"}`	upserts the file bucket `db.bucket`; the `.files` suffix marks it as a file bucket
`PATCH /db/bucket.files {"number": 1}`	modifies the properties of the file bucket `db.bucket`;
`GET /db/bucket.files`	returns the properties and the paginated files of `db.bucket`
`GET /db/bucket.files/file`	returns the properties of file `file`
`GET /db/bucket.files/file/binary`	returns the binary content of file `file`
`POST /db/bucket.files properties={"a": 1} file=<binary>`	creates an immutable file in bucket `db.bucket` note this is a multipart request
`PUT /db/bucket.files/file properties={"a": 1} file=<binary>`	creates the immutable file `file` in the bucket `db.bucket` note this is a multipart request
`DELETE /bucket.files/coll/file`	deletes the file `file`