firebase-st

A simple Smalltalk layer on top of the Google Firebase REST API, since no working gRPC/Protobuf libraries exists yet for Smalltalk.

Installation

Metacello new
    repository: 'github://psvensson/firebase-st:master';
    baseline: 'Firebase';
    load

Usage

You need to generate a service account credential file in JSON format, which can be done either from the Firebase project settings page or from the Google Cloud console; https://cloud.google.com/iam/docs/creating-managing-service-account-keys

certificateString := 'opencobalt-firebase-adminsdk.json' asFileReference readStream contents.

"push put and patch uses dictionaries as objects"
obj := Dictionary new.
obj at: #abc put: 'foobar'.

"queries are also objects, and which keys can be any of; startAt endAt limitToFirst limitToLast equalTo orderBy)"
query := Dictionary new.
query at: #orderBy put: '"$key"'.

firebase := Firebase new: certificateString .
rtdb := FirebaseRtdb new: firebase.
(rtdb putPath: '/foo/bar' obj: obj ) onSuccessDo: [ :res | Transcript show:'put result=',res asString;cr ].
(rtdb get: '/foo' query: query ) onSuccessDo: [ :res | Transcript show:'put result=',res asString;cr ].

Methods for rtdb

Rtdb is the original firestore real-time database. All data for a project resides in one large JSON structure, so be careful what you get or query, as you might get back quite a lot of data. Queries can only be made at one property at a time, which limits the use a bit.

• getPath: path "get data under path"
• deletePath: path "delete data from path"
• get: path "get data at path but not under"
• query: query "see example above"
• putPath: path obj: obj "add data under path"
• pushPath: path obj: obj "add data to array on object path"
• patchPath: path obj: obj "update data on path"

putPath adds new object to a path, pushPath pushes a new member of an array or an object of an existing path, pachPath replaces an existing path's proeprties with those proivded in the obj.

Firestore Usage

Firestore is the new firebase datastore with a more traditional noSQL way of organizing data. You create collections (tables) where you store documents (which can contain sub-collections). You can query on multiple properties and if an index needs to be created for the query, you will get a handy url back in an error message telling where to go to add one.

|certificateString firebase firestore data data2 data3 query res|
data := Dictionary new.
data at: 'hej' put: 'svejs2'.
data at: 'foo' put: #(a b c).
data2 := Dictionary new.
data2 at: #x put: #y.
data3 := Dictionary new.
data3 at: #deep put: #stuff.
data at: #z put: data3.
certificateString := 'service_account.json' asFileReference readStream contents.

firebase := Firebase new: certificateString .
firestore := Firestore new: firebase.

res := firestore create: 'bar' id: 'whoop2' document: data.
res onSuccessDo: [ :s | Transcript show:'Success: ',s asString;cr ].
res onFailureDo: [ :e | Transcript show:'Failure: ',e asString;cr ].
res

The above code in a playground will create a new collection named 'bar' (if it doesn't exist already) and adds a nontrivial JSON document with id 'whoop2'.

|certificateString firebase firestore data query res|
certificateString := 'service_account.json' asFileReference readStream contents.

firebase := Firebase new: certificateString .
firestore := Firestore new: firebase.

query := Dictionary new.
query 
	at: #from put: #bar
	;at: #select put: #(hej)	
	;at: #where put: #('hej.EQUAL.svejs')
	;at: #direction put: 'ASCENDING'.

res := firestore runQuery: query.
res onSuccessDo: [ :s | Transcript show:'Success: ',s asString;cr. ].
res onFailureDo: [ :e | Transcript show:'Failure: ',e asString;cr ].
res

This is a complex query with one 'where' clause which also filte rout just the proeprties you want to seee using 'select'. A list of possible where operators can be found here; https://firebase.google.com/docs/firestore/reference/rest/v1beta1/StructuredQuery#FieldFilter

query := Dictionary new.
query 
	at: #from put: #bar
	;at: #offset put: 3	
	;at: #direction put: 'ASCENDING'.

res := firestore runQuery: query.
res onSuccessDo: [ :s | Transcript show:'Success: ',s asString;cr. s inspect].
res onFailureDo: [ :e | Transcript show:'Failure: ',e asString;cr ].
res

And the above is a simple query which uses offset, to enable a simple form of pagination.

Methods for Firestore

• list: path pageSize: pageSize pageToken: pageToken orderBy: orderBy "All can be nil except for 'path'"
• create: path id:id document: document "path is the colelction name and id (if not nil) is the desired id. If nil, an id will be generated"
• get: path "Get the document on the path, like 'bar/whoop2'"
• patch: path document: document "Update an existing document"
• runQuery: query "Query properties can be any of from, select, where, direction, limit, offset and orderBy